Agro-vision IoT-enabled Crop Pest Recognition System based on VGG-16

It is more difficult to implement recognition of crop pests in the field than the generic object recognition task because of the high variation among crop pest species which usually leads to the misclassification of crop pests from the same class. Furthermore, in deploying these recognition systems to the field, intensive hardware setups are mostly used. To improve the recognition accuracy, this paper proposes an IoT-enabled crop pest recognition system (CPRS) that utilizes Visual Geometry Group (VGG-16) which uses the multi-level learning features obtained from pest images for the automatic classification of crop pests. The Accuracy result of 88.72% obtained validates the reliability and robustness of the proposed method as capable of detecting the presence of pests in different leaf image samples.