基于夜间辅助驾驶的图像预处理方法研究

辅助驾驶中的前车车辆检测及灯语识别能够有效减少夜间行车时交通事故的发生。提出一种夜视图像的预处理方法,能够对夜视图像中存在的白色眩光和红色眩光进行有效去除,为夜间行车情况下的车辆检测及灯语识别算法创造良好的应用环境。基于暗通道先验去雾理论进行计算的简化,针对夜视图像中的红色眩光问题提出相应的去除方法,能够在保证处理效率的前提下,保证对夜视图像中的白色眩光和红色眩光的去除效果。

Research on Image Preprocessing Method Based on Aided Driving At Night

Front vehicle detection and lamp language recognition in assisted driving can be used to effectively reduce traffic accidents during night driving.This paper proposes a preprocessing method for night vision images,which can be used to effectively remove white and red glare in night vision images,and create a good application environment for vehicle detection and lamp language recognition algorithms at night.The simplification of the dark channel prior defogging theory,and the corresponding removal method for the red glare problem in night vision images can be used to effectively remove the white and red glare in the night vision images under the premise of ensuring processing efficiency.