Optimization of Object Detection Algorithm in Low Light Scene

Abstract

The development of deep learning technology has given rise to the prosperity of the field of target detection. The existing target detection algorithms show extraordinary performance and robustness in the scene with sufficient light, but they do not perform well in the low light scene. In order to improve the detection performance of the existing target detection algorithm in the low light scene, a low light enhancement module is designed. According to the idea of multitask learning, the low light enhancement is integrated into the back end of the detection algorithm, so that the algorithm can learn the target detection and brightness enhancement tasks at the same time to improve the performance of target detection. The proposed method can effectively improve the detection performance in the low light scene without increasing the reasoning cost of the algorithm.