USING GROUND-BASED SPECTRAL MEASUREMENTS FOR VEGETATION CHARACTERIZATION

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: COMPLETE
• Funding Source: HATCH
• Reporting Frequency: Annual
• Accession No.: 0216142
• Project Start Date: Sep 15, 2008
• Project End Date: Sep 14, 2013
• Program Code: [(N/A)]- (N/A)

Recipient Organization
UNIVERSITY OF FLORIDA
G022 MCCARTY HALL
GAINESVILLE,FL 32611

Performing Department
Gulf Coast Research and Education Center, Bradenton

Non Technical Summary
Vegetation studies using spectral measurements have been the subject of a considerable amount of research in the last few decades. Of the topics that have been explored, understanding vegetation characteristics and response has been of particular concern due to the dynamic nature of vegetation and the need to accurately describe its decompositions and response (Gao & Goetz, 1995; Daughtry, et al., 2000; Goetz & Prince, 1996; Liu et al., 1997). Variations in the amount of emitted, absorbed, and reflected electromagnetic energy arise from the particular plant species, the condition of the plant, and the in-situ effects of the data collection process. Each of these factors contributes to a plant's spectra to form a unique pattern, or signature. Although identification of plants using lab spectral measurements is important for many applications, the integration between field and remotely sensed spectral measurements is essential for the success of many applications. The need to establish the relationship between plant status and spectral measurements is of interest to many state and national agencies as well as private industry. Being able to utilize spectral data for the early identification of invasive plant species or optimizing the use of fertilization in agricultural lands could save the economy billions of dollars, not to mention the positive impact on human health and the environment. Agencies such as the water management districts are interested in not only being able to accurately classify agricultural crops for accurate estimation and projection of water demand, they are also interested in studying the response of crops under different management schemes to recommend ones with the best water conservation and environmentally-sound practices. Fortunately, recent advances in spectral measurement technology made the acquisition and use of such technologies feasible. However, this put tremendous pressure on the research community to develop field experiments and analysis algorithms needed for implementation. Understanding the relationship between plant physiology and spectral response through field measurements will improve the remote sensing data acquisition and analysis techniques, and enable accurate use in large-scale implementations. In this context, the proposed use of a hyperspectral imaging sensor currently utilized for airborne remote sensing data acquisition, accompanied by field spectrometer measurements to analyze vegetation response, will help close the gap between field and remote sensing data collection. Furthermore, successful development could lead to the development of a ground-based hyperspectral imaging system that assists agriculture and natural resources applications along roads or waterways.

Animal Health Component

(N/A)

Research Effort Categories

Basic

(N/A)

Applied

(N/A)

Developmental

(N/A)

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
402	7210	2080	25%
402	7310	2020	25%
402	7310	2080	25%
402	7210	2020	25%

Knowledge Area
402 - Engineering Systems and Equipment;

Subject Of Investigation
7310 - Experimental design and statistical methods; 7210 - Remote sensing equipment and technology;

Field Of Science
2020 - Engineering; 2080 - Mathematics and computer sciences;

Keywords

spectral

image sensors

hyperspectral analysis

ground-based sensors

spectral separability

dimension reduction

Goals / Objectives
This proposal aims at developing methodologies and algorithms to study the relationship between plant status under different management treatment and ground-based spectroscopy. In achieving such objective, the following goals are identified: I. Perform a comprehensive literature review This objective aims at performing a comprehensive literature review of the experiments and results of the research that utilize field and imaging spectroscopy for vegetation classification and characterization. This literature should include both field and lab research results in addition to recent remote sensing technology applications. By reviewing research procedures and results and physical plant characteristics, the field experiments can be designed to address the practical challenges of the implementation. II. Acquire, Calibrate and Test spectral measuring equipment Two sets of equipment are proposed in this research: field spectrometers and hyperspectral imaging sensor. This task involves equipment requisition, configuration, calibration, and testing. The development of geometric correction for the imaging sensor is also proposed in this task. III. Design field experiment to collect spectral measurements This objective aims at designing a field experiment to collect spectral measurements using field spectrometer and hyperspectral imaging sensor. The design includes identifying plant management procedures, frequency of observations and accompanied reference lab measurements. IV. Use spectral imagery to quantify plant status Measurements will be acquired and analyzed. Image correction algorithms will be implemented. Hyperspectral image analysis techniques such as dimensionality reduction using Minimum Noise Fraction Transform algorithm (Chen, et al., 2003) and endmembers identification procedures will be implemented. The results will be analyzed and compared to standard lab tests for assessment.

Project Methods
The objectives of this project will be achieved through the following methods: I.1 Perform Literature review This includes mining the literature for field and lab experiments, analysis methodologies and algorithms, types of plants involved, etc. The output literature review should be comprehensive enough to pave the road for the following research tasks. II.1 Acquire field spectrometer and hyperspectral imaging sensor The methodology proposed in this project utilizes field spectrometer and hyperspectral sensor measurements in the VNIR region of the spectrum. Acquiring and configuring the equipment is a crucial task in the proposed implementation. This includes equipment selection and testing in addition to preparing the logistics needed for successful and reliable data collection in the field. The proposed sensor is a Visible-Near Infrared (VisNIR) 450-900nm linear array hyperspectral sensor. II.2 Testing different procedures to reduce at sensor radiance observations to reflectance Reducing at at-sensor radiance observations to reflectance value is necessary for identifying certain spectral patterns. Several methods can be used to achieve this task. Some of these methods will be tested and compared. II.3 Perform preliminarily geometric correction to ground hyperspectral imaging sensor Many hyperspectral imaging sensors utilize rotating mirrors to capture static scenes. This type of image acquisition is accompanied by many geometric errors. This task is concerned with understanding and correcting these geometric errors. III.1 Design plant treatment procedure A testing site will be used for the study. Different plant types/plant treatment methods are suggested for analysis. Other field measurements will also be suggested. The frequency and time for data acquisition will be decided in this task. III.2 Design spectral data collection procedure Procedures to collect field data are identified and tested in this step. This includes sensor configuration and mounting and powering logistics, capturing elevation(s), calibration techniques, and quality control procedures. IV. Collect data using field spectrometers and hyperspectral imaging sensor Actual measurements using field spectrometer and hyperspectral imaging sensor will be captured several times according to established design. Initial in-situ checks of the measurements will be performed. The data are stored and prepared for the image analysis step. IV.2 Analyze acquired spectral data and reporting results Observations will be analyzed by removing dark current and stray light errors from the signal after utilizing the reference object observations. The reflectance values will be computed, and statistically analyzed for multiple plant samples. The data will be analyzed to quantify the effect of different plant treatment procedures using field spectral measurements. The potential for extrapolating obtained results to airborne and space-born remote sensing hyperspectral, multispectral, and LiDAR data will be investigated.

Progress 09/15/08 to 09/14/13

Outputs
Target Audience: Target audience are growers and scientists interested in utilizing spatial and image analysis in precision agriculture applications. Local government agencies interested in spatial data collection technologies are also part of this project audience. A symposium was conducted to highlight used technologies. The symposium was attended by growers, spatial technology service providers, researchers and local government representatives. Public audience and growers were also reached through two magazine publications discussing project activities and outcomes. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Several undergraduate and graduate students participated in this project. Through this participation, they received professional and technical training. Students from different departments including the School of Forest Resources and Conservation and Computer and Information Science and Engineering department participated in activities related to integrating hardware equipment (imaging and navigation sensors), image acquisition and calibration software development, conducting image acquisition and analysis experiments, and analyzing acquired images and navigations data. Two magazine articles were published for the public community highlighting research output and accomplishments. How have the results been disseminated to communities of interest? National and international conference presentations were delivered by project team in addition to several refereed journal publications (see the Products Section). These presentations/publications aimed at disseminating project output to interested audience including researchers and the public. Two articles were published in regional magazines (see the Products Section). These articles targeted the growers’ community in addition to the general public. A symposium was organized in 2009 by the PI to discuss the advances in precision agriculture technology and applications, where this project activities were highlighted. The symposium was well attended with audience from the growers, researchers, and local government communities. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? This project aims atdeveloping methodologies and algorithms to study the relationship between plant status under different management treatment and ground-based spectroscopy. This was fulfilled through designing, developing, testing and implementing a multi-purpose ground based imaging system. The system is capable of capturing high spatial resolution and hyperspectral imagery for vegetation studies. Each of the 4 goals identified to achieve the project aim was fulfilled. The following is a description of the accomplishments realized for each goal: I. Perform a comprehensive literature review: A comprehensive literature review was performed to survey hyperspectral imaging sensors, system design, and use in vegetation analysis applications. The review identified a gap in close-range hyperspectral sensor availability for vegetation analysis such as assessing plant stress and precision agriculture. The review identified the technical difficulties associated with implementing such systems, which include issues related to integration with navigation sensor for accurate georeferencing, image quality, platform design and mounting, and image analysis algorithms. II. Acquire, Calibrate and Test spectral measuring equipment We developed the Vegetation Mobile Mapping System (VMMS), a low-cost hyperspectral sensing system that is supported by consumer-grade digital camera(s). The system was developed using off-the-shelf imaging and navigation components mainly for ground-based applications. The system integrates a variety of components including timing and positioning GPS receivers and an Inertial Measurement Unit (IMU). The system was designed to be modular and interoperable allowing the imaging components to be used with different navigation systems. The technique used for synchronizing captured images with GPS time was presented. A relative radiometric calibration technique utilizing images of homogeneous targets to normalize pixel gain and offset parameters was used. An empirical spectral calibration method was used to assign wavelengths to image bands. Data acquisition parameters to achieve appropriate spatial coverage were presented. III. Design field experiment to collect spectral measurements The system was tested in ground-based data collection, filtering, and analysis experiments that included water quality and vegetation studies.Several field experiments were conducted to: (1) develop and test algorithms to filter acquired imagery (2) quantify tomato vegetation health and (3) estimate chlorophyll-a (chl-a) concentration in aquaculture ponds. Field experiments were accompanied by ground truthing through vegetation assessment and/or lab analysis of collected samples. More than 30 images were collected for the chl-a experiment and 15 images for the tomato vegetation analysis. Several factors such as collection time, weather condition, the use of calibration targets were accounted for in designing the field experiments. IV. Use spectral imagery to quantify plant status An algorithm based on Wavelet Transform and Adaptive Frequency domain filtering was modified, tested and implemented to accommodate high resolution hyperspectral imagery. The algorithm was compared to other existing ones and showed superior filtering results with minimal introduced artifacts. Two experiments were conducted, as stated above, to quantify tomato vegetation subject to disease and to evaluate Chl-a concentration in aquaculture ponds. The algorithms used in the analysis utilized individual narrow bands acquired by our hyperspectral sensor in the red and infrared regions of the spectrum. The tomato vegetation experiment was conducted to evaluate the use of our systems to detect and quantify bacterial disease in tomato vegetation. Our experiments involved capturing imagery on three different dates along the plant cycle at 4, 6, and 8 weeks age for five different trials under different management conditions and was accompanied by field evaluation accomplished by plant pathology experts. The Normalized Difference Vegetation Index (NDVI) using two bands (λ = 668 and 736 nm) was used in this analysis.These two bands represent the chlorophyll-a strong absorption and fluorescence bands in the red and infrared region and define the two end of the chlorophyll red edge area of the spectrum. In the second experiment to estimate Chl-a concentration as an indicator of algae bloom and a measure of water quality, the images were acquired simultaneously with water samples from 30 aquaculture ponds. A 3-band index utilizing bands (λ = 680, 700 and 769nm) produced a correlation coefficient of 0.93 with chl-a concentration obtained by lab analysis of water samples.

Publications

• Type: Journal Articles Status: Published Year Published: 2011 Citation: Abd-Elrahman A., R. Pande-Chhetri, G. Vallad, 2011. Design and Development of a Multi-Purpose Low-Cost Hyperspectral Imaging System. Remote Sensing 3(3):570-586.
• Type: Journal Articles Status: Published Year Published: 2011 Citation: Panda-Chhetri, R., A. Abd-Elrahman, 2011. De-striping hyperspectral images using wavelet transform and adaptive frequency domain filtering. International Society of Photogrammetry and Remote Sensing Journal 66(5):620-636.
• Type: Journal Articles Status: Published Year Published: 2011 Citation: Abd-Elrahman, A., M. Croxton, R. Pande-Chhetri, G. Toor, S. Smith, and J. Hill, 2011. Chlorophyll-a estimation using in-situ hyperspectral imaging system in aquaculture water bodies. ISPRS Journal of Photogrammetry and Remote Sensing 66: 463-472.
• Type: Journal Articles Status: Published Year Published: 2013 Citation: Panda-Chhetri*, R.and A. Abd-Elrahman, 2013. Filtering high-resolution hyperspectral imagery in a maximum noise fraction transform domain using wavelet-based de-striping. International Journal of Remote Sensing 34:2216-2235
• Type: Conference Papers and Presentations Status: Published Year Published: 2010 Citation: Development of high-resolution mobile hyperspectral imaging system. ISPRS Commission IV Topics IV/3 Mapping from High Resolution Data, Orlando, FL, November 14-17, 2010.
• Type: Conference Papers and Presentations Status: Published Year Published: 2010 Citation: Estimation of water quality parameters in freshwater aquaculture ponds using hyperspectral imaging system. ASPRS Specialty Conference Topics: Ocean Monitoring and Modeling, Orlando, Florida, November 14-17, 2010
• Type: Journal Articles Status: Published Year Published: 2009 Citation: Abd-Elrahman, A., M. Thornhill, M. Andreu, and F. Escobedo, 2010. A community-based urban forest inventory using online mapping services and consumer-grade digital images. International Journal of Applied Earth Observation and Geoinformation 12:249-260.
• Type: Journal Articles Status: Published Year Published: 2013 Citation: Panda-Chhetri*, R.and A. Abd-Elrahman, 2013. Filtering high-resolution hyperspectral imagery in a maximum noise fraction transform domain using wavelet-based de-striping. International Journal of Remote Sensing 34:2216-2235.
• Type: Other Status: Published Year Published: 2009 Citation: Abd-Elrahman, A., 2009. Vegetation mobile mapping system is the latest in the precision agriculture revolution. Research Report at Citrus - Vegetable Magazine, June/July, p. 30.
• Type: Other Status: Published Year Published: 2013 Citation: Jim Frankowiak, 2013. The Cutting Edge, In the field, Hillsboroughs Agricultural Magazine, Sep15-Oct15, 2013, Vol. 9, Issue 11, pp35.
• Type: Other Status: Published Year Published: 2010 Citation: Abd-Elrahman, Foster, E., A., Pande-chhetri, Wiseman, A., and Vallad, G. 2010. Preliminary assessment of tomato disease using ground-based hyperspectral image analysis. Florida Ag Expo, Balm, Florida, November 9th, 2010

Progress 10/01/11 to 09/30/12

Outputs
OUTPUTS: During this period, experiments on modeling vegetation Leaf Area Index (LAI) in urban/rural areas using moderate resolution hyperspectral imagery (Hyperion)were conducted. More than 53 forest sampling plots in the Tampa bay area were used in the study. remote sensing imagery involved 30m resolution Hyperion image. Eight spectral indices including Simple Ratio (SR), Normalized Difference vegetation Index (NDVI), and Soil Adjusted Vegetation Index (SAVI) were tested to find best index-band combinations. PARTICIPANTS: Naveen Anne Michael Andreu Nicole Hewitt TARGET AUDIENCES: Nothing significant to report during this reporting period. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.

Impacts
Our results supported the findings of previous studies that demonstrated the possibility to model LAI using spectral indices extracted from remote sensing imagery. The SR, SAVI*SR, and NDVI*SR indices produced best results with different band combinations. Unlike other researcher results conducted in dense forest landscape, our study showed moderate correlation between spectral indices and vegetation LAI. Most previous studies of LAI in diverse landscape utilized high resolution hyperspectral imagery and achieved results inferior to this conducted over dense vegetation. In our case, we used 30m resolution imagery, which in addition to landscape heterogeneity could have affected the results. We are still evaluating the use of local spatial autocorrelation coefficients in the LAI prediction models to account for spatial homogeneity in urban/rural areas.

Publications

• Abstracts: Anne, N. and Abd-Elrahman, A, M. Andreu, D. Lewis, and N. Hewitt, 2012. Estimating forest biophysical parameters in spatially heterogeneous urban-rural landscape gradient using Hyperion and Landsat images. American Society for Photogrammetry and Remote Sensing annual conference, Sacramento, California, March 21-24.

Progress 10/01/10 to 09/30/11

Outputs
OUTPUTS: An experiment was conducted to test the performance of ground-based hyperspectral imaging sensor in evaluating plant (tomato) strees due to bacterial infection. Plots were established in the fall of 2010. The trials were established along 300 ft long beds with 5 ft center-to-center bed spacing prepared. Tomato seedlings were transplanted at 18 inch spacing to 3 separate 24 foot bed lengths for each treatment plot with 6 feet between plots. Treatment trials were performed which include vascular wilt treatments, blight treatments and viral treatments. Plants were periodically monitored for the incidence and severity of diseases over time using visual assessment. A total of 15 visual assessments (performed in three dates 2-weeks apart) of 5 plant beds were conducted and synchronized with imaging spectroscopy data collection. Imaging Spectroscopy: Hyperspectral images were collected using Specim/Imperx hyperspectral sensor system developed in our Geomatics Lab mounting on a camera boom carried on board a vehicle. Digital images are also captured using Nikon D300 camera mounted on the same boom. A 4 x 5 ft PVC frame is made which is used to hold the spectralons and calibration plates and to mark field boundary for each imaging site by fitting it onto the buried PVC pipes. 5 such defined sites were imaged with the mentioned sensors 3 times (bi-weekly) starting at 3 weeks after transplanting. Hyperspectral images were subjected to preprocessing, which comprises reducing stripes and noise using method of filtering selected MNF bands with wavelet-frequency adaptive filter, subtracting dark current and, computing reflectance value using reference (calibration) observations. Then the reflectance data are analyzed using NDVI Vegetation indices (Delalieux et al. 2009; haboudane et al. 2002) , as expressed below: NDVI = (NIR - RED) / (NIR + RED) = [R(739)-R(671)] / [R(739) + R(671)] Areas/volumes of various classes of Vegetation indices, as defined by NDVI indices, are calculated and their percentages with respect to whole frame area are computed. Tomato disease severity is then correlated with the percentage vegetation indices. Results: Regression between average NDVI within each marked bed versus Disease Severity (DS) shows some moderate level of correlation (r = 0.64). A second order polynomial gave a better fit. The results were in contrary to expectations as we expects reduced NDVI with increased disease level. We believe that the average NDVI (plant health measure) is dominated with the effect of plant growth and increased canopy area with time, especially that tomato diseases are often spread from the bottom up. In order to accurately correlate reflectance of surface canopy with disease status, the data needs to be normalized against increased canopy size due to plant growth. Efforts are underway to improve the performance of the ground-based hyperspectral sensor. A Novatel SPANCPT GPS/INS system was purchased to improve the geometric correction and georeferencing performance of the system. Equipment integration and testing is being implemented. PARTICIPANTS: In addition to PI, Dr. Gary Vallad contributed to this project by providing tomato trial sites and visual assessment of the disease. Undergraduate (Gregory Baksis) and graduate (Roshan Pande-chhetri) students participated in image analysis and reporting. TARGET AUDIENCES: Researchers involved in remote sensing and image analysis applications who study plant phonology and pathology and eventually growers and farmers looking for disease detection and best management practices for pest control activities. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.

Impacts
The system proves its feasibility to detect plant disease through image analysis. Methods developed in this regard can be implemented from ground-based imaging systems and/or an airborne platform.

Publications

• Abd-Elrahman A., R. Pande-Chhetri, G. Vallad, 2011. Design and Development of a Multi-Purpose Low-Cost Hyperspectral Imaging System. Remote Sensing 3(3):570-586.
• Panda-Chhetri, R., A. Abd-Elrahman, 2011. De-striping hyperspectral images using wavelet transform and adaptive frequency domain filtering. International Society of Photogrammetry and Remote Sensing Journal 66(5):620-636.
• Abd-Elrahman, A., M. Croxton, R. Pande-Chhetri, G. Toor, S. Smith, and J. Hill, 2011. Chlorophyll-a estimation using in-situ hyperspectral imaging system in aquaculture water bodies. ISPRS Journal of Photogrammetry and Remote Sensing 66 (4) 463-472.

Progress 10/01/09 to 09/30/10

Outputs
OUTPUTS: We experimented with multiple ground-based sensors including consumer grade digital camera and hyperspectral sensor. We tested the sensors on three different applications: (1) extracting metric and qualitative information for urban forest inventory; (2) Modeling chlorophyll-a (chl-a) concentration in aquaculture ponds; and (3) Assessing bacterial disease effect on tomato vegetation. The experiments showed that low-cost consumer grade cameras were able to provide information needed for urban forest inventory and model chlorophyll-a content using 2-band/3-band models and special calibration target design (R^2 = 0.98). The results were disseminated through presentations and posters in international and national professional and scientific events. PARTICIPANTS: Marry Thornhill is a graduate student who contributed to this work as part of her thesis. Michael Andreu and Francisco Escobedo are UF faculty member and urban forest specialists, who advised on field work acquisition and techniques used by urban forest crews. Roshan Pandi-chhetri (a graduate student working on spectral data analysis) and Matt Croxton (research assistant) contributed to the spectral analysis of the imagery to build chl-a models. Scot Smith, Gurpal Toor, and Jeff Hill advised on different aspects regarding water quality analysis and field site selection. TARGET AUDIENCES: Remote Sensing and image processing community, Urban Forest specialists, Water quality Specialists PROJECT MODIFICATIONS: Not relevant to this project.

Impacts
Our developed techniques for urban forest inventory can be used to supplement or replace urban forest crew observation. They also can be used as effective techniques for quality assessment and control. Our spectral analysis of chl-a content in aquaculture ponds provided a technique that enables accurate estimation of chl-a content using close-range spectroscopy in eutrophic and/or hyper-eutrophic waters.

Publications

• Journal Publications: Abd-Elrahman, A.H., M. Thornhill*, M. Andreu, F. Escobedo, 2010. A community-based urban forest inventory using online mapping services and consumer-grade digital images. International. Journal of Applied Earth Observation and Geoinformation, 12:249-260.
• Abstracts: Croxton, M., A. Abd-Elrahman, R. Pand-chhetri*, G. Toor, J. Hill, 2010. Estimation of Water Quality Parameters in Freshwater Aquaculture Ponds Using Hyperspectral Imaging System. ASPRS Specialty Conference Topics: Ocean Monitoring and Modeling.
• Abd-Elrahman, A., R. Pand-chhetri*, 2010. High-resolution mobile hyperspectral imaging. ISPRS Commission IV Topics IV/3 Mapping from High Resolution Data.

Progress 10/01/08 to 09/30/09

Outputs
OUTPUTS: Ground-based Vegetation Mobile Mapping System (VMMS) components were identified according to design specifications. The components include hyperspectral sensor, consumer-grade digital camera and navigation sensors. The design for the integration scheme of the sensor components was developed including the mobile mapping platform and the synchronization scheme. All components were acquired. Individual sensor testing was conducted and integrated system testing is underway. Image enhancement techniques were developed to overcome image typical image defect such as sensor gain/offset striping. Experimental hyperspectral data sets were acquired for vegetations (flower beds, right-of-way vegetation). A symposium was organized to demonstrate the advances in geo-spatial technologies including remote sensing technologies on precision agriculture. The VMMS system configuration, potential and initial results were presented. Overall, the project fulfilled its objectives for the first implementation year. Work is underway to accomplish second year goals. PARTICIPANTS: Amr Abd-Elrahman (PI): responsible for System design and integration and supervise graduate students. Roshan Pande-Chetri (graduate student): processing of captured images and developing image enhancement algorithms. Mary Thornhill (graduate student): Data collection and analysis Matthew Croxton (research assistant): Data collection and analysis Siddharth Goyal (gradate student): application development Sandeep Chinni (graduate student): application development TARGET AUDIENCES: Not relevant to this project. PROJECT MODIFICATIONS: Nothing significant to report during this reporting period.

Impacts
A mobile mapping platform is designed and implemented. The platform can be mounted on a moving vehicle and used to collect imagery suitable for vegetation analysis. A system synchronization technique was developed to synchronize captured images with GPS time. A new technique for image de-striping was developed to enhance captured hyperspectral images.

Publications

• Abd-Elrahman, A., M. Thornhill, M. Andreu, and F. Escobedo, 2009. A community-based urban forest inventory using online mapping services and consumer-grade digital images. International Journal of Applied Earth Observation and Geoinformation (pending).
• Abd-Elrahman, A., 2010. Development of a ground-based hyperspectral mapping platform. International Journal of Remote Sensing (pending).
• Abd-Elrahman, A., R. Panda-Chetri*, and S. Smith, 2010. Utilizing wavelet analysis and adaptive filtering to minimize striping effect in hyperspectral imagery, Photogrammetric Engineering and Remote Sensing (pending).