Shadow Detection Algorithm for Moving Vehicles Based on Color Space and Texture
Yanbin Fan · Ceshi jishu xuebao · 2013
According to the shadow characteristics of moving vehicles,the two shadow detection algorithms respectively based on the RGB color space and texture are improved,and a shadow detection method integrating color and texture is presented.In such an algorithm,the fact is used that the brightness of pixels in shadow region is lower than that in background,while the colors and textures remain the same.By computing and comparing the color distortion,brightness distortion and first-order gradient,the shadow pixels are recognized,and then the shadow pixels and non-shadow pixels are distinguished.Additionally,the effect of shadow detection can be further improved,which the foreground objective after removing the shadow is reconstructed by using the morphological method.Finally,the efficiency of vehicle detection can be increased by setting the ROI area near the driveway in video.The experimental results show that the proposed approach has good performance in shadow detection for moving vehicles.