A convolutional neural network with average pooling for chickpea disease detection and classification

Chickpeas rank among the top legumes worldwide, yet diseases can hit them hard, slashing both quality and yield. Accurate disease classification matters a lot for managing them well. In the Proposed research work use a Convolutional Neural Network (CNN) with Average ‎Pooling to identify and categorize diseases of chickpea plants. Primary Data used in this work are Chickpea and Fusarium-22 datasets Pre‎Pre-processing was done through demonising to maintain the purity of the data. To balance the classes, Data aug-mentation is done, which boosts the dataset ‎size, with key features drawn out using the AlexNet architecture, before the CNN-Average Pooling steps in to classify the diseases. The ‎CNN automatically snatches crucial details from chickpea images, while the Average Pooling layer slices down spatial dimensions to interesting patterns that are more useful. This arrangement brings precise disease classification, making it beneficial for agriculture and crop management. Tests show the CNN-Average Pooling hitting 99.60% accuracy on the Chickpea dataset and 99.53% on Fusarium-22, exceeding ‎options like CNN-LSTM, the GLCM-Color Histogram combo, and DenseNet-201‎