Automated Cotton Leaf Identification Using Feature Selection Techniques

Agriculture is the backbone of any prosperous nation. Pest infestations and bacterial or viral illnesses cause significant economic losses in the cotton farming commercial, costing Indian farmers an average of 10-20% of their annual income. Cash crops include cotton and other valuable agricultural products. Cotton is highly susceptible to the vast majority of crop-damaging diseases. Several diseases affect crop production by attacking the leaves. The early diagnosis of diseases helps prevent additional damage to crops. Many diseases can afflict cotton, including leaf spot, nutrient insufficiency, powdery mildew, leaf curl, and many others. Correctly diagnosing a condition is crucial for taking appropriate action. Accurately diagnosing plant diseases requires. The suggested model based on the Biattention process makes accurate diagnosis of cotton leaf diseases possible. Also, useless features lower categorization precision. These issues are tackled by the IGWO (Improved Grey Wolf Optimization) method. We photographed cotton leaves in the field for our analysis. There are 2385 images in the dataset, including both leaves. The dataset was expanded with the use of data increase techniques. A meta-learning strategy has been devised and applied to deliver high precision and generalization. The projected model has a higher accuracy on the Cotton Dataset, at 97.45%.