Low-quality picture target detection method based on bilateral grid and YOLO

The invention discloses a low-quality picture target detection method based on bilateral grids and YOLO. The low-quality picture target detection method comprises the steps of enhancing an input picture and performing target detection by using a YOLO target detection model. According to the method, on one hand, the low-illumination target detection capability is improved, the target detection accuracy of the low-illumination image is enhanced, and on the other hand, the target detection speed is increased. 本方法公开一种基于双边网格与YOLO的低质量图片目标检测方法，包括了对输入图片进行增强和利用YOLO目标检测模型进行目标检测的步骤。本发明一方面提高了低照度目标检测能力，增强了对低照度图像进行目标检测的准确度，另一方面加快了目标检测速度。