【摘要】 主要研究无人机降落图像的地表积水的识别。首先分析了积水的反射特性,建立了具体的检测方法,通过灰度变换、二值化、数学形态学中的腐蚀等算法对积水的图像进行预处理,通过轮廓连接与提取得出积水边缘信息;然后对提取出的边缘信息进行方形拟合并选定感兴趣区域;最后对前后两帧图像进行角点检测,选定合适的阈值,根据反射景物出现前后角点变化识别积水。结果表明,此种方法可以准确、快速地将积水提取出来,为无人机提供安全着陆信息。更多还原

【Abstract】 This paper mainly studies the recognition of water accumulation on flat ground in UAV landing image.Firstly,the reflective characteristics of accumulated water are analyzed,and a specific detection method is established.The image of accumulated water is preprocessed by gray transformation,binarization and corrosion in mathematical morphology.The edge information of accumulated water is obtained by contour connection and extraction.Then,the extracted edge information is fitted by square and the region of interest is selected.Finally,the corner points of the first and second frames of images are selected.Detection,selection of appropriate threshold,according to the reflection of the scene before and after corner changes to identify water accumulation.更多还原