基于感知编组的遥感图像港口提取方法

Harbor Detection Method of Remote Sensing Images Based on Perceptual Organization

遥感图像中港口识别是遥感海洋应用研究的重要方面,准确、高效地自动提取港口目标对于提高遥感图像自动解译能力具有重大意义。由于人造目标按照某种结构方式有组织地构成,而人类视觉的感知编组可以获取图像中相关组织及结构,因而感知编组的特性有助于识别人造目标,该文将感知编组方法用于提取港口目标。通过建立突堤区域编组、疑似港口突堤端点连线编组以及基于产生式规则的港口编组机制,提出了基于多级感知编组的港口提取方法。实验结果表明,相对于已有的港口检测方法提取的港口目标不准确、效率不高等问题,新方法提取的港口具有目标标识正确、轮廓完整以及运算速度快等特点。

Harbor recognition is an important aspect for remote sensing ocean application research. Accurate and fast harbor detection has great significance for improving the ability of automatic interpretation for remote sensing images. Man-made Object is composed of some components according to certain structure,besides perceptual organization of human vision can acquire related organization and structure of images in the case of ignoring prior knowledge of image content. Therefore perceptual organization is helpful for recognizing man-made object,this paper introduce perceptual organization principle into harbor detection. This paper proposes a kind of harbor detection method based on perceptual organization by establishing group of jetty areas,group of lines between jetty butts and harbor organization mechanism based on production rules. Compared with an existing harbor detection algorithm that has inaccurate harbor contour result and inefficiency,our experiments show that novel harbor detection method has an accurate result,a whole contour and faster operation.