A Novel Approach for Identification of Weeds in Paddy By using Deep Learning Techniques

Weed is an unwanted plant which grown in agriculture land. The land not cultivated will be fully covered by Weeds. Management of weed the major concern for farmer because reduce crop production quantity. There are many methods to control weeds, one those manual plucking expensive it takes more time, consumes human work. Second applying any chemicals suggested external experts. This may cause damage cultivated. Identifying weeds early stage growth and destroying them through proper method most important increasing production. We proposed efficient identifying classifying paddy field using Deep learning-based computer vision techniques. applied Semantic Segmentation model trained our with SegNet different batch size 16,32,64 obtained a highest accuracy 94.223 dropout value 0.1 set 32.