Abstract: Based on the continuous development and integration of agricultural technology and information technology, residual neural network pest image recognition system on Asp.NET Core MVC architecture was proposed to solve the low accuracy and efficiency of crop pest recognition in Hebei province. Firstly, the system collects the image information of target classification by mobile acquisition terminal and network image crawler, and uses data enhancement technology to expand the sample library to obtain the data set of neural network training model. Then by building machine learning framework, ResNet-50, ResNet-101, ResNet-152 residual network models are introduced to train the data set and verify its accuracy. Finally, the most accurate training results model is applied to crop pest classification service system. The identification model has good applicability and robustness, which can provide identification and diagnosis functions for pests of main crops in Hebei province.

Key words: agriculture, pest, image recognition, data set, model