Project : USDA ARS

ARS Home » Southeast Area » Stoneville, Mississippi » Cotton Ginning Research » Research » Research Project #438162

Research Project: Development and Evaluation of Novel Technologies to Improve Fiber Quality and Increase Profitability in Cotton Processing

Location: Cotton Ginning Research

2025 Annual Report

Objectives
1. Develop methods and devices to enable the reduction of plastic contaminants in commercially harvested cotton. 1.1. Develop a UAV-based intelligent system to identify and remove plastic particles in cotton field. 1.2. Develop a sensor and control system to remove plastic contamination in ginning process. 2. Develop and evaluate tools and methods to enable the commercial preservation of cotton fiber quality and increase ginning efficiency. 2.1. Develop and evaluate sensors for cotton moisture measurement in real time in situ. 2.2. Detect moisture in cotton module using UAV-based platform. 2.3. Develop and evaluate air-bar lint cleaner to increase the turnout and preserve fiber quality. 2.4. Develop a sensing and control system to automatically adjust ginning process for optimal ginning efficiency. 3. Develop methods to enable the use of commercial cotton gin trash and seeds for bio-products and bio-energy. 3.1. Develop new methods to process gin trash for bio-products and energy. 3.2. Investigate moisture dynamics in cotton seeds.

Approach
The Cotton Ginning Research Unit seeks to develop cotton ginning technologies to maximize fiber quality, increase ginning efficiency, and minimize the environmental impact of ginning. Plastic contaminants in U.S. cotton are rapidly increasing in recent years and have become a serious threat to U.S. cotton industry by reducing marketable quality. New sensing and control systems and ginning machinery are needed to clean the contaminants, improve fiber quality and ginning efficiency, and increase cotton producers’ profitability. Researchers will develop and evaluate sensing and control systems to remove plastic contaminants from cotton and develop new tools for accurate cotton moisture measurements. UAV (unmanned aerial vehicle) remote sensing will be used as a platform to find and remove the plastics from cotton fields and to detect moisture in cotton modules. Optical sensors, data processing, automatic controls and the like will be designed and built to detect and remove the plastic materials during gin processing. Moisture sensors, coupled with improved measurement of mass-flow rate and new models, will be developed and tested to accurately determine moisture of seed cotton, cotton lint, and cotton seeds in real time. Using the data gathered, an improved control system will be designed and fabricated to optimize ginning efficiency. Additional research includes developing and evaluating new lint cleaning technology to better preserve fiber quality and increase the ginning turnout. Studies on new methods to use gin trash for bio-energy will also be conducted in this project.

Progress Report
As part of Objective 1, researchers developed systems to detect contamination in cotton fields and collaborated on systems to remove plastics in the gin. One way plastic gets into bales of cotton is trash in the field picked up during harvest. ARS researchers in Stoneville, Mississippi, addressed this by developing an Artificial Intelligence (AI) system for detecting plastic and other trash in the field. This required collecting and annotating field data from 2021 to 2023 using an Unmanned Arial Vehicle (UAV). The annotations involved boxing and naming objects in thousands of images. Parallel efforts also studied the interaction between lenses, flight heights, and object sizes. That interaction is important because large objects are easier to detect but low flight times decrease expenses. The final dataset had over 24,000 objects representing a range of conditions. Researchers used data from 2021 and 2023 to train the system and 2022 data to test it. Holding out a year of data for testing provides a more accurate performance estimate for future crop years. Testing showed detection precision above 80% is possible which means there are few false positives. A low false positive rate helps limit product loss when acting on detections which is beneficial for precision agriculture. A collaboration with university partners in Starkville, Mississippi, has focused on a robot for removing detected plastic. A related study with university partners in Tifton, Georgia used a camera to identify sources of bark in ginned lint. By using location, production, and harvest data, the study integrated imagery data with the “harvest identification” HID file. The results from this study help to develop best practices for improving cotton fiber quality. For plastic detection and removal in the gin, ARS researchers worked with university partners in Starkville, Mississippi. A proof-of-concept system for detecting plastic in seed cotton was designed to operate just after the module feeder dispersing cylinders. At this location, the module feeder has dispersed seed cotton into smaller blocks of material, and the cotton is moving relatively slowly. The system works by imaging the seed cotton as it tumbles. Tumbling the cotton stirs the material, reveals more surface area, and increases the chance of exposing hidden plastic. The imaging system used a custom AI model trained to detect plastic. Testing on the system showed the AI model could run at up to 60 frames per second. Laboratory scale testing on the overall system performance showed that the AI system could obtain up to 93 percent accuracy. As part of Objective 2, researchers developed energy and moisture sensing prototypes, took steps to count seeds in real-time and developed systems to estimate cotton fiber quality in the gin. The prototype low-cost energy sensing tool works by collecting gin stand power and a new software tool extracts the ginning energy from the power data. These new tools enabled research on variables affecting energy such as sample size, saw sets, feeding conditions and ginning rate. These factors are important for reducing energy costs associated with ginning. The energy work also enables the development of lower ginning energy cotton. Therefore, researchers also studied real-time cotton seed counting in parallel with power measurements to quickly calculate metrics used by breeders like seed index. These studies established the ability to track individual seeds using AI based object tracking and low-cost components. In addition, researchers worked with university partners in Clemson, South Carolina, to study the effects of drought tolerance traits on cotton ginning, cottonseed, and fiber quality. ARS researchers worked on moisture testing with university partners from Starkville, Mississippi. This led to the development of a prototype capacitive sensor and paddle sampler for cotton. The sampler was set up in a seed cotton duct for testing and calibration. The tests also included an off-the-shelf capacitive sensor. The sampler used a mass-flow sensor that included a light source and light sensor that detected cotton flowing. A controller and a laptop computer running a custom program controlled the system. The range of moisture tested was 4 to 14%. Both sensors provided repeatable data with the custom capacitive sensor having a lower price point. Researchers looked at sensing moisture in round cotton modules using both thermal images and temperature sensors. The thermal imaging portion involved capturing images both at ground level and at an elevated viewpoint like a low flying UAV. Solar heating of the plastic wrap affected the thermal images. There were some correlations between moisture probe values and the infrared maximum temperature, but the relationships changed based on time of day. This makes it hard to compare data across time. In parallel to this experiment, a collaboration with ARS researchers in Lubbock, Texas included logging the temperature of 50 cotton modules. The sensors were placed into modules of known moisture levels shortly after harvest and removed just prior to ginning. Modules with a moisture content of approximately 10% or greater tended to rise in temperature for at least the first several days after construction while those under approximately 7% tended to cool off. This relationship may be useful for detecting higher moisture content modules so that they can be handled with best practices for preserving fiber quality and minimizing losses. To increase turnout and preserve fiber quality, ARS researchers work on a grid bar with a new shape. The bar was intended to allow for easier removal of foreign matter attached to the fiber. However, the grid bar captured more material and suffered from buildup of cotton. The design proved to be unfeasible to deploy without developing some method to remove the buildup. Compressed air was attempted to remove the buildup but that only showed limited success. ARS researchers in Stoneville, Mississippi, developed two experimental systems for estimating fiber quality in terms of leaf grade and micronaire in a cotton gin. The systems can provide information in real-time. This information could be used to adjust the ginning process without the time lag associated with the USDA classification of samples. Leaf grade estimates came from images of cotton bales just before bagging. Micronaire was based on data collected from the press. The systems collected data at two commercial gins over multiple ginning seasons. The data is being used to refine the process. As part of Objective 3, ARS researchers worked on adding value to byproducts and issues affecting seeds. Researchers focused on finding uses for cotton gin byproducts (CGB) that improve the profitability of cotton ginning. Previous efforts on CGB usage found many methods have economic challenges. Researchers in Stoneville, Mississippi, started by looking at CGB as a feedstock for soil amendment pellets. Tests on pellets made from CGB compost showed a durability over 90%. Early CGB pellet tests suggested that composting affects quality. Further studies looked at the effect of weather and composting on CGB properties. Weather effects were researched in multi-year studies at both Stoneville, Mississippi, and at Las Cruses, New Mexico. Efforts also included finding low-cost additives that can help produce CGB-based fertilizer pellets. Beef manure was the main additive considered. Researchers also initiated projects on co-composting CGB with dairy manure in collaboration with university partners in College Station, Texas. A collaboration with university partners in Tyler, Texas focused on converting CGB into liquid fuels and high-value chemicals. To investigate mechanical damage on cotton seeds, researchers collected cotton seeds from multiple scales of ginning equipment. In collaboration with ARS researchers in Lubbock, Texas, the commercial scale seed samples were delinted. An x-ray system tested the delinted seeds for damage while the seed moisture was tested using a standard gravimetric oven method. As part of a collaboration, samples were also collected from a small-scale research gin and delinted using a chemical method. Visual damage analysis of those seeds was done using a microscope. This work generated data that is critical for establishing moisture dynamics in relation to seed damage. To develop an industrial hemp economy in the United States, ARS researchers in Stoneville, Mississippi worked with many industry and university partners at the request of the Agricultural Marketing Service. The industrial hemp industry needs to improve the processing of hemp stalks into fiber and hurd. Addressing these needs requires knowledge from the ginning industry. Similarly, the industry has a need for the ability to characterize industrial hemp fiber and hurd to develop a domestic industrial hemp economy. Researchers applied their knowledge of processing natural fibers and pneumatic transportation of fibers to develop a novel continuous-flow hammermill decortication process for separating industrial hemp fiber and hurd. Researchers have applied the techniques and knowledge of cotton fiber characterization to begin to develop methods to characterize industrial hemp fiber. The fiber is highly variable in length, strength, fineness and non-fiber content and represents a significant challenge for routine measurements. Methods to assess these properties of industrial hemp fiber have been developed by modifying commercially available tools. The developed methods have limitations and are, thus far, suitable for use on highly processed fiber samples, but not for the complete range of fibers being marketed. However, these initial steps are crucial to supporting the emerging domestic industrial hemp industry.

Accomplishments
1. Multi-platform cotton ginners calculator. Maintaining and modifying a cotton gin requires diverse knowledge contained in handbooks and research papers. Some of the documents may not be readily available when needed. To address this a simple tool was developed. ARS researchers at Stoneville, Mississippi developed a multi-platform web-based tool that makes key calculations easier to complete. The cotton ginners calculator provides easy access to calculations commonly used in cotton gins. These include calculations related to airflow, mechanical systems, and dryers. The calculator consists of multiple modules which allow stakeholders to quickly make calculations without needing to consult complex equations, charts, or lookup tables. The multi-platform modular aspect of the calculator ensures that it is available to the widest possible group, and it is easy to upgrade.

2. AI based field contamination detection. Plastic continues to be a significant problem for the USA cotton industry, costing over $300 million per year. This work addresses plastic found in seed cotton before it arrives at the gin. Multiple vision systems were trained to detect contamination in the cotton field and a dataset of over 24,000 annotated objects was curated. Testing on data held out from the crop years used for training showed the vision system had a low false positive rate. The system also detected most of the contamination. Testing on data from a crop year not included in training showed that the model works effectively when operating on data it had not previously seen. The dataset itself also provides great value to the industry by enabling future state-of-the-art developments in trash detection.

3. Composting and co-treatment of byproducts of cotton ginning and cattle industries to make value-added product. Millions of tons of byproducts are generated annually by cotton gins and beef cattle operations. These byproducts are high in nutrient content and could become a profitable co-product. Composting is a common practice in both industries, but research on the process for cotton gin byproducts (CGB) is limited. The application of these byproducts as soil amendment is also significantly limited by the low density, and concerns with the presence of weed seeds in the byproducts. ARS researchers at Stoneville, Mississippi conducted a multi-state, multi-year, data intensive CGB and cattle manure composting study. These studies were the first documented investigation of co-pelletizing cotton gin byproducts with beef manure. The research on composting CGB is the first to investigate and continuously monitor industrial-scale composting processes for CGB. The efforts have illuminated the potential for co-treating and co-utilizing byproducts of two big industries in the US.

4. 2024 Roadmap for economically feasible markets for cotton gin byproducts (CGB). For decades, researchers and experts have explored the potential of various methods and technologies to add value to CGB. In 2023, the industry stakeholders recognized a critical need for robust documentation of all past and current research in this area. To address this, ARS researchers at Stoneville, Mississippi collaborated with other ARS researchers across multiple locations and an international expert to compile decades-long record of relevant research and industry practices. This work provides guidance to the industry on the future of sustainable CGB utilization. The project was recognized by key stakeholders, including the National Cotton Council, who recommended that it be published as chapter in the Cotton Ginners Handbook. The handbook is used throughout not only the US cotton industry but serves as a guide for the global cotton ginning industry.

U.S. DEPARTMENT OF AGRICULTURE

Cotton Ginning Research: Stoneville, MS