Progress 07/01/22 to 02/28/24

Outputs
Target Audience:The target audience of this project is the natural fiber industry. The natural fiber industry lacks the tools needed to develop elite spinnable germplasm and to develop efficient and automated supply chains, which include the processing of the crop into fiber. The target audiences can be split into two groups, the more established cotton fiber industry and the growing bast fiber industry. The bast fiber industry is largely led by hemp fiber production. Within the cotton fiber industry, our target audience is cotton breeders, agronomists, and cotton merchants. The target audience within the bast fiber industry also includes breeders and merchants but is also extended to include stakeholders involved in decortication processes. The target audience also includes spinning mill managers and process engineers that are wanting to use natural fibers as a raw material in their yarn. Our efforts were to provide these target audiences with an MVP that addresses their need for a fiber quality tester. We also engaged these groups through customer discovery interviews. Changes/Problems:We experienced many challenges during this project. We were able to overcome most of these challenges and were able to develop an MVP that meets the target goals. What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest?We are currently disseminating the results by word of mouth, including presentations at conferences. We will develop a more robust marketing strategy after we vet the MVP this fall in the fiber quality laboratory. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Goal 1: We developed an MVP that will subsample a sample of natural fibers and present the subsample to a camera. The camera images the subsample, and the software evaluates the quality of the subsample. The quality metrics are then reported to the user. The device will be evaluated in a natural fiber testing laboratory after harvest in the fall. The device can subsample and image the fiber stages of hemp processing. The MVP can also subsample and image all the cotton samples we evaluated. Goal 2: The MVP can subsample and image hemp obtained from hemp fiber processing and provide quality metrics. Some stages of hemp processing are not fibrous and are not appropriate for this MVP. The MVP can evaluate some important hemp fiber quality characteristics based on traditional image analysis techniques; including length, length uniformity, and color. The machine learning approach can also detect differences between hemp samples. Quality metrics are made available to the user after analysis. Goal 3: The MVP can subsample and image cotton fiber representing different genetic and environmental backgrounds and provide quality metrics that differentiate them. The primary tool used by the cotton industry for quality evaluation is the High Volume Instrument (HVI). The HVI provides a measurement of length, length uniformity, strength, micronaire, color, and reflectance of a sample of cotton. The MVP can provide comparable measurements of fiber quality. Goal 4: We spoke with hemp growers, researchers, and hemp fiber company owners. We also spoke with the ASTM hemp fiber working group, a group developing standards for hemp fiber quality measurement. We also spoke with cotton growers, breeders, agronomists, and merchants.

Publications

Progress 07/01/22 to 06/30/23

Outputs
Target Audience:We have had minimal interactions with external audiences up to this point. We are in the process of prototyping our MVP and have not had an external prototype trial. We have continued with customer discovery, but not to the degree that would occur in an I-Corps cohort. The potential customers we interacted with were cotton breeders, physiologists, and agronomists. Changes/Problems:We do not have significant changes. We will not need to deviate significantly from the MVP described in the original proposal. We will likely end up needing to use more expensive computing hardware than we expected. This is largely due to the poor supply of inexpensive alternatives. This should not negatively impact our ability to deliver on this project but will affect our manufacturing costs. What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest?The natural fiber industry is more aware of the work we are doing due to our participation in grower meetings, but we have not published any of our results. Publishing is not a goal of this project. What do you plan to do during the next reporting period to accomplish the goals?Goal 1: We are finishing up the prototype. Part of our goal is to have something that works on many different fibers and is robust enough to be used in a commercial setting. We are not there yet. We will continue to prototype hardware with additive manufacturing until we are able to refine the settings to meet our goals. Goal 2: We are currently evaluating different approaches to accomplish this task. If an embedded approach does not work, we will rely on attaching the system to a more expensive laptop. This will drive up manufacturing costs, but will allow us to have a completed MVP. Goal 3: We have the material we need for this goal. Our approach will be similar to Goal 2 above. Goal 4: Continue to talk with potential customers. We need to reach out to more potential customers in the hemp and other natural fiber industry

Impacts
What was accomplished under these goals? Goal 1: We made significant progress in this arena. Early progress was hampered by equipment problems. One of the machines we obtained for manufacturing some of the parts did not function as expected. We developed a different manufacturing approach while trying to overcome this problem. Our early prototypes were almost exclusively printed using additive manufacturing, such as filament printing. Our current prototype is built around aluminum extrusions. There are some custom parts that are still being refined that rely on in-house manufacturing. This allows for quick prototyping. The primary parts requiring this prototyping approach are the sampling bucket, most of the sampling jaw, and parts of the cleaning mechanism. The switch to more universal parts has resulted in a more robust frame and a shorter manufacturing and prototyping cycle. We are still refining the sampling jaw, but we have made significant progress and do not expect this to hold back progress. Goal 2: This has been a major stumbling block for our project. We were expecting to use a raspberry pi based system for an on-board software interface. The global supply chain limits the availability of these boards and we do not believe we can depend on these boards in manufacturing. We are currently exploring alternatives. Alternatives include generic forms of the boards and/or some other type of embedded system. We are concerned about the reliability of generic boards and are primarily focused on other embedded systems. The main drawback of other systems is the price. While there were many setbacks for this objective with regard to the electronics and computing, we believe advances in deep learning networks create even more opportunities than we originally envisioned. Goal 3: This objective follows from objective 2. We are progressing, but this objective is not complete. We arranged for cotton samples from 2023 production year for use in training and validation of the system. Goal 4: We have continued to conduct customer discovery. We talked with cotton breeders, physiologists, and agronomists. There is a general consensus that a new approach is needed for fiber quality assessment. Spinning natural fibers for evaluation is too expensive and requires too much material; current fiber quality assessment methods are not adequately predicting spun yarn quality. We feel like we have the potential to fill this void. We discussed potentially partnering with a breeder to validate the utility of our prototype in germplasm development. We need to work on customer discovery in other natural fiber industries, such as hemp.

Publications