Image recognition method based on deep learning and application in rice disease recognition

Abstract

The invention discloses an image recognition method based on deep learning and an application in rice disease recognition. The method comprises the steps of obtaining an image training set containinga target object; performing data enhancement processing on the training set image by adopting image amplification and image contrast adjustment; obtaining a trained deep learning network, the traineddeep learning network being obtained by training an image training set and a constructed to-be-trained deep learning network, and the construction and training of the to-be-trained deep learning network being realized based on an auxiliary model; acquiring an image to be identified; identifying a target object in the image; according to the invention, the building and training processes of the deep learning network are completed by adopting an auxiliary model; an existing network model trained based on a big data set is utilized, part of weight parameters and a network layer are selected to build a to-be-trained network model, network fine tuning training is carried out by inputting an image training set, the training time is remarkably shortened, and the classification accuracy is improved.

Description

technical field [0001] The invention relates to the field of image recognition, in particular to an image recognition method based on deep learning and its application in rice disease recognition. Background technique [0002] The diagnosis and identification of rice diseases is of great significance to improve the quality of rice. Applying image processing and machine vision technology to rice disease identification has incomparable advantages over traditional manual diagnosis and identification methods, and improves the ability of crop disease monitoring and early warning. [0003] In the process of rice disease identification based on image recognition, the problems encountered include: [0004] (1) The background information of the rice disease image is huge, which brings difficulties to the segmentation process of the target area in the image; [0005] (2) There are various types of rice diseases, and the types and locations of diseases are different in the same growt...

AI Technical Summary

Problems solved by technology

[0006] (3) In the process of using the deep learning network model, a large number of rice disease images are required for training to achieve recognition accuracy. However, at present, rice disease images do not have such a large number of data sets, and the ideal classification effect cannot be achieved.

Images