IDENTIFICATION AND CLASSIFICATION OF PADDYAILMENTS BY MODIFIED INCEPTION ARCHITECTUREUSING NETWORK LEARNING TECHNIQUE

Agriculture productivity is something on which the economy highly depends. India is an agronomical country where about two-third of population depends on agriculture. Among which, paddy is India’s preeminent crop and the staple regime of east and south India. India is the second-largest producer of rice. It also extorts vastly to the GDP of the country. The important fact is out of 98% of paddy crop grown in India, 76% of them are affected by diseases. Paddy has become more significant to feed the ever-growing populations. Early recognition of diseases, pests, and treatment is crucial. Proper care should be taken in this area to avoid serious effects on the crops and due to which productivity and quality are very much affected. This study aims to capture the paddy ailment images in which the output can be trained and tested by modified inception architecture. As a result, the diseased images are classified and identified based on their convolution filtering method. The use of a modified inception model will increase the performance of the model prediction and its robustness. Here the different combination of blocks will be trained in parallel and finally, concatenation of the blocks is done to improvise the efficiency of the size of the model. This proposed model mainly focuses on paddy leaf diseases like Bacterial leaf blight, Blast, Brown spot, False smut, Grain discoloration, Leaf streak, Sheath blight, Sheath rot, Tungro. This proposed model presents an architecture for image processing technique that is utilized for the identification and classification of paddy ailments using Network learning techniques.