Progress 01/01/05 to 12/31/05
Outputs
Activities are related to information system development for production agriculture. While individual projects are described separately, there is considerable overlap in the various activities in terms of objectives and the usefulness of any results. Efforts continued with the COTMAN cotton management expert system to evaluate the potential for using site-specific datasets for optimizing selection of sampling sites. Historical site-specific data was used in selecting representative sample sites in four fields, three in Texas during 2005 and a historical Arkansas dataset where intensive COTMAN observations had been recorded in 2003. A technique was developed for using multiple years of normalized yield data to identify regions of the field that had been consistently average in production. The mode was selected as a more appropriate descriptive statistic for normalization due to the common occurrence of yield maps skewed toward lower yield values. The use of the
historical data resulted in less variability in the range of cotton development parameters when considering various combinations of averaging two sampling points. A second activity focused on maintaining seed cotton quality during storage. Lint and seed quality can be maintained for long periods while stored in a module if the seed cotton is protected from moisture. This protection requires both a well shaped module that will quickly shed rainfall, and a cover that will resist moisture penetration. Module cover testing included continued standardized testing of water impact penetration, hydrostatic resistance and outdoor weathering resistance. Small-scale module testing was performed in order to evaluate module cover resistance to contact from seed cotton foreign matter and wind effects. Vinyl and film module covers allowed zero to minimal water penetration. Woven, poly covers allowed minimal to high water penetration. One finding was that module shape is critical for prevention of
water accumulation on the tarp surface. We evaluated a prototype feedback system to improve module shape. Visco-elastic performance of seed cotton indicates an increased mass of seed cotton should be placed along the centerline of modules to create a rounded surface allowing rainwater runoff. Observations were made of modules that experienced an extended period of rain on the Southern High Plains in 2004. Lint loan value decreased on average $440 per module when worn module covers were used. Poorly shaped modules caused lint loan value to decrease on average $200. Modules with good covers and shapes had no decrease in quality compared to modules ginned before the rainfall. A third project was focused on the development of systems for variable application of pesticides in orchards. Efforts were directed toward the development of a decision support system that would generate site-specific application maps and a sprayer control system that could implement the maps. Collaborative efforts
are underway with commercial providers of existing precision agriculture hardware and software to extend the capabilities of those systems.
Impacts
Results from studies under this project can have a significant impact on management decisions in agriculture. The techniques demonstrated for selecting sampling sites for use with the COTMAN expert system provide the opportunity for producers to improve field management with the use of COTMAN while minimizing the cost of obtaining that data. The impact of inappropriate protection of seed cotton during storage has been documented and the resulting lost of income to producers is estimated to have been $12.5 million or more for a single rainfall event in the Southern High Plains. These losses were entirely preventable, and the improvement of module shape and cover quality emphasized in this project can prevent these losses.
Publications
- Akbar, M. A.*, A. L. Kenimer, S. W. Searcy and H. A. Torbert. 2005. Soil water estimation using electromagnetic induction. Transactions of ASAE 48(1):129-135.
- Stabile, M.C.C. and S.W. Searcy. 2005. Use of site-specific data for COTMAN site selection. Proceedings of the 2005 Beltwide Cotton Conferences. pp. 1934-45.
- Simpson, S. L. and S. W. Searcy. 2005. The benefits of replacing used module covers. 2005 Proceedings of the Beltwide Cotton Conferences, pp. 3029-44.
- Stabile, M. C. C. and S. W. Searcy. 2005. Real-time plant height mapping and variable rate application of growth regulators. 2005 Proceedings of the Beltwide Cotton Conferences. pp. 532-44.
- Hardin, IV, R. G. and S. W. Searcy. 2005. Design and evaluation of an operator feedback system for the module builder. 2005 Proceedings of the Beltwide Cotton Conferences. pp. 574-80.
- Searcy. S. W. 2005. Plant height impacts on spindle picked lint quality. 2005 Proceedings of the Beltwide Cotton Conferences. pp. 560-7.
- Clouse, R. L. and S. W. Searcy. 2005. Cotton model simulation of applied water-yield relations for the Texas high plains. 2005 Proceedings of the Beltwide Cotton Conferences. pp. 568-73.
- Simpson, S. L. and S. W. Searcy. 2005. Cotton module cover standardization. ASAE Paper No. 051096. American Society of Agricultural Engineers. St. Joseph, MI.
- Bronson, K. F., Booker, J. D.,Officer, S. J., Lascano, R. J., Maas, S. J., Searcy, S. W., Booker, J.. 2005. Apparent electrical conductivity, soil properties and spatial covariance in the U.S. Southern High Plains. Precision Agriculture. 6(3): 297-311.
|
Progress 08/12/99 to 08/11/05
Outputs
Activities are related to information system development for production agriculture. While individual projects are described separately, there is considerable overlap in the various activities in terms of objectives and the usefulness of any results. The final year of the project to evaluate the COTMAN cotton management expert system with site-specific datasets for optimizing selection of sampling sites was completed. The site selection protocol established in project was used to determine sites within fields in Burleson Co., Texas. Consistent with previous results, the use of the historical data resulted in less variability in the range of cotton development parameters when considering various combinations of averaging two sampling points. The results of this past year are being combined with previous results to determine final recommendations for cotton producers who wish to use COTMAN in their management strategy. A second activity focused on maintaining seed cotton
quality during storage. Lint and seed quality can be maintained for long periods while stored in a module if the seed cotton is protected from moisture. This protection requires both a well shaped module that will quickly shed rainfall, and a cover that will resist moisture penetration. The module shape feedback system was revised and tested by placing it on a module builder operated by a custom harvest crew. The system used magnetic sensors to monitor the movement of the carriage, replacing the ultrasonic sensor used in previous years. The new design worked well for maintaining the position of the carriage. The system accurately displayed the shape of the module being formed, as determined by measurement of module shape following completion of the module. The workers were positive about the value of the feedback system, stating that it was particularly valuable when operating at night and for determining the shape at the far end of the module builder. Further analysis of a dataset of
modules that experienced an extended period of rain on the Southern High Plains in 2004 was performed. The number of modules examined was expanded, and the analysis extended to include the impact of poorly protected seed cotton on ginning rates. Producers lose value through the decreased lint loan value of approximately $400 per module when worn module covers were used and $200 when modules were poorly shaped. In the gin, the combination of a poor cover and shape reduced turnout ratio from 34 to 26 percent, and ginning rate from 42 to 19 bales per hour. Well protected modules (good covers and shapes) had no decrease in quality compared to modules ginned before the rainfall.
Impacts
Results from studies under this project can have a significant impact on management decisions in agriculture. The techniques demonstrated for selecting sampling sites for use with the COTMAN expert system provide the opportunity for producers to improve field management with the use of COTMAN while minimizing the cost of obtaining that data. The impact of inappropriate protection of seed cotton during storage has been documented and the resulting lost of income to producers is estimated to have been $12.5 million or more for a single rainfall event in the Southern High Plains. These losses were entirely preventable, and the improvement of module shape and cover quality emphasized in this project can prevent these losses.
Publications
- Hardin, R. G. and S. W. Searcy. 2006. Development of an operator feedback system for the module builder. Proceedings of the 2006 Beltwide Cotton Conferences pp. 411-19.
- Simpson, S. L., S. W. Searcy and A. Brashears. 2006. Performance of module covers under extended storage. Proceedings of the 2006 Beltwide Cotton Conferences pp. 627-37.
- Clouse, R. W. and S. W. Searcy. 2006. Selection of COTMAN sampling points using historical site-specific data sets. Proceedings of the 2006 Beltwide Cotton Conferences pp. 475-86.
- Clouse, R. W. and S. W. Searcy. 2006. Comparison of soil water relationships of cotton simulation models in the Texas High Plains. Proceedings of the 2006 Beltwide Cotton Conferences pp. 454-68.
- Hardin, R. G. and S. W. Searcy. 2006. Development of an operator feedback system for the module builder. ASABE Paper No. 061081. American Society of Agricultural and Biological Engineers. St. Joseph, MI.
|
Progress 01/01/04 to 12/31/04
Outputs
Activities are related to information system development for production agriculture. While individual projects are described separately, there is considerable overlap in the various activities in terms of objectives and the usefulness of any results. Our cotton height sensing system was used with the COTMAN cotton management expert system to evaluate the potential for using site-specific datasets for optimizing selection of COTMAN sampling sites. Recommended sampling practices are typically not followed due to time and expense. Using plant height or yield maps to select regions with average conditions can improve the accuracy of management decisions made with COTMAN. The different selection methods were evaluated based on the date of peak nodal development and date of nodes above white flower (NAWF) less than 5. Results showed that, when selecting sampling sites using site-specific data, average field conditions can be described with fewer than the recommended number
of sampling locations. Regardless of method used for site selection, the date of peak nodal development varied less than two days, and the date of NAWF>5 by four days. Another project activity was focused on maintaining seed cotton quality during storage. Lint and seed quality can be maintained for long periods while stored in a module if the seed cotton is protected from moisture. This protection requires both a well shaped module that will quickly shed rainfall, and a cover that will resist moisture penetration. Characterization of visco-elastic properties of seed cotton was the first step toward development of an operator feedback system to aid in construction of well shaped modules. The feedback system utilizes knowledge of applied compressive force and deformation for seed cotton. Factors were tested to determine their effects on seed cotton during compression, creep loading, and recovery. Initial density did not significantly affect the compressed density. Picker harvested
cotton was compressed to a greater density than stripper harvested cotton, but expanded more during recovery, resulting in similar final densities. Multiple compressions increased the density, but the increase was not physically significant after the third compression. Higher moisture content resulted in a slightly increased seed cotton density after compression. These physical properties were used to develop an operator feedback system for the module builder. The feedback system was developed and underwent limited testing. Water resistance and weathering of new and used module covers have been studied over the past two years. Laboratory testing has proven that, during rainfall and when rain water ponds on top of the cover, small holes in module covers allow considerable amounts of water to penetrate the cover and enter the seed cotton. The effectiveness of newer and older module covers was determined at a west Texas cotton gin in 2004. Modules that had a good shape and good covers
retained good color grades and profits. Modules that were both poorly-built and had poor covers resulted in the worst color grades and the largest losses, ranged on average from $30 to $478.
Impacts
Results from studies under this project can have a significant impact on management decisions in agriculture. The techniques demonstrated for selecting sampling sites for use with the COTMAN expert system provide the opportunity for producers to obtain the demonstrated cost reductions available with the use of COTMAN while minimizing the cost of obtaining that data. The knowledge gained about the physical properties of seed cotton during compression is enabling a feedback system that can potentially prevent significant loss of lint and seed value. In addition, publication of that work is adding knowledge in an area where no previous research has been published. The knowledge gained about the performance of cotton module covers after various levels of exposure and use will allow the development of a cover replacement strategy for that will retain significant value for cotton producers.
Publications
- Harmel, R. D., A. L. Kenimer, S. W. Searcy, and H. A. Torbert. 2004. Runoff water quality impact of variable rate sidedress nitrogen application. Precision Agriculture 5: 3, 247-261
- Fox, G. A.*, G. J. Sabbagh, S. W. Searcy and C. Yang. 2004. An automated soil line identification routine for remotely sensed images. Soil Science Society of America Journal 68: 4, 1326-1331.
- Akbar, M. A.*, A. L. Kenimer and S. W. Searcy. 2004. Estimating Soil Profile Depth With Apparent Electrical Conductivity for a Texas Vertisol. Transactions of ASAE 47(4): 1087-1092.
- Clouse, R. L. and S. W. Searcy. 2004. Sensitivity of gossym cotton growth model to soil variability and site-specific management. 2004 Proceedings of the Beltwide Cotton Conferences, pp. 874-884. National Cotton Council of America, Memphis, TN.
- Geiger, D.W., S. W. Searcy and D.D. Fromme. 2004. Precision agriculture: a tool for COTMAN. 2004 Proceedings of the Beltwide Cotton Conferences, pp. 760-769. National Cotton Council of America, Memphis, TN.
- Simpson, S.L., C.B. Parnell and S.W. Searcy. 2004. Systems analysis of ginning season and seed cotton transport. 2004 Proceedings of the Beltwide Cotton Conferences. pp. 1021-1026. National Cotton Council of America, Memphis, TN.
- Simpson, S.L. and S.W. Searcy. 2004. Performance of module covers in resisting moisture penetration. 2004 Proceedings of the Beltwide Cotton Conferences. pp. 3040-3048. National Cotton Council of America, Memphis, TN.
- Hardin, IV, R.G., S. W. Searcy and S.L. Simpson. 2004. Visco-elastic properties of seed cotton and their effect on module shape and density. 2004 Proceedings of the Beltwide Cotton Conferences. pp. 748-754. National Cotton Council of America, Memphis, TN.
|
Progress 01/01/03 to 12/31/03
Outputs
Activities under this project are related to sensor and information system development and control, with application in production agriculture. While individual projects are described separately, there is considerable overlap in the various activities in terms of objectives and the usefulness of any results. Most of the research activity in 2003 was focused on improving cotton production through the application of precision agriculture techniques and lint quality preservation during module storage. The previously developed cotton height sensing system was used to evaluate the potential for integrating the COTMAN cotton management expert system into a subfield management strategy. COTMAN was developed as a whole field management tool, and requires an extensive sampling effort to accurately characterize a field. The time and expense of this sampling has limited the adoption of COTMAN by producers. Combining the ability of the height measurement system to map plant
development over the entire field with the knowledge base of COTMAN has the potential to reduce costs and extend the utility of COTMAN. Irrigated and dryland fields in two regions of were monitored in 2003. Subfield regions were identified from the height maps and sampled for COTMAN parameters. Results showed that the fields in Wharton Co. did not have enough variability to justify subfield management. In Burleson Co., the variability was greater, and justified differential management. Results also showed that using height maps to select sites of average development, the total number of samples required could be reduced to one third of the recommended number. A second project focused on the development of cotton modules and covers that could maintain seed cotton quality during storage. Efforts were focused on the characterization of visco-elastic properties of seed cotton (as necessary background for development of an operator feedback system) and the resistance of commercial module
covers to moisture penetration. Results on the seed cotton property analysis show that the dominant factor in determining module shape is the distribution of cotton mass. With a known operating pressure, the mass can be determined by measuring the displacement of the tramper ram. A three month exposure of module covers to ambient weather resulted in dramatically different moisture resistance, even within the same model of cover. The tests document the dramatic increase in moisture penetration as covers experience greater wear. New and lightly used covers provided excellent protection and would result in small changes in moisture content during a four hour rain storm. Each increased level of wear (moderate, heavy, abused) would resulted in a projected 10 percent increase in moisture content within the top layer of cotton.
Impacts
The integration of cotton height mapping with the COTMAN management tool demonstrates the utility of collecting height maps while performing field operations such as spraying. Using the height information to select sampling sites for plant observation can potentially reduce the cost of using COTMAN. Multiple height maps for a field will provide rate of growth information for the different field regions, and could potentially be used in management of plant growth regulators. Examination of cotton modules stored in gin yards showed that approximately one half of those modules had a shape that would cause water to collect on top of the module cover. The combination of a poorly formed module and a cover that does not resist moisture penetration will result in high moisture content cotton, and likely decrease in value. Based on observations of modules and the ages of the covers used on them, the potential for significant quality loss is high. The seed cotton quality
preservation project will continue with an effort to document actual quality losses and to inform the producers and ginners of the potential losses that are likely occurring.
Publications
- G.A. Fox*, G.J. Sabbagh and S.W. Searcy. 2003. Radiometric normalization of multitemporal images based on image soil lines. Transactions of ASAE. 46(3):851-859.
- Clouse, R.W. and S.W. Searcy. 2003. Analysis of site-specific irrigation with GOSSYM simulation model. 2003 Proceedings Beltwide Cotton Conferences pp. 625-631. National Cotton Council of America, Memphis, TN.
|
Progress 01/01/02 to 12/31/02
Outputs
Activities under this project are related to sensor and information system development and control, with application in production agriculture. While individual projects are described separately, there is considerable overlap in the various activities in terms of objectives and the usefulness of any results. Much of the research activity in 2002 was focused on the IMPACT Center, a site for long term evaluation of information based management tools and techniques. The site provides a laboratory for the evaluation of newly developed techniques and an area where these technologies can be demonstrated to producers on a scale similar to their situations. Investigations undertaken in 2002 include conservation tillage, variable rate seeding and cotton variety trials. The conservation tillage study includes tillage treatments of conventional, ridge till and no tillage with a corn/cotton rotation on both irrigated and rain-fed production areas. 2002 was the first year of a
multi-year project. The variable rate seeding of corn was examined by comparing four different fixed seeding rates with varying seeding rate by soil derived polygons. Cotton varieties were compared in large plots. All studies were evaluated by harvested weights from each treatment replication and by yield mapping. A new project on maintaining seed cotton quality during module storage was initiated in 2002. The economics of cotton ginning are causing fewer gins to process more cotton. As a result, seed cotton remains in modules for longer periods, and is exposed to potential water damage if the module covers allow water to penetrate. The objectives of the project are to develop an electronic feedback system to module operators that will improve the creation of moisture resisting module shapes, and to evaluate the performance of new and used module covers in resisting water penetration. The project was initiated in the fall, and the intent is to have a prototype feedback device
developed for evaluation in the 2003 harvest season.
Impacts
The activities under this project are intended to address real problems faced by Texas producers. Producers are faced with a wide array of new technologies and management strategies, but have little information about which of those will work under Texas conditions. The inter-disciplinary, long term investigations underway at the IMPACT Center will provide some of those answers. The development of new devices and techniques for protecting cotton in module storage addresses a need clearly documented during the 2001 and 2002 harvest seasons when significant numbers of cotton modules were damaged by heavy rains.
Publications
- No publications reported this period
|
Progress 01/01/01 to 12/31/01
Outputs
Activities under this project are related to sensor and information system development and control, with application in production agriculture. While individual projects are described separately, there is considerable overlap in the various activities in terms of objectives and the usefulness of any results. Much of the research activity in 2001 was focused on a new research facility located near College Station. The IMPACT Center was established to provide a site for long term evaluation of information based management tools and techniques. The site provides a laboratory for the evaluation of newly developed techniques and an area where these technologies can be demonstrated to producers on a scale similar to their situations. I am serving as one of the managers of the IMPACT Center, and am responsible for maintaining the GIS database, implementing the precision agriculture technologies, and collaborating with other faculty who are conducting studies at the site. A
cotton height sensor was previously combined with a biological model of cotton physiology to create a real-time, biomass-based variable rate application system for cotton growth regulators. A cotton physiological model called MPERT utilized the number of plant nodes and height to predict the needed application rate during the early linear phase of plant growth. During 2001, the height sensor was used to track the growth rate of cotton as a function of field location. Cotton yield mapping systems for cotton strippers and pickers were evaluated in 2001. Cooperative evaluations were performed with the commercially available Ag Leader system on a cotton picker used in the Brazos River Valley, and an experimental prototype that was installed on cotton strippers was evaluated in the Central Blacklands area and the Southern High Plains.
Impacts
Both the real-time chemical application system and the cotton yield mapping system for strippers have potential for adoption by producers in the near term. The establishment of the IMPACT Center will facilitate the transfer of information to local producers on both the utility of new research techniques, such as the variable rate Pix application, and the value of adopting commercially available technologies such as yield mapping. In 2001, the initial year of operation, precision agriculture techniques were used to document the effect of variety and tillage practices.
Publications
- Beck, A. D.*, S. W. Searcy and J. P. Roades*. 2001. Yield data filtering techniques for improved map accuracy. Applied Engineering in Agriuclture 7(4):23-431.
|
Progress 01/01/00 to 12/31/00
Outputs
Activities under this project are related to sensor and information system development and control, with application in production agriculture. While individual projects are described separately, there is considerable overlap in the various activities in terms of objectives and the usefulness of any results. A crop height sensor was combined with a biological model of cotton physiology to create a real-time, biomass-based variable rate application system for cotton growth regulators. A model called MPERT utilized the number of plant nodes and height to predict the needed application rate during the early linear phase of plant growth. A computer was programmed to process the plant height sensor data, calculate the needed rate of growth regulator and transmit the desired set points to a sprayer controller. Variable rate application of Pix Plus was made on cotton based upon plant size. Experiments were conducted with fixed rates, variable rates and no regulator. The
performance of the sensing and application systems was acceptable, with minimal differences between plant heights on adjacent passes of the machine. Two fields of approximately 20 acres each were sprayed with the system. No significant differences in yield were found between the three treatments. Cotton yield mapping systems for cotton strippers were evaluated in 2000. The Texas A&M experimental system based on weighing and a commercial system with modified optical flow sensors were installed on a stripper. Both systems were evaluated against manual yield estimates and basket dump weights. The commercial system was found to have dramatically improved performance over the previous years installation. The modified sensors were sealed to prevent blockage of the light beams, and were placed in the outlet of the field cleaner. The resulting yield maps had approximately 80% of the points where map and hand sampled yields were equivalent. The experimental system had experienced failures in
hardware and software during testing. After eliminating data that was clearly flawed, 60% of the compared yield points were equivalent to hand samples. A technique was developed to allow comparisons of a time series of aerial images of crops. Producers would like to compare a series of images to determine the rate of growth and relative health over the course of a growing season. However, comparison of multi-band images taken at different times is generally not possible without some form of intensity correction, normally standardized reflectance panels placed in the image. For much commercial remote sensing, standard panels are not available. We developed a technique for normalizing images based upon the soil line theory. The algorithm is based on the assumption that the soil line should be standard from one image to the next. The soil lines are determined in each available image, and then normalized to one selected image. In comparison to histogram matching, a commonly used
radiometric correction technique, the soil line transform normalization technique resulted in significantly better performance in matching the expected change in pixel reflectance with changes in crop biomass.
Impacts
Both the real-time chemical application system and the cotton yield mapping system for strippers have potential for adoption by producers in the near term. Although no advantage to variable rate Pix application was shown in our studies in 2000, work in other states (MS, NC) has shown economic value, if the information cost for determining the correct rate are not too high. With further development, the existing system could be economically feasible. The availability of a reasonably accurate yield mapping system for cotton strippers has been a limiting factor for cotton producers in the Southwest who are interested in precision farming techniques. The result in 2000 with the modified sensor design indicates that useful yield maps can be obtained with cotton strippers. The success of the soil line transform normalization technique means that more accurate crop observations can be made with a series of multi-band remote images.
Publications
- Kataoka, T., A. D. Beck and S. W. Searcy. 2000. An algorithm for estimating plant height using a light screen. ASAE Paper No. 003063. American Society of Agricultural Engineers. St. Joseph, MI.
- G.A. Fox, G.J. Sabbagh, S.W. Searcy, and M. Eriksson. 2001 Radiometric normalization of multi-temporal images using the Soil Line Transformation (SLT) technique. Journal of Remote Sensing of the Environment. In press.
- Searcy, S. W. and A. D. Beck. 2000. Real-time assessment of cotton plant height. Proceedings of the 5th International Conference on Precision Agriculture. In press.
- Yu, M., E. Segarro, H. Li, R. J. Lascano, C. Chilcutt, L. T. Wilson, K. Bronson and S. W. Searcy. 2000. The economics of precision agricultural practices in cotton production. Proceedings of the Beltwide Cotton Conferences. pp. 369-375.
- Ruble, M. K. and S. W. Searcy. 2000. A GIS interface for cotton growth models that incorporates midseason adjustments. Proceedings of the Beltwide Cotton Conferences. pp. 456-459.
- Searcy, S. W., J. P. Roades and A. D. Beck. 2000. Cotton yield mapping: Texas experiences in 1999. Proceedings of the Beltwide Cotton Conferences. pp. 404-408.
- Torbert, H. A., S. W. Searcy, A. L. Kenimer and J. Roades. 2000. Precision farming effects on corn productivity and water quality. Proceedings of the Second International Conference on Geospatial Information in Agriculture and Forestry. January 10-12, 2000, Lake Buena Vista, FL. ERIM International Inc. Ann Arbor, MI. Vol. 2, pp. 495-502
|
Progress 01/01/99 to 12/31/99
Outputs
Activities under this project are related to sensor and information system development and control, with application in production agriculture. While individual projects are described separately, there is considerable overlap in the various activities in terms of objectives and the usefulness of any results. A crop height sensor sensor developed in 1998 was refined and expanded for addition crops. The sensor was originally developed for assessing the height of cotton plants. In 1999 the sensor was expanded for use with corn and sorghum, and the height estimate algorithm was modified to use a probability based filter for determining crop height. The sensor was successfully used in corn and cotton in 1999. Variable rate applications of harvest aid chemicals was made on cotton based upon the size of the plants. Experiments were conducted in the Brazos Valley, where favorable rainfall conditions resulted in rank cotton growth in some areas. The plant height sensor was
used to map the cotton height. The field was classified as normal or rank growth, and the map used to determine the rank areas of the field where two applications of defoliant were needed. The normal cotton received one application. Last season moisture stress caused easy defoliation in 1999, and as a result, the defoliation treatments did not show a significant difference. As expected, the rank cotton which received only one application had a lower defoliation rating, but the difference was not enough for statistical significance. The number of green bolls in the treatment areas was significant, with the rank cotton with only one application having significantly more green bolls than the rank cotton with only one application. The linking of geographic information systems (GIS) with crop growth models would allow the site-specific management of the crop to have a predictive capability. The use of sensing techniques that could generate in-season calibrations of the model results could
improve the accuracy of the system. In 1999, two efforts were made to link cotton models to the Arc-View GIS. Arc-View was successfully integrated with ICEMM (Integrated Crop Ecological Management Model), a cotton model optimized for the Texas Coastal Bend region. In addition, work is underway to integrate the GOSSYM cotton growth model with Arc-View. This work is in cooperation with USDA-ARS, which initiated the linkage of these software packages. This research is still underway. Cotton yield mapping systems for cotton strippers were evaluated in 1999. The Texas A&M experimental system based on weighing and a commercial system using optical flow sensors were installed on a stripper. Both systems were evaluated against manual yield estimates and basket dump weights. The commercial system was found to have a highly varying calibration factor. The experimental system had improved reliability, and user friendliness, but had a significant number of yield estimates that deviated from the
hand samples.
Impacts
The project with the most immediate potential for impact is the development of an accurate cotton yield mapping system. Yield maps generated with the system described the variability in yield observed in the field. The plant height sensor can potentially affect the rate and manner in which cotton chemicals are applied. The linkage of a cotton growth model with a GIS can provide growers with more accurate estimates of plant status, yield and economic thresholds.
Publications
- Stewart, J. H., S. W. Searcy and J.A. Landivar. 1999. Cotton-height measurement for variable- rate Pix (mepequat chloride) application. Proceedings of the Beltwide Cotton Conferences. pp. 547-554.
- Landivar, J.A., S. W. Searcy and J. H. Stewart. 1999. A variable-rate chemical application system for cotton in south Texas. Proceedings of the Beltwide Cotton Conferences. pp. 851- 853.
- Beck, A. D., J. P. Roades and S. W. Searcy. 1999. Post-process filtering techniques to improve yield map accuracy. ASAE Paper No. 99-1048. American Society of Agricultural Engineers. St. Joseph, MI.
- Jessip, J. L., S. W. Searcy and G. J. Sabbagh. 1999. An algorithm for the delineation of management units based on hydrologic properties. ASAE Paper No. 99-3126. American Society of Agricultural Engineers. St. Joseph, MI.
|
Progress 01/01/98 to 12/31/98
Outputs
Activities under this project are related to sensor development and control, with application in production agriculture. While individual projects are described separately, there is considerable overlap in the various activities in terms of objectives and and the usefulness of any results. The development of a crop biomass assessment system for use on field machinery was completed in 1998. The system is based on the continuous measurement of cotton plant height along a row within the field. Cotton biomass can be estimated using knowledge of crop phenology, emergence date and measured crop height. Two height measurement techniques, contact with feelers and non-contact with optical sensors, were evaluated. The optical and contact systems had a 95% confidence intervals for predicted values of 3 and 5 cm, respectively. A primary use of biomass estimates would be the determination of optimum application levels of cotton growth regulators. An evaluation of the sensitivity
of the optimum rate to the height prediction errors resulted in a mean application rate error of 7.3 ml/ha. Since this level is less than the accuracy of most sprayers, the height measurements are sufficiently accuracy. A two product variable rate application system based on direct injection was assembled on a high clearance sprayer and used to apply cotton growth regulators. Fixed and variable rates of mepequat chloride and an experimental growth regulator were applied in strips across a 40.5 ha field. The height of the cotton in each strip was measured manually, and the variable rate treatments were applied based on the height. The height/application rate relationship was determined using a cotton growth model. The field experienced a drought during 1998, and as a result, no significant differences were found between the treatments. However, the variable rate application of mepequate chloride did require approximately 30% less material. Efforts were continued on the development of a
cotton yield mapping system for a cotton stripper based on a weighing technique. Load cell mounts were redesigned for improved accuracy and easier installation. Alternative data processing algorithms were developed and are under evaluation. Maps from the 1998 harvest season are still being processed and evaluated
Impacts
(N/A)
Publications
- Searcy, S. W. and J. P. Roades. 1998. Evaluation of cotton yield mapping. Proceedings of the Beltwide Cotton Conferences. pp 636-639.
- Searcy, S.W. 1998. Evaluation of weighing and flow-based cotton yield mapping techniques. Proceedings of the 1998 Precision Agriculture Symposium. in press.
|
Progress 01/01/97 to 12/31/97
Outputs
Activities under this project are related to sensor development and control, with application in production agriculture. While individual projects are described separately, there is considerable overlap in the various activities in terms of objectives and and the usefulness of any results. Investigations were initiated into the development of a crop biomass assessment system that could be used on field machinery during normal field operations. The biomass is being estimated using knowledge of crop phenology, emergence date and measured crop height. Two methods, contact with feelers and non-contact with optical sensors, are being researched. The systems have been evaluated under laboratory conditions, with the contact sensor providing the greater reliability. The optical system experienced difficulty with sensitivity and ambient light interference. Both systems will be field evaluated in 1998 on both corn and cotton. A variable rate application system was assembled on
a high clearance sprayer and used to apply cotton growth regulator (mepequat chloride). The use of the growth regulator is intended to limit vegetative growth, thus creating a more uniform plant size. The height of the cotton was measured manually over a 100 acre field, and the growth regulator was applied in response to the height. The height/application rate relationship was determined using a cotton growth model. Application rates varied from 4-8 oz/ac. The application map was created in GIS software, and the rates were automatically controlled using a GPS receiver. The COV for crop height was decreased from 12.2 on June 3 to 7.6 on August 5. Due to drought during the latter part of the growing season, the value of the growth regulator was limited. This study is continuing in 1998. A new project was initiated in October, 1997 on the development of a multispectral reflectance sensor to predict nitrogen stress. Activities initiated include the measurement of nitrogen stress on
different colored corn varieties and the development of a multispectral sensor that could be programmed to measure the most useful wavelengths. This project will continue through 2000.
Impacts
(N/A)
Publications
- Landiar J. A. and S. W. Searcy. 1997. Evaluation and Evelopment of a Precision Management System for the Management of Crops in South Texas: Annual Report. Texas Agricultural Experiment Station, Corpus Christi, TX.
|
Progress 01/01/96 to 12/30/96
Outputs
Research continued under the BARD sponsored project, "Adaptive Sorting of Agricultural Produce" to develop a programmable, multi-spectral imaging system. An accousto-optical tunable filter (AOTF) has been selected as the active, tunable element. The entire wavelength selectable, imaging system consists of a CCD camera, the AOTF, a radio frequency generator, a digital to analog voltage output board and a computer. Testing with the system was limited. A reflectance spectrophotometer was purchased and installed. The spectrophotometer was used to determine the reflectance spectra of normal and defective carrot flesh for two varieties. The BARD project has been completed, and no further development of the system is anticipated at this time.
Impacts
(N/A)
Publications
|
Progress 01/01/95 to 12/30/95
Outputs
Research continued under the BARD sponsored project, "Adaptive Sorting of Agricultural Produce" to develop a programmable, multi-spectral imaging system. An accousto-optical tunable filter (AOTF) was selected as the active, tunable element. The AOTF wavelength is selected by imposing a radio frequency signal on the crystal. The entire wavelength selectable, imaging system consists of a CCD camera, the AOTF, a radio frequency generator, a digital to analog voltage output board and a computer. Software was written to control image formation so that the operator can select the desired wavelengths with mouse actions. Considerable effort went into the calibration of the system. Since this is one of the earliest applications of AOTF for imaging processes, several problems were experienced with out-of-spec. performance of the device. Subsequent testing with the system has shown that image degradation can result from inaccurate creation of the desired radio frequencies.
Development and testing work is continuing through the final year of the BARD project.
Impacts
(N/A)
Publications
|
Progress 01/01/94 to 12/30/94
Outputs
Research activity has been concentrated on the BARD funded project entitled "Adaptive sorting of fresh produce." These efforts have been concentrated in two areas, the establishment of a large database for the evaluation of fresh market carrots and the development of a multi-spectral, programmable imaging sensor. Carrots were obtained from a packing shed in the Rio Grande Valley of Texas. 2788 carrots were analyzed for quality features such as size, shape, splits and surface defects. Using a 650 nm optical filter and an improved version of the INSPECT software previously developed at TAMU, inspection was performed at a rate of two carrots per second with an accuracy of more than 90%. This compares favorably with the 85% accuracy typically found in manually graded carrots. Various technologies were examined for construction of the multi-spectral imaging sensor. Desirable characteristics for tile design included rapid switching between wavelengths, coverage of the
visible and near-infrared ranges, and ease of processing the multi-spectral image, given a rapid motion of the object being inspected. Technologies considered included opto-acoustical tunable filter (AOTF), linearly variable filters and grating spectrophotometers. The AOTF was determined to be the best solution, and was purchased for combination with a CCD camera. The system has been constructed and is currently undergoing testing. Incorporation into the INSPECT system will take place in 1995.
Impacts
(N/A)
Publications
|
Progress 01/01/93 to 12/30/93
Outputs
Research has continued into the application of machine vision technologies. In August 1993, a BARD funded project on adaptive sorting of produce was initiated. This project is concentrating on advancement of the previously developed carrot grading system and the development of an adaptive multi-spectral imaging sensor. The previously developed algorithm for grading fresh market carrots used run length encoding and adaptive thresholding to identify defects such as surface rot, splits, cracks, forking or over/under size. The original computer code was designed to function with a line scan camera, but was developed and tested using a two dimensional camera. During this year, efforts have been made to convert the system to operation with a line scan camera. Modifications are still underway at this time of reporting. In the coming year, the completed system will be tested with a large quantity of carrots in order to debug the algorithms and to develop dataset that will be
used by Israeli cooperators. Another objective of the project is to develop a multi-spectral imaging sensor. Biological materials have unique interactions with electromagnetic energy, and those interactions are frequently used to determine properties of the material. For example, braised apple flesh can be detected by measuring the reflectance of near infrared wavelengths. Unfortunately, simple reflectance can be affected by many other variable, making single wavelength measurements less sensitive.
Impacts
(N/A)
Publications
|
Progress 01/01/92 to 12/30/92
Outputs
Research has continued into the application of machine vision technologies. These efforts have concentrated in two areas, the automated inspection of fresh market carrots and automated guidance of field machines. Efforts toward the inspection of fresh market carrots have been directed toward publishing previous work and writing proposals for external funding. Manuscripts published or accepted for publication cover research on neural network classification of carrot tip shape, algorithms for description of carrot shapes and defects, comparison of statistical and neural network classifiers for quality measures, and accuracy of automated carrot grading. A proposal was written to the Binational Agricultural Research and Development Foundation for funding of research on a highly flexible machine vision based inspection process which could adapt itself to the sample population which was to be inspected. This proposal includes the concept for a multi-spectral sensor which
could select the optimum wavelengths for feature discrimination under software control, and a neural network based classifier which could monitor the products being inspected and select the most appropriate feature vector for classification. The efforts directed toward the automated guidance of field machines has resulted in a significant improvement in the algorithm used to determine guidance information.
Impacts
(N/A)
Publications
|
Progress 01/01/91 to 12/30/91
Outputs
Research under this project is concentrating on the application of machine vision technologies for agricultural problems. Effort has been concentrated in two areas, development of grading algorithms for fresh-market carrots and guidance of agricultural vehicles. Work on the development of carrot grading algorithms, funded by BARD, was completed this past year. Algorithms were successfully developed for measuring length and diameter of carrots, shape, brokenness, and surface defects. Results indicate that these algorithms provide improved grading accuracy over manual methods now employed. Further work is necessary to develop a prototype grading system. Further funding will be sought for this project. The research on guidance algorithms for agricultural vehicles has concentrated on generalizing the earlier guidance algorithms. Previous work had concentrated on situations where straight row crops were available. This algorithm has been generalized to incorporate curved
rows and situations where only a single edge is available. A computer simulation of the scene in front of the tractor has been generated and is being used to test the accuracy of the guidance algorithm. Testing of the guidance algorithm continues with simulated and videotaped images. In-field testing of the improved algorithm is planned for 1992.
Impacts
(N/A)
Publications
|
Progress 01/01/90 to 12/30/90
Outputs
Work was concentrated on two projects. First of these was funded by BARD and concerned the grading of fresh market carots with machine vision. The second was continuing effort on development of automated guidance for field machinery. Significant progress was made on the grading of fresh market carrots. Techniques were developed to classify carrots tip shapes using both statistical and neural network classifiers. Results showed the neural network to be superior to statistical classification. Other work included the incorporation of the size, shape and defect determinations into a single grading system. Results of that investigation indicates that accuracy is at least as good as human graders. A determination of machine grading accuracy was limited by the accuracy of humans to determine the appropriate grade for individual carrots. Work continued on the development of a automated guidance system for agricultural field machinery. Test were conducted with a
steering control system installed on the tractor. Dynamic problems were identified and the systems is under modification. New guidance determination algorithms are being developed which will generalize the situations in which the automated steering system could be used.
Impacts
(N/A)
Publications
|
Progress 01/01/89 to 12/30/89
Outputs
Work under this project has consisted of continued development of the vision guided tractor and the development of a carrot inspection system. These two separate projects will be described individually. Development of the machine vision guided tractor consisted of the installation and initial testing of the guidance system on a production model tractor. Earlier efforts had been confined to laboratory testing of the guidance algorithms. During 1989 the control system was put in place on the tractor and used in testing in steering response in the field. This testing was designed to development equations which would relate known wheel angles to the resulting turn within the field. Final analysis of this data is in progress. Further testing is needed on the control system and its interaction with the guidance algorithm. During the coming year tests will be made on standardized courses and under field conditions. The development of a carrot inspection system has progressed
with emphasis on the algorithms that will detect the appropriate characteristics. Algorithms have now been developed to detect carrot length, carrot width, carrot shape and defects such as rot and black spots. Testing of these algorithms is continuing. The development of a specialized sensor is in progress. This sensor is designed to allow rapid inspection of the carrot as they pass a sensing point in a conveying system. The inspection algorithms and sensor will be integrated and tested on handling equipment under development in Israel.
Impacts
(N/A)
Publications
|
Progress 01/01/88 to 12/30/88
Outputs
This Hatch project encompasses the activities on several subprojects. They willbe reported individually. The development of an automated guidance system for agricultural machines has continued with publication of previous results and the development of the steering controller. This project is partially supported by Ford/New Holland by the loan of a tractor. The power steering system of this tractor has been modified to accept a controller which can be computer controlled. The modified tractor has been used for empirical testing of the tractors turning performance. Development of an automated shrimp deheading system has continued, with further investigations into the optimum features that can be visually determined. Although the ratio of carapace length to total length is potentially the most accurate, it is too difficult to determine accurately with the vision system. Location of the hepato pancreas is the best method. Efforts are continuing to interest industry
to participate in this project. A project was initiated to develop a machine vision based carrot grader. In order to provide the basic knowledge needed, spectrophometric studies were done of various carrot defects. This information was used to determine wavelengths that are most appropriate for discrimination between defective and normal carrots. Machine vision Algorithms are being developed to utilize this spectral information for the location and evaluation of carrot defects.
Impacts
(N/A)
Publications
|
Progress 01/01/87 to 12/30/87
Outputs
Work has been continuing on the full range of activities covered under this project. More emphasis has been placed on the computer vision topics than the data logging research within the past year. Computer vision work has continued on automated guidance of agricultural vehicles. Improvements have been made to algorithms for determining guidance and heading offset parameters for row crop images. Initial work was begun for development of a steering controller for implementation of the automatic guidance system with testing planned for 1988. Work has been completed in utilization of computer vision to identify proper location for trimming stem material from commercial processing carrots. Algorithms were developed to identify and locate the stem/root joint. Tests showed that processing speeds of 6 to 8 carrots per second were obtainable. Investigations are ongoing for vision-controlled deheading of shrimp. Algorithms were developed to identify optimum cutting
location and the vision and waterjet system technologies were integrated and tested. A processing rate of 0.5 shrimp per second with cutting accuracy of 1 mm was achieved. Further testing and improvements are planned for 1988. Work has continued on data logging research with an effort to generalize the data logger so that it can be used for many applications. A data acquisition system was developed for a tractor mounted cone penetrometer capable of storing up to 32 tests in the field. The data logger was also utilized to record rainfall.
Impacts
(N/A)
Publications
|
Progress 01/01/86 to 12/30/86
Outputs
Work has been continuing on the full range of activities covered under this project. During the past year, more emphasis has been placed on computer vision activities with a lessening effort on data logging research. The tractor performance monitoring and optimization research has been continuing with both the John Deere 4440 and 4450 tractors. Emphasis has been on the use of the on-board data logging equipment to monitor tractive performance. Indirect measures of engine load are also being investigated. A front wheel assisted tractor has been shown to have 5-15% improvement in work rate and fuel efficiency over 2 wheel drive operation. Work was completed on the ultrasonic based sugarcane cutting height control system. After field testing, it was determined that the signal to noise ratio for the ultrasonic signals was too low to reliably extract cutting height information. As a result the controller experienced cutting height oscillations of +/-5 cm, much greater
than the desired level of +/-1 cm. Funding for this work was discontinued and the project has been terminated. An on-animal data logger was field tested during 1985. The first printed circuit prototype was debugged and changes made to improve reliability. A second printed circuit board is being developed and will be extensively tested during 1987. Testing to date has provided valuable animal activity data for range scientists.
Impacts
(N/A)
Publications
|
Progress 01/01/85 to 12/30/85
Outputs
Research has continued on a range of projects involving microcomputer based control and data acquisition. In addition, research has progressed on computer vision applications, with new areas of application planned. The tractor performance optimization project has continued with extensive field testing of the John Deere 4440 on commercial farms. Data has been gathered on a variety of implements and a feedback system for providing efficiency information to the operator was tested. A three point hitch hynamometer was constructed and used in field testing. A second tractor with front wheel assist, John Deere 4450, has been instrumented and will be used to investigate on-farm utilization of front wheel assist. Improved signal processing algorithms were developed for use with the ultrasonic-based sugarcane cutting height control system. The new algorithms were designed to make the system less sensitive to noise in the ultrasonic echoes. Preliminary tests indicated
potential for success and further tests will be done in the 1985-85 harvest season. An on-animal data logger was constructed for use in field experiments by range scientists. The initial prototype was field tested and a second prototype has been manufactured for use in expanded experiments during 1986. The use of computer vision for automated guidance of agricultural machinery has continued with efforts concentrated on extraction of guidance information from recorded images.
Impacts
(N/A)
Publications
|
Progress 01/01/84 to 12/30/84
Outputs
Work has continued on Several applications under this project including tractor performance monitoring and optimization (TPMO), sugarcane cutting height control, on-animal data logging, computer vision and crop establishment. The instrumented tractor has been used on commercial farms and data on energy usage has been analyzed. Results show a significant fuel savings which could result from proper operation. An information feed back system has been installed on the tractor and is undergoing test. Results show potential fuel savings of 15-40%. Tests were conducted with the modified hydraulic system on a sugarcane harvester. Ability of the crop gathering roller to float on the ground was successfully demonstrated. Control system simulations identified acceptable algorithms for processing the ultrasonic echoes used to detect the cut height of stalks. Field testing of the ultrasonic unit has shown poor correlation between the echoes and actual stalk height under
operating conditions. Work is continuing on methods of improving the echo signal and improved signal processing. A data logger has been developed for use on free roaming animals. It is being used in range management research and is undergoing initial testing. The data logger is monitoring animal biting stepping and micro-environment. A computer vision laboratory has been established. Current application is the development of automated guidance for agricultural vehicles.
Impacts
(N/A)
Publications
|
Progress 01/01/83 to 01/30/84
Outputs
Research is progressing on several projects including tractor performance monitoring (TPM), sugarcane cutting height control and automated transplanting. An additional topic which was initiated in 1983 was the utilization of image processing technology for control of agricultural equipment. In the TPM project the hardware and software were completed for an on-tractor, microcomputer-based data acquisition system. In-field testing was initiated and will be continued during 1984. Modifiations have been made to a commercial sugarcane harvester in order to accommodate a height control system. Work is continuing on that machine. Computer simulations of control algorithms have been conducted and will be implemented in the controller. Work on the automatic transplanting system continues. A transplant forming machine has been developed and is in debugging stages. A new project to investigate the application of vision systems to agriculture has been initiated. Current
efforts are directed towards acquiring the necessary equipment to perform the research.
Impacts
(N/A)
Publications
|
Progress 01/01/82 to 12/30/82
Outputs
Research is continuing in several areas including automated control of surface irrigation, tractor performance monitoring, cutting height control for a sugarcane harvester and automated transplanting. In initial field tests, water sensors were connected to a computer via handwired telemetry and water advance times in furrows were recorded. Radio telemetry valves and sensors have been ordered and will be utilized in testing during 1983 in testing of the rate of advance algorithm being used to determine water flow rates. As part of a cooperative agreement with John Deere Co., a tractor has been provided to facilitate measurement fuel use, engine speed, wheel slip, axle torque and drawbar load. A commercial data logger has been installed and is being used while a microcomputer is being constructed and programmed for this application. All hardware has been obtained and software development is underway. Development of a cutting height controller for sugarcane harvesters
has been proceeding based on laboratory work. Modifications of a commercially available ultrasonic detector have been made and field tests will be conducted in early 1983. The work has consisted of developing software to appropriately control the ultrasonic transmitter and receiver. Research has commenced on an automated transplanting system based on a bandoleer concept. The system consists of bandoleer forming machinery of use in a gree house and an automated transplanter for field use.
Impacts
(N/A)
Publications
|
Progress 01/01/81 to 12/30/81
Outputs
Work has been initiated on utilization of microcomputers in three different applications: automated control of surface irrigation, monitoring of tractor performance, and cutting height control for sugarcane harvesters. Computer hardware has been acquired and sensors developed for a hardwired telemetry system to detect and control the rate of advance of water down a furrow. This system will be installed and used to test various control algorithms. The controlling software is being developed and system testing will be conducted spring/summer 1982. Based on information gathered, a larger radio telemetry system will be installed for larger scale testing. Development of a microcomputer controlled data collection system for tractor performance has been initiated. A cooperative agreement has been reached with John Deere Co. for the loan of a tractor for use in this research. A system for data recording which will not require operator interaction is being constructed.
The project will provide documentation of current use cycles and aid in optimization of tractor performance. The monitoring and control of cutting height on sugarcane harvesters with an ultrasonic detector is being investigated. Past work has shown a commercially available ultrasonic level controller to be capable of detecting sugarcane stalk height. Modifications are being made to the ultrasonic device to adapt it to the sugarcane applications.
Impacts
(N/A)
Publications
|
|