CAJ | 학술논문

阴影检测在车辆检测中具有关键作用.针对目前车辆检测算法中阴影常被误检为前景目标的问题,提出了一种基于多源信息的阴影检测算法.首先进行前景目标的提取;然后分别进行基于边缘、HSV颜色空间和光学不变性信息的阴影检测;最后根据三种信息检测结果的综合分析确定阴影点.实验结果表明,该算法具有很好的阴影检测效果.
음영검측재차량검측중구유관건작용.침대목전차량검측산법중음영상피오검위전경목표적문제,제출료일충기우다원신식적음영검측산법.수선진행전경목표적제취;연후분별진행기우변연、HSV안색공간화광학불변성신식적음영검측;최후근거삼충신식검측결과적종합분석학정음영점.실험결과표명,해산법구유흔호적음영검측효과.
Shadow detection plays a key rule in the vehicle detection. Aiming at the problem that shadows are detected incorrectly as foreground objects in the vehicle detection, a novel shadow detection algorithm based on multi-source information was proposed hereby. Firstly, foreground targets are gained. Secondly, the shadow detection algorithms based on edge, HSV color space and optical invariant information are executed, respectively. Finally, the shadow points are determined by comprehensive analysis of the three information detection results. The experiment results show that the proposed algorithm is highly effective for shadow detection.