Early Detection and Prevention of Mustard Downy Mildew Disease using a Hybrid CNN-LSTM Model
Vinay Kukreja, Rishabh Sharma, Satvik Vats
Abstract
Mustard downy mildew is a widespread disease that poses a significant risk to the production of high-quality mustard products, which in turn affects the agricultural industry. The early discovery and prevention of the illness are necessary if significant crop losses are to be avoided. The proposed research suggests a hybrid convolutional neural network (CNN) and long short-term memory (LSTM) model identify three thousand pictures of mustard crops based on the six different severity levels of the disease known as mustard downy mildew. When it came to determining the extent of the illness, the model that was suggested was extremely accurate, achieving a rate of 92.36%. The model's incorporation of CNN and LSTM layers enabled it to effectively manage the complicated visual patterns caused by the illness. This was made possible by the successful recording of both the spatial and temporal characteristics of the images. The suggested model has the potential to serve as an efficient instrument for the early detection and prevention of mustard downy mildew disease, which has the potential to make a substantial contribution to the development of environmentally responsible agricultural practices. According to the results of this research, a hybrid CNN-LSTM model may have the ability to increase agricultural outputs as well as the quality of mustard oil production.