Highway pavement disease recognition inspection system based on deep learning

Abstract

The invention discloses a highway pavement disease recognition and inspection system based on deep learning, and relates to the technical field of automatic detection and recognition of highway pavement diseases. The mobile image acquisition module is deployed on an inspection vehicle and is used for continuously acquiring an image sequence of a highway pavement in a moving process; and the image processing module is in communication connection with the mobile image acquisition module and is used for processing the acquired image sequence in real time through a built-in pavement disease identification deep network model. According to the invention, through the adversarial domain adaptive network, the model is actively enabled to learn characteristics insensitive to environmental changes, environmental invariance is realized, and the stability performance in different environments is effectively ensured; the two types of features are decomposed into an independent submerged space through a multi-loss mechanism constrained feature decoupling encoding-decoding device and a forced model, so that pure and material-independent disease features are obtained, the generalization ability of cross-material pavement deployment of a system is improved, and the problem of false detection or missing detection of diseases is reduced.

Images