Abstract:

Abstract: For most people, farming is both a livelihood and a way of life. In most parts of the world, farming is central to cultural practices and customs. More efficient use of time and resources, as well as increased profitability, could be possible in the agricultural sector with the application of modern farming techniques. An innovative framework for plant leaf disease classification is provided by the proposed CGANmodel, which combines OTSU and SURF. One method for improving and preparing images in the CGANmodel is the contrast-limited adaptive histogram equalisation. The SURF technique uses scale-invariant feature transformation to extract local features, while the OTSU algorithm speeds up picture segmentation without previous knowledge of the pictures. These algorithms are employed by the suggested model. By employing an image-generation technique, CGAN expands the input plant village dataset, which it then uses to detect and categorise a wide range of plant leaf diseases. Fungal, viral, and bacterial illnesses are the three main types of leaf diseases. More than 300 illnesses are included in these categories. Out of 18,161 different crop species, both major and small, at least 200 illnesses were found in the suggested study. Researchers use the Python Jupyter program in conjunction with the Kaggle Plant Village Dataset and farmer-collected leaf samples to conduct their study. With the suggested framework, we reach an accuracy rate of 99.2%.

Description:

.