鱼眼图像的目标物体角点检测方法

为实现宽视场角图像中目标物体的定位,提出了一种鱼眼图像的目标物体角点检测方法。首先对含有目标物体的鱼眼图像进行等面积切分,利用目标物体即网格图像灰度直方图的明显特征确定目标区域;然后根据实际应用对Harris算法进行改进,并将其应用至目标物体的角点检测中,确定目标物体位置。实验结果表明,能有效确定目标区域,视觉冗余度由90.3%下降到47.8%,目标物体角点检测准确率为83.8%,可以为实现宽视野范围目标的位置与距离同时测量奠定基础。

Target object corner detection method for fisheye image

In order to realize the positioning of target objects in wide-angle angle images, a target object corner detection method for fisheye images is proposed. The method firstly divides the fisheye image containing the target object into equal areas, and the target area is determined by the obvious features of the gray histogram of the target object, that is, the grid image. Then the Harris algorithm is improved according to the practical application, and applied to the corner detection of the target object to determine the position of the target object. The experimental results show that this method can effectively determine the target area, the visual redundancy decreases from 90.3% to 47.8%, and the target object corner detection accuracy is 83.8%. This method can lay the foundation for simultaneous measurement of the position and distance of a wide field of view target.