A Comparative Study of ResNet50, EfficientNetB7, InceptionV3, VGG16 models in Crop and Weed classification

With the growing technology and the use of the same in the field of farming, it is becoming imperative that technology be used in the classification of crops and weeds as it is a growing menace and will help the farmers to identify and distinguish between them accurately. The problem of weeds in the cultivation of crops is a serious one and hampers the growth of food crops. As such the use of machine learning algorithms to ease the labor-intensive industry is not only justified but also efficient. In the following study, a comparative analysis of the popular pre-trained deep learning models, namely ResNet50, EfficientNetB7, InceptionV3, and VGG16 has been presented. From the study, it can be concluded that in terms of accuracy, EfficientNetB7 performs the best followed by ResNet50 which is more lightweight as compared to the rest and more suitable. On the other hand, it is seen that VGG16 may be more suitable for scenarios where computational resources are scarce. Our findings indicate that while all the models tested are effective in crop and weed classification, selecting the most appropriate model depends on the specific requirements of the application.