顾及局部畸变的侧扫声呐图像多特征保形镶嵌研究

High-quality side scan sonar image with accurate location and clear shape outlines play a basic and guarantee role in marine engineering and marine resource development, etc. However, side scan image still exist local distortion with changing type and size everywhere, this result in that the adjacency bands are hard to realize the same goal conformal mosaic through the simple offset and rotation. Thus, the study on side scan sonar image mosaic considering local distortion is of great importance in eliminating local distortion, solving the target misplacement and ghosting..Based on the pre-processing of the side scan sonar data, the reconstruct of the side scan sonar image is firstly conducted considering the tow fish motion state, whose parameters and model are then estimated and corrected separately using the finite measuring information and cross correlation coefficient changing characteristics between different pings. After the first step, the research of multi-features joint registration is developed and to make sure that the registration relationship can be used even when poor and complicated characteristics exist on the sonar image. Lastly, the blocking mechanism is induced according to the above characteristics, and the linear method of transformation parameter or the thin plate spline elastic transformation method is adopted to control the local distortion and the conformal mosaic of common targets. At last, a high quality large area side scan sonar image is formed with accurate position, clear contour and without misplacement and ghosting..The theory and method proposed by this research, have an important theoretical and practical significance in the improvement of side scan sonar system measurement results application level.

位置准确、目标轮廓清晰的高质量侧扫声呐海底地貌图像对海洋工程实施、海洋资源开发等均具有重要的基础保障作用。然而，侧扫声呐图像各处存在类型和大小不断变化的局部畸变，造成相邻条带图像难以通过简单的平移和旋转来实现共视目标的保形镶嵌，开展顾及局部畸变的侧扫声呐图像保形镶嵌研究，对消除局部畸变、解决目标错位和重影问题具有重要意义。.项目在侧扫声呐测量信息预处理的基础上，首先开展顾及拖鱼运动状态的侧扫声呐条带图像重构研究，构建拖鱼运动状态参数的估计和修正模型；在此基础上，开展相邻条带图像的多特征联合配准研究，实现配准关系在贫特征和复杂特征区域的全覆盖；最后，引入分块机制，实现局部畸变的控制和共视目标的保形融合，最终形成目标位置准确、轮廓清晰无错位和重影的高质量大区域侧扫声呐图像。.项目研究提出的理论和方法对于提高侧扫声纳系统测量成果应用层次具有重要的理论和实际意义。

位置准确、目标轮廓清晰的高质量侧扫声呐海底地貌图像对海洋工程实施、海洋资源开发等均具有重要的基础保障作用。然而，侧扫声呐图像各处存在类型和大小不断变化的局部畸变，造成相邻条带图像难以通过简单的平移和旋转来实现共视目标的保形镶嵌，开展顾及局部畸变的侧扫声呐图像保形镶嵌研究，对消除局部畸变、解决目标错位和重影问题具有重要意义。.项目在侧扫声呐测量信息预处理的基础上，首先开展顾及拖鱼运动状态的侧扫声呐条带图像重构研究，构建拖鱼运动状态参数的估计和修正模型；在此基础上，开展相邻条带图像的多特征联合配准研究，实现配准关系在贫特征和复杂特征区域的全覆盖；最后，引入分块机制，实现局部畸变的控制和共视目标的保形融合，最终形成目标位置准确、轮廓清晰无错位和重影的高质量大区域侧扫声呐图像。.项目研究提出的理论和方法对于提高侧扫声纳系统测量成果应用层次具有重要的理论和实际意义。