煤仓煤位双目视觉检测自适应标定技术研究

Study on new adaptive calibration technique of coal level detection in coal bin with binocular vision

为了研究视觉图像测距检测煤位的理论和方法,运用基本的针孔成像模型理论,提出了一种非接触式高效的基于双目视觉图像自适应标定的煤仓煤位深度测量方法。通过结合Faugeras标定方法以及卫星定位原理,根据预设定的视觉传感器坐标以及在双目图像中提取的关键轮廓角点,代入三维坐标自适应公式,运用最小二乘法得出视觉传感器的内外参数,并使用线性优化对内外参数值进行优化,得到最终的摄像机内部和外部参数。最后,根据采集图像的像素信息,确定煤仓煤位的深度值,对整体参数RMS测量误差与张正友标定算法进行了对比分析。该方法不需要任何已知的摄像机参数,简单方便地求解。实验结果表明该方法精度较高,能满足煤仓煤位检测的需要,且使煤位检测标定过程实现了自动化。

In order to study the theory and method of coal level detection with visual image ranging, using the basic principle of pinhole imaging model, we proposed a non-contact and efficient coal bin coal depth measurement method based on adaptive calibration of binocular vision image. Combining Faugeras calibration algorithm and satellite positioning principle, the preset visual sensor coordinates and the key contour corner points extracted from the binocular vision images are substituted into the three-dimensional coordinate adaptive formula;the least squares method is used to obtain the internal and external parameters, and linear optimization method is used to optimize the internal and external parameters ; the final internal and external parameters are obtained. According to the pixel information of the ac- quired images, the depth value of the coal bin coal level is determined. In order to verify the accuracy of the new method,the RMS measurement errors of the overall parameters were compared with those obtained using the Zhang Zhengyou＇ calibration algorithm. The proposed method does not require knowing any camera parameters, and the solution process is simple and convenient. The experiment results show that the new method has high precision, can meet the requirements of coal bin coal level detection and achieve the automization of the coal level calibration process.