Plant Disease Identification Using Deep Learning Classification Model: CNN


Sakshi Mangal, Pratiksha Meshram
Department of Information Technology, Narsee Monjee Institute of Management Studies, Shirpur, Maharashtra, India.


Agriculture in India is not all about getting food for livelihood from agriculture. Large portion of Indianeconomy comes from agriculture as India is one of the largest exporters of food, grains and other agricultural products. More than 70% of rural population of India is dependent on agriculture of their livelihood. 42% of crop gets damaged every year which leads to huge loss to Indian farmers, out of which 15.7% of crop is damaged due to pest. Thus, early detection of plant disease is very important to protect complete plant from getting damaged. The older technique in which diseases are detected when changes are noticeable on plant leaf is not efficient as till that time almost whole plant gets damaged thus new technique like image processing and computer vision algorithms should be used to detect disease at early stage. The technique which is implemented should provide accurate and precise detection of disease, so that the pesticide and insecticide which are being used does not damage quality of soil and also spreading of large amount of pesticides and insecticide damages directly or indirectly the health of crop. Damage in crop quality or quantity leads to losses thus early and accurate detection of plant disease are very important. Thus, for early disease detection we used Laplacian filter and Unsharp masking technique as image processing technique and Canny edge detection as image segmentation technique. The classification model which we are implementing in this project is convolution neural network which is a deep learning classification model. [1][6][7][8][10][12][14]