Road surface target lightweight enhancement method based on high-speed sensing

The invention discloses a pavement target lightweight enhancement method based on high-speed sensing, and the method comprises the steps: firstly obtaining a pavement image through a video sensor of a highway, then designing a generative adversarial network for image enhancement, and then carrying out the lightweight of the generative adversarial network in a plurality of modes, and the parameter quantity, the storage volume and the training time of the generative adversarial network are reduced, so that the generative adversarial network is more efficient and convenient. The method has practical significance for road monitoring and maintenance engineering. 本发明公开了基于高速传感的路面目标轻量化增强方法，本方法首先利用高速公路的视频传感器获取路面图像，然后设计用于图像增强的生成对抗网络，之后使用多种方式对生成对抗网络进行轻量化，减少生成对抗网络的参数量、存储体积、训练时间，使其更为高效和便捷。本发明对于道路监测和养护工程具有实际意义。