|Hughs, Sidney - Hughs Ed|
Submitted to: National Cotton Council Beltwide Cotton Conference
Publication Type: Proceedings
Publication Acceptance Date: 1/5/2000
Publication Date: 6/30/2000
Citation: N/A Interpretive Summary: Research at the Southwestern Cotton Ginning Research Laboratory (SWCGRL) has resulted in improving our ability to identify trash categories of bark (fibrous), bark (non- fibrous), leaf, and pepper in cotton. Eighteen Agricultural Marketing Service (AMS) trash level check boxes were used. The correlation coefficient between AMS and SWCGRL trash content measurements is 0.9988 indicating the technique proposed can be implemented to predict the trash content. In addition to the total trash area, the amount of each trash type can be computed.
Technical Abstract: Soft computing techniques such as neural networks and Fuzzy Logic based approaches are discussed in relation to their effectiveness in the identification of various trash types (non-lint material/foreign matter) in ginned cotton. The performance of a hybrid neuro fuzzy structure, namely the Adaptive Network-Based Fuzzy Inference System (ANFIS) to classify trash types is compared to other techniques. Shape descriptors like area, solidity and extent are used to distinguish between the trash types. The correlation between Agricultural Marketing Service (AMS) and Southwestern Cotton Ginning Research Laboratory (SWCGRL) trash content is computed.