IoT-Enabled Deep Learning Framework for Early Pest Detection using Drone Imagery and Image Segmentation Techniques

One of the major participants in the global agricultural precinct is India. Despite the fact that agriculture has many sophisticated technology, there is currently not any appropriate technology to manage pest-related issues. This suggested approach uses Deep Learning (DL) to explain IoT for early pest detection in agricultural fields. Data collection, pre-processing, data augmentation, segmentation, and detection make up the following suggested model. The data is collected from the drone images, after that pre-processing is done to get the clear image with removal of noise and get smooth edges using Bilateral Filter (BF), data augmentation is done to get more number of training images, after that, segmentation is followed by U-net to segment the insect from its background and finally, AlexNet is used to detect the pest or not. The proposed model is applied in python platform also it attains accuracy of 99.47% compared with existing techniques.