Revolutionizing Agriculture by Automated Plant Disease Detection and Management Using Advanced Imaging Techniques and Xception Neural Network

The number of agricultural items produced, productivity, economic prosperity, and crop yield are all significantly reduced as a result of plant diseases. Considering that agricultural products account for 70% of India's GDP, it is imperative to control and reduce plant disease losses. Controlling plants from the very beginning of their life cycle is essential to reducing the incidence of these diseases. The only method of supervision used in the past was eye observation, which necessitates a large time, resource, and specialist knowledge commitment. Therefore, it is essential to automate this process by employing a disease detection methodology. It is essential to design an imaging-based illness detection system. Plant diseases are identified by the use of image processing. The process of detecting diseases include taking photographs, pre-processing them, segmenting them, collecting pertinent information from them, and then categorising them. This study looked at the identification of plant illnesses using photos of leaves and the treatment of identified diseases with pesticide recommendations. Specific methods for dividing and extracting plant pathogens during detection were also included in the debate. A comparison study found that Xception outperformed other pretrained deep learning models with an astonishing accuracy of 97.82%. In terms of accuracy, Xception performs better than other well-known models as ResNet50 (97.45%), MobileNetV2 (95.84%), InceptionV3 (96.57%), and VGG16 (94.35%). This paper emphasises how cutting-edge neural network architectures may improve disease detection precision, support more sustainable crop management practices, and boost agricultural output.