为满足结构静态试验三维变形测试研究需要,建立了立体视觉测量方法的三维数学模型,得到了结构变形的离面及面内位移公式;研究了圆形标识点的边缘检测、特征识别及中心拟合算法,并应用极线约束准则完成了标识点的精确匹配跟踪.在上述图像处理算法的基础上,编译了基于MATLAB平台可独立运行的立体视觉结构变形分析软件.使用集成的立体视觉测量系统完成了室内及碱液环境作用6个月后的UV-GFRP加固混凝土梁的破坏性试验,并与万能试验机的测量结果进行比较分析.结果表明:2种测量方法的荷载一位移曲线走势一致,吻合较好,证明了立体视觉方法用于结构静态试验三维测量的可行性.%To meet the measurement requirement of three-dimensional deformation of structure static test, the three-dimensional mathematical model of stereo vision measurement method was founded, and the out-of-plane and in-plane displacement equations of structure deformation were obtained. The algorithms of edge detection, character recognition and center fitting were studied for circle identifying points, and the epipolar constraint criteria were used to match and track the identifying point precisely. The structure deformation analysis software of stereo vision which could run independently was compiled on the base of above imaging processing algorithms with MATLAB. The destructive tests of reinforced concrete beams strengthened with UV-GFRP were performed with the stereo vision measurement system integrated, which were put in the indoor environment and alkaline solution for six months separately. The test results were compared with the data of universal testing machine. Results show that the load-displacement curves are consistent in trend and in good agreement between two methods. It is verified that the stereo vision method is feasible for measuring the three-dimensional deformation of structure static test.
展开▼
