深海auv图像序列拼接关键技术分析-key technology analysis of image sequence mosaic for deep sea auv.docxVIP

浙江理工大学硕士学位论文深海 浙江理工大学硕士学位论文 深海 AUV 图像序列拼接关键技术研究 II万方数据 II 万方数据 ABSTRACT In the 21st century, with the increasing depletion of land resources, exploration and development of marine resources are more and more urgent. But people cannot explore on the spot because the technology that can be utilized is limited. To get a good knowledge of the undersea environment, relying on the existing deep-sea exploration equipment and operations cannot meet the needs. The autonomous underwater robot (AUV) is commonly used to detect the undersea situation. In the various types of data acquired by AUV, the image data has the advantage of being clear and concise. The images acquired are usually limited to the view field and resolution ratio, . However, the images can be spliced into the images with a larger view field and high resolution ratio by image stitching technology. In this paper, the deep-sea image sequences research is conducted by the following aspects: The paper is conducted on the purpose of improving the efficiency and stability of the target images stitching. We propose a stable and less-time consuming method of rapidly pairing deep-sea sequence images on the basis of analyzing the continuity time and space of the sequence images shot by underwater robots. Firstly, we analyze the adjacent sequence pictures with the same intersection area by employing the detected feature pair points, and then calculate the parameters of the coordinate transformation of the camera. Secondly, Kalman filter with the aim of the known camera motion parameters to forecast the camera movement trend is establish. At last, the detected feature points regionally are matched on the basis of the estimation of intersection area of pair images by using the coordinate transformation above, the frequency of matching feature points is reduced, and the matching efficiency is improved. The experimental results show that the proposed method ensures a good matching effect and the splicin