基于双目视觉的输送带表面损伤定量分析方法

Quantitative analysis method of conveyor belt surface damage based on binocular vision

为了对输送带表面损伤进行定量分析以有效判断其损伤程度,并根据损伤程度选择检修方法和时间,为工作人员检修提供依据,提出了一种基于双目立体视觉的输送带表面损伤定量分析方法。该方法首先对双目相机进行相机标定,对输送带表面损伤图像进行图像处理校正、立体匹配,基于视差原理根据三角测量原理获取目标点的三维空间信息,通过数学计算模型对损伤程度进行定量分析,提出了改进的Census变换和SGM(semi-global matching)算法相结合的立体匹配算法。结果表明:提出的改进算法匹配效果有所提升,和Census变换算法及梯度算法进行对比,误匹配率分别降低了5.26百分点和3.92百分点,得出的视差图精度最高;通过提出的方法对输送带表面损伤进行定量分析,在输送带损伤长度、宽度、深度三者误差中,损伤深度误差最大为8.5%,仍在可接受范围。

In order to quantitatively analyze the surface damage of conveyor belt and effectively judge its damage degree,maintenance methods and time are selected according to the damage degree to provide basis for staff maintenance.A quantitative analysis method of conveyor belt surface damage based on binocular stereo vision was proposed.In this method,the binocular camera was calibrated first,the conveyor belt surface damage image was processed and corrected,and the three-dimensional matching was carried out.The three-dimensional spatial information of the target point was obtained based on the parallax principle and the triangulation principle,and the damage degree was quantitatively analyzed through the mathematical calculation model.A stereo matching algorithm based on improved Census transform and SGM(semi-global matching)is proposed.The results show that the matching effect of the proposed algorithm is improved.Compared with Census transform algorithm and gradient algorithm,the mismatching rate is reduced by 5.26 percentage points and 3.92 percentage points,respectively,and the parallax map accuracy is the highest.The proposed method is used to quantitatively analyze the surface damage of the conveyor belt.Among the errors of the length,width and depth of the conveyor belt damage,the maximum error of the damage depth is 8.5%,which is still within the acceptable range.