【摘要】 针对现有视频车辆检测算法受光照、阴影等环境因素影响大,漏检和误检率高的问题,提出了一种视频车辆检测算法.有别于传统算法使用运动特征进行车辆检测,该算法使用边缘特征和对性特征定位车辆.算法首先对图像进行灰度化、平滑去噪等预处理,使用Sobel算子垂直方向掩模计算图像感兴趣区域内的边缘梯度,确定候选区域;而后根据车辆图像垂直边缘具有对称性的特点,分析候选区域的对称性强弱,并计算其对称轴位置和车辆宽度.使用边缘强度、对称性和宽度这3个约束条件对候选区域进行验证.道路实验结果表明,该检测算法有效、可靠,具有良好的鲁棒性.更多还原

【Abstract】 The existing vehicle detection algorithms are subject to many environmental influences,such as different lighting conditions,shadows.To solve these problems,a novel vehicle detection algorithm was proposed.Different from traditional methods,which use motion features to detect vehicles,the proposed method used edge and symmetry features to locate possible vehicles.First,the color-image was preprocessed with grayscaling,smoothing.The vertical gradient components of the image were found using the vertical mask of Sobel operator.Then the candidate areas with sufficient amount of edges were determined by calculating the gradient of edges.After finding possible vehicle candidates,the proposed method measured the prominent vertical symmetry of candidate areas and verified all possible candidates,as symmetry is an important feature of vehicle.The proposed method also figured out the vertical axis of symmetry and the width of the candidate vehicle.Three important restrictions obtained from above steps,including edges,symmetry and width,were used to validate the candidate vehicles.The experimental results show that the proposed algorithm is effective,reliable and robust.更多还原