一种基于边缘特征提取的景像匹配方法

下视景像匹配制导中,实时图和基准图的拍摄高度、方位和图像比例尺不完全一致,两图的轮廓形状存在旋转和变形等差异,边缘特征提取方法受图像灰度分布影响存在变形和移位,易导致匹配错误。用侧抑制竞争方法,图像边缘特征提取快速准确,不发生移位,对灰度变化也不敏感。在此基础上,用扩展边缘特征法,克服轮廓变形对边缘匹配的影响,提高精度。试验结果表明,此方法模型简单、抗几何变形能力强且便于实现。

A Scene Matching Navigation Method Based on Image Edge Character Pick-up

Since the real-time image and norm image shoot height,orientation and image size are different,the two image's outline exist rotation and distortion,and so obtained image's edge character is influenced by picture's gray distribution,so that the match result is usually not correct.Using side restrain compete method can obtain image edge character fast,correctly,moreover the character results have no shift,and are not sensitivity to gray.Then by using enlarged edge character method,outline distortion can be overcomed and correct image can be matched.The tests results show this method is simple,and can strong resist against geometry disorder and can be realized easily.