【作者】 吴丽蓉；

【导师】 徐安；

【作者基本信息】 同济大学 ， 检测技术与自动化装置， 2007， 硕士

【摘要】 图像是人类获取信息的主要途径。本文基于墙体渗漏判断的实际应用，主要探讨了如何采用图像数据处理技术为检测建立更实用更精确的数字化处理平台。本文完成了以下主要工作：(1)根据数据采集的工作环境和采集要求，使用模块化和平台化设计技术思想搭建了墙体渗漏数字化检测装置，同时完成与上位机的通信与传输，并在上位机内完成数据的二维重组与三维重构。(2)对摄像机标定，图像滤波，特征点匹配，双目视差，几何变换，边缘检测等一系列数字图像及数据处理算法进行了分析并应用于墙体渗漏检测的数字化处理，提供墙体渗漏判定的可靠的依据。(3)在上述研究的基础上，对检测现场的采集数据和实景图像的数据融合方法进行了探讨，并应用模糊控制理论对渗漏原因和程度做出科学客观的分析。经实验验证，研制的采集子系统和处理子系统能够完成需要的基本功能，图像数据处理过程中采用的算法满足了应用需求的精确度。更多还原

【Abstract】 The technology of digital image processing has applied widely in various fieldssuch as medical treatment and diagnosis, resource and environment, weather andtraffic detection and so on, and has created great social values. This paper mainlydiscusses that how to build more utility and more accurate digital procession platformfor wall water-leakage detections using digital image procession technology, which isbased upon the application of wall water-leakage evaluation.The main work shows as follows:(1) According to the working environment and acquiring demand, modular andplatform technology is applied to the designation and realization of wallwater-leakage, both the hardware and software structure. Meanwhile, buildcommunication between micro-chip and PC machine, recompose and display theoriginal data in PC.(2) Analyzing, comparing and applying several data process and digital imageprocess algorithms, such as stereo vision system, edge detection, neural network,fuzzy control, feature identification and so on.(3) Based on the researches mentioned above, discuss data fusion method tosampled data and field images, and introduce the fuzzy control theory into the projectto generate a final water-leakage cause evaluation report.The experiment proves that both the data acquirement system and the dataprocess system meet the needs, and the algorithms used in image processing reach theprecision requirement.更多还原