无人船监视图像反光区域检测与去除方法及实验验证

A detection and removal method for specular regions in USV monitoring images and experimental verification

受海上环境、太阳高度角和无人船航向等因素的影响,无人船监控图像中海面反光现象不可避免,进而干扰后续视频处理,例如海天线的提取和运动目标的跟踪等。本文提出了一种无人船视频监控海面反光区域的检测及去除方法。首先,利用RGB颜色通道中的最大值与最小值获取像素点的亮度比例关系,检测海面反光区域中的高亮像素点。然后,针对检测到的高亮像素点进行形态学膨胀操作,获取较完整的海面反光区域。最后,结合反光区域中像素点亮度值随其位置分布的变化特点,使用加权后的邻域像素进行高亮成分抑制,实现反光区域的去除,并完成了无人船监视图像处理实验验证。实验结果显示,所提出的方法可有效检测并去除海面反光区域。

Specular regions on the sea surface of unmanned surface vehicle （USV） monitoring images are inevitable because of the effects of factors such as maritime environment, solar elevation angle, and USV’s heading. They affect the visual processing later, such as the extraction of sea skylines and tracking accuracy of moving targets. In this paper, a method is proposed for the detection and removal of specular regions on the sea surface of USV visual surveillance. First, the highlight pixels in specular regions are detected using the intensity ratio obtained from the maximum and minimum values of RGB channels. Second, a complete image of specular regions is obtained by morphological dilation for the detected highlight pixels. Finally, weighted neighborhood pixels are used to suppress the highlight components according to the feature in which the intensity of pixels varies with their position distribution in the specular regions, thus achieving a removal result for specular regions. Experimental verification was carried out using a USV. The results show that the proposed method can efficiently detect and remove the specular regions on sea surface.