水下机器人视觉系统图像处理研究

发布时间：2018-05-28 04:42

本文选题：水下机器视觉 + 图像处理　；参考：《天津理工大学》2017年硕士论文

【摘要】：人类地球的七成以上都是由海水覆盖,海洋生态资源的使用对于全球经济起着至关重要的作用。近几个世纪,由于人口数量的骤增,人类面临着严重的资源匮乏问题。世界各国都加强了对海洋环境的探索、海洋科技的发展和海洋资源的利用。人类社会资源的可持续发展要求人们必须加大对海洋产业的开发利用。其中,水下机器人技术是人类借助现代科技技术来探测神秘海底世界的重要手段之一。为了更好的发展水下机器人技术,人们需要开发更为先进的水下机器视觉技术进行准确清晰的海底资源探测和开采。因而,水下目标图像清晰化的处理技术在海底勘探、水下物体轨迹跟踪、定位等领域中至关重要。光通过水体介质传播过程中的衰减,为水下图像成像带来了一系列的问题。水下目标在成像时会受到海水水质及悬浮物等物理性质的影响,使采集到的图像模糊、对比度较差。水下视频与图像增强算法的研究,一直是机器视觉领域中备受关注的方向,具有十分重要的理论意义和实际应用价值。本课题从水下机器视觉系统的外部机械安装和视觉系统图像清晰化处理两条技术路线出发,解决水下机器人视觉系统中遇到的难题。针对水下图像成像过程中遇到的问题进行分析,确定影响水下机器视觉系统进行拍摄的关键因素,结合实验结果分析,给出了较为实际的一种处理思路。设计并优化了一种新的图像处理算法,通过实验验证取得了良好的效果,对水下机器人视觉系统的研究提供了指导性建议。通过对比度的调整实现了水下图像可视度的增强和图像细节信息的有效放大,使得由于水底光照不均匀带来的问题能得到有效缓解。本文利用水下机器人摄像机采集到焊缝图像,并通过算法的优化对水下降质图像进行预处理。从中提取出焊缝位置信息,为水下机器人对管道焊缝缺陷的检测提供数据支持。通过一系列的算法对图像进行预处理,利用减影法分割出焊缝区域并对裂纹缺陷的有效信息进行检测处理。该方法的实施为水下石油管道焊缝处缺陷检测流程提供依据。
[Abstract]:More than 70% of the earth is covered by sea water. The use of marine ecological resources plays a vital role in the global economy. In recent centuries, due to the rapid increase in population, mankind is faced with a serious problem of lack of resources. Countries all over the world have strengthened the exploration of marine environment, the development of marine science and technology and the utilization of marine resources. The sustainable development of human social resources requires people to increase the exploitation and utilization of marine industry. Among them, underwater vehicle technology is one of the important means for human beings to explore the mysterious undersea world by means of modern science and technology. In order to better develop underwater vehicle technology, people need to develop more advanced underwater machine vision technology for accurate and clear exploration and exploitation of submarine resources. Therefore, the processing technology of underwater target image clarity is very important in the field of underwater exploration, underwater object trajectory tracking, positioning and so on. The attenuation of light propagating through water media brings a series of problems for underwater image imaging. The underwater target will be affected by the physical properties such as the water quality of seawater and suspended solids when imaging, which makes the collected images blur and the contrast is poor. The research of underwater video and image enhancement algorithms has been the focus of attention in the field of machine vision, which has very important theoretical significance and practical application value. Based on the two technical routes of external mechanical installation and image clarity processing of underwater machine vision system, this paper solves the problems encountered in underwater vehicle vision system. Based on the analysis of the problems encountered in the underwater image imaging process, the key factors affecting the underwater machine vision system are determined, and a practical approach is given based on the analysis of the experimental results. A new image processing algorithm is designed and optimized, and good results are obtained through experiments, which provide guidance for the research of underwater vehicle vision system. The enhancement of the visibility of underwater images and the effective amplification of the details of the images are realized by adjusting the contrast, which can effectively alleviate the problems caused by the uneven illumination at the bottom of the water. In this paper, the image of weld seam is captured by underwater vehicle camera, and the image of water descent quality is preprocessed by optimization of algorithm. The position information of weld seam is extracted from it to provide data support for underwater vehicle to detect pipeline weld defect. The image is preprocessed by a series of algorithms, the weld zone is segmented by subtraction method and the effective information of crack defects is detected and processed. The implementation of this method provides the basis for the inspection flow of the weld seam of underwater oil pipeline.
【学位授予单位】：天津理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP242

【相似文献】