论文部分内容阅读
针对海岸线背景下的海面小目标的自动检测问题展开讨论,提出一种目标所在感兴趣区域的自动提取算法,将水域和复杂背景分离,并对分割后的二值图像进行标记,海面区域即水域应该是标记块中面积最大的一块,利用此特征提取出感兴趣区域——水域。对感兴趣区域进行轮廓跟踪后再用原始灰度进行扫描填充,在得到的简单背景下对目标进行局部中值滤波,进而用F-R准则将感兴趣舰船目标检测出来。最后拟利用感兴趣舰船目标区域的方差值计算感兴趣目标出现的置信度估计值,若置信度大于90%,则认为是真目标。实验结果给出了实验处理时间和所识别出的各个目标的置信度,表明了本文算法的有效性。
Aiming at the problem of automatic detection of small sea targets in the coastline, this paper proposes an automatic extraction algorithm for the region of interest where the target is located, separates the water and the complex background, and marks the binary image after segmentation. The sea area Should be the largest block in a block, using this feature to extract the region of interest - waters. After contouring the region of interest, it is scanned and filled with the original grayscale. The local median filter is applied to the target in the simple background, and then the target of the ship of interest is detected by the F-R criterion. Finally, we will use the variance of the target area of the ship of interest to calculate the confidence estimate of the target of interest. If the confidence is greater than 90%, then it is considered as the true target. The experimental results show the experimental processing time and the confidence of each target identified, which shows the effectiveness of the proposed algorithm.