Advancing Plant Disease Detection: Exploring the Efficiency of Feed-Forward and Backpropagation Neural Network

As of September 2021, approximately 58% of agriculture employed a large portion of India's people, either directly or indirectly. It contributes more to our country's GDP and also employs the majority of India's population. The major difficulty in agriculture is the accurate detection and diagnosis of plant disease and plant stress. Plant disease identification is vital for preventing yield and quantity losses in agricultural products. The plant disease diagnosis is critical for farmers to monitor manually as it requires time and knowledge of plant pathogens. Plant diseases are caused by viruses, bacteria, fungi, etc. This is detected using image processing by capturing the image of the leaves and processing it in an image processing platform. Based on image processing techniques, an algorithm for feature extraction and classification was developed. The goal is to detect, identify, and quantify the symptoms of disease, provide solutions, and detect the rate of spreading of the disease by which the farmers can preplan their harvesting to prevent losses. In existing farming technology, the articles are released by which the farmers cannot identify all the diseases and its solution. This plant disease detection system based on leafimage processing techniques is a highly beneficial and low-cost technology, especially for assisting farmers in monitoring vast plantation regions. This comes under the scheme of Sub Mission on Plant Protection and Plant Quarantine (SMPP).