电力监控全景图像拼接及接缝优化技术研究与实现.docxVIP

PAGE 21 电力监控全景图像拼接及接缝优化技术研究与实现 摘 要 电力监控全景图像拼接理论、技术和算法实践是目前图像处理领域的研究热点，它可以很好的解决当前工业领域大面积监控需求的痛点和难点：现实中因成像设备的限制、地理位置、视角制约等问题导致监控不能直接拍摄到超大面积的监控图像，只能采用人工手动观察来自多机位的不同显示画面。由于机位数量多，位置分散，因此需要电力监控全景图像，该研究主要是针对电力行业中对于环境的全面了解，对输电线路周边环境进行多角度图像采集，并采用图像拼接算法对图像完成拼接，优化接缝，使得接缝过渡平滑。 所谓的电力监控全景图像拼接就是场景之中两张以上具有部分重叠的照片彼此之间的关系匹配。通过在有一定的联系的图片之间进行采样融合可得到一张包含多图像信息、广角、完整新的高清图像，有利于改善全景图像采集设备（例如广角镜头或鱼眼镜头）图片边缘易畸变的缺点，解决拼接融合区域过程中产生的巨大阴影和不自然的图像连接问题。 本文以两台监控摄像机从不同的位置和视角获取到的电力现场图像为研究对象基础，深入探究基于SIFT的尺度不变特征变换匹配算法。传统的拼接技术是选取一幅参考图像，其余所有图像都同参考图像对齐、融合成全景图，该方法由于积累误差的存在会造成拼接后的图像存在扭曲和失真现象，且相邻图像之间接缝优化效果很不理想。而SIFT算法主要针对局部进行描述符识别，图片之间除识别点外的整体允许有显著不同，适合用于处理电力监控生成的图像。通过MATLAB软件建模对电力监控图像进行深度的识别，包络图片预处理，特征点提取、检测、描述（图片相似值）和匹配，图像融合以及降噪以此来生成电力监控全景图，并在机械性拼接后的全景图上进行去接缝（光学优化）等步骤，为电力监控全景视图处理提供解决方案。 关键词：全景图像；接缝优化；SIFT算法；信息可视化展示； Research and Realization of Power Monitoring Panoramic Image Stitching and Seam Optimization Technology Abstract Power monitoring panoramic image splicing theory, technology and algorithm practice is the current research hotspot in the field of image processing, it can well solve the current industrial field of large area monitoring needs of pain points and difficulties: in reality, due to the limitations of imaging equipment, geographical location, viewpoint constraints and other problems lead to monitoring cannot directly capture the monitoring image of the large area, can only be manually observed manually from multiple locations of different display screens. Due to the large number of stations and scattered locations, the need for power monitoring panoramic images, this study is mainly for the power industry's comprehensive understanding of the environment, multi-angle image acquisition of the environment around the transmission line, and the use of image stitching algorithm to complete the image stitching, optimize the seams, make the seam transition smooth. The so-called power monitoring panorama image stitching is the relationship between two or more partially overlapping photos in the