河流水下高精度微形貌自适应视觉测量机理与模型研究

With 40% large and medium-sized reservoirs in dangerous period, people have been paid highly attention to regular safety checks of dams based on underwater detection. 3D micro-topography detection of underwater structures of dams with high precision and resolution can early detect the security hidden troubles, the reinforcement measures are taken actions to avoid the emergent fatal accidents, which can ensure the safe operation of the water conservancy projects. However, measurement resolution of existing underwater topography detection technology is low, which cannot meet the demand of micro-topography detection of river underwater structures. 3D topography detection based on vison can obtain the measurement resolution and accuracy of millimeter level. But because of the dynamic and complex characteristics of river water environment, there are serious problems of image degradation and nonlinear measurement difficultly to achieve. Underwater structured light system is studied as the research object in the project proposal. Light transmission mechanism in river water is firstly discovered. Adaptive underwater imaging theory is proposed, including real-time light source design, compensation and image restoration. Influence of refraction phenomenon on the model of vision measurement is discussed. Model of measurement based on refraction is built of underwater structured light system. Those studies will provide the new solutions to common problems in underwater vision system, and play an active role in promoting the development of underwater topography measurement technology.

随着我国近40%的大中型水库进入病险高发期，以水下探测为主的堤坝安全检测引起人们高度重视。通过对堤坝水下结构高精度、高分辨率微形貌探测，提前发现堤坝安全隐患，及时采取工程措施，防止事故发生，确保水利工程安全运行。然而现有的水下形貌探测技术，测量分辨率较低，无法满足堤坝微形貌探测需求。基于视觉的水下形貌探测技术作为一种新的测量方法，可以获得毫米级别的测量分辨率和精度。但由于河流水下环境动态复杂的特点，该技术还存在图像降质严重和非线性测量难以实现等问题亟待解决。本项目以水下结构光系统为研究对象，在研究河流水下光传递机理的基础上，提出以在线光源设计、补偿和图像复原为主的自适应水下成像理论；分析折射现象对视觉形貌测量模型的影响机理，建立基于折射的测量模型，搭建适用于河流水下的结构光测量系统。研究成果可为解决河流水下视觉系统共性问题提供新的解决方法，为水下探测技术的发展起到积极的推动作用。

我国拥有水库大坝将近10万座，其中近40%的大中型水库进入病险高发期，以堤坝安全检测目的的水下缺损探测越来越重要。通过对堤坝水下结构高精度、高分辨率微形貌探测，提前发现堤坝安全隐患，及时采取工程措施，防止事故发生，确保水利工程安全运行。然而现有的水下形貌探测技术，测量分辨率较低，无法满足堤坝微形貌探测需求。基于视觉的水下形貌探测技术作为一种新的测量方法，可以获得毫米级别的测量分辨率和精度。围绕河流水下高精度微形貌自适应视觉测量研究了水下光传递模型和该模型的在线微倾斜刀口测量方法，提出了适合河流水下图像的自适应增强算法，构建了便携式河流水下结构光形貌测量系统。该系统在2m的近距离测量过程中可以得到1.2mm的测量精度，并避免了水下标定操作，提高了系统的便携性，有利于扩大视觉测量系统在水下工程探测、水下三维形貌测量、水下结构缺陷检测等情况中的应用，为水下视觉测量系统的设计提供新的思路和解决方法。在水利智能传感和信息化方面也进行了初步的研究，为后续进一步服务水利信息化打下良好的基础。