基于改进Hough变换和图搜索的油库目标识别

为了识别遥感图像中圆形油库目标，首先改进了基于梯度信息的圆形检测Hough变换方法，提取出图像中的圆形油库。然后根据油库的空间分布关系，提出利用深度优先的图搜索策略对检测到的圆进行分组，剔除虚警目标，最终实现油库目标区域的定位。改进的Hough变换通过利用梯度的方向信息和降低参数空间维数的方法降低了算法执行时耗费的时间和占用的存储空间，提高了圆检测的效率，同时用图搜索技术来排除虚假目标和定位目标区域，降低了虚警率，提高了识别精度。实验表明，该方法能够快速准确地识别油库目标，适用于不同分辨率的可见光遥感影像。

Oil Depots Recognition Based on Improved Hough Transform and Graph Search

In order to identify the circular oil depots from remote sensing images, a developed Hough transform method based on gradient information is proposed to extract circular oil tanks firstly. Then, the depth-first graph search strategy is employed to group the detected circles and eliminate the false alarms according to the spatial distribution of the oil depots. Finally, the target areas of oil depots are localized. The improved Hough transform reduces the time and space consumption by using the gradient direction information and reducing the dimension of parameter space, and improves the efficiency of circles detection. The graph search strategy can exclude the false targets and locate the target areas, which improves identification accuracy. The experimental results indicate that the proposed algorithm can recognize the oil depots targets fast and accurately, which is suitable for optical remote sensing images of different spatial resolutions.