Adaptive Enhanced Detection Network for Low Illumination Object Detection

Any small environmental changes in the driving environment of a traffic vehicle can become risk factor directly leading to major safety incidents. Therefore, it is necessary assist drivers automatically detecting factors during process using algorithms. However, besides making more difficult for judge changes, performance automatic detection networks low illumination scenarios also be greatly affected and cannot used directly. In this paper, we propose model based on deep learning test optimization image enhancement problems. The overall structure includes dual discriminators, encoder–decoders, etc. consists two main stages. first stage, input scene adaptively converted into standard through lighting conversion module. second assessed factors. experiments show that network overcome impact has high accuracy.