基于计算机视觉的立木枝干直径自动测量方法

立木枝干直径是智能型立木整枝机器人工作的重要工作参数.该文提出了一种基于计算机视觉原理的立木枝干自动测量方法.该方法利用已知大小的标定尺简化无线性畸变CCD摄像机标定过程.首先，将标定尺与被测立木紧贴在一起并通过CCD摄像机获取它们的图像；其次，从图像中通过二次模板匹配方法检测标定尺并统计标定尺像素数；然后，计算出图像每一个像素代表的实际尺寸；第四，通过立木枝杈点检测分离出树干和树枝；最后，将统计立木枝干直径的像素数与每一个像素代表的实际尺寸相乘就可以很容易地计算立木枝干直径.该文测量了12组立木枝干的直径，平均绝对误差为0.67 cm，平均相对误差为1.9%.实验结果表明，计算机视觉的立木枝干直径自动测量方法能够比较精确地解决智能型立木整枝机器人中的立木枝干自动测量问题，但是该方法在进行复杂背景立木枝干直径测量时精度需要进一步提高. 

Computer vision based method of automatic measurement of trunk and branch diameters of standing trees

The diameters of trunks and branches of the standing tree are important working parameters of intelligent pruning robots.This paper presents a method based on computer vision to measure the diameters of trunks and branches of the standing trees.A gauge,whose size was known,was used to simplify the calibration of CCD camera without nonlinear aberrations.Firstly,the images of gauge and standing tree were acquired by CCD camera.Secondly,the template was detected by twice template matching,the number of pixels in the area of gauge in the image was counted.Thirdly,the actual size of a pixel in image was got by dividing the size of gauge by the number of pixels in the area of gauge.Fourthly,trunks and branches of standing tree were detected by the joints of them,and then the numbers of pixels of trunks and branches were counted.Finally,the actual diameters of trunks and the branches of standing tree were got by multiplying the numbers of pixel of trunks and branches by the actual size of a pixel.The diameters of 12 trunks and branches were got in the experiments,and the mean absolute error was 0.67 cm,while the mean relative error was 1.9%.The results of experiments show that the method presented in the paper can exactly get diameters of trunks and branches of standing trees.When the diameters of trunks and branches of standing trees in a complex background are measured by this method,the precision will be improved.