拖曳作业模式会引起侧扫声纳条带图像错位和畸变,并给基于地理编码的条带图像整体拼接以及海底目标的识别和判读带来问题。提出了一种基于共视目标的相邻侧扫声纳条带图像分段拼接方法。该方法通过对侧扫声纳条带图像预处理、基于共视目标的分段匹配、基于小波变换的公共覆盖区图像信息融合,最终实现了相邻侧扫声纳条带图像拼接,最大限度地削弱了错位和畸变对声纳图像质量的影响。试验验证了该方法的有效性。
The towing operation mode results in dislocation and distortion of SSS images, as well as creating difficulties for image mosaicing by coordinates, and target recognition and determination. Therefore, this paper proposes an image segmented mosaic method based on corresponding features on the seabed. Through preprocessing of side scan sonar image, seg- mented matching by mutual features, and image fusion in the public coverage area based on wavelet transformation, this method can correctly fuse adjacent SSS images, going a ,long way toward solving the problem of dislocation and distortion found in the traditional methods fnr image mosaic. A test verifies the effectiveness of the proposed method.