基于视觉技术的林木生长状态无损测量方法

为提高林木生长状态测量的准确性，克服传统测量方法的不足。从视差处理的角度出发，将图像处理技术与视觉理论有机结合，根据不同时间点采集到的树木图像信息，判断出一段时间内树木生长状态的变化情况。试验过程中，将待采集图像的树木上标出红色的矩形信息点，并利用双摄像机针对标出的特征点进行采集，然后将2幅树木图像进行对比研究, 分别计算出在一定时间间隔内树木上信息点空间信息，从而确定该段时间内信息点的位置变化。试验结果表明，针对标定的信息点，传统测量方法在高度和粗度分别增长5.63 mm和5.75 mm，二者数值相近，与林木生长实际不符；而基于视觉技术的测量，高度和粗度分别增长2.4 mm和1.5 mm，高度变化是粗度变化的1.6倍，与林木的实际增长过程是一致的。所以，基于视觉技术的测量方法，能够很好地实现对树木的无损测量研究，判断出树木的生长状态变化情况。

Nondestructive Measuring Methods of Tree Growth State Based on Vision Technology

Departing from the angle of disparity processing,the image processing technology and the theory of binocular vision were combined,and according to tree image information collected at different time points,judgment was made about the change of tree growth state within a period of time.Meanwhile,the traditional measuring method was applied to measure the depth change and changes in horizontal and vertical directions of trees,thus changes in calibration points of trees were judged within a given period.In the experiment,red rectangular information points on the tree were marked and two cameras were used to capture feature points,then the color information of that of the images collected was extracted,according to which the information of the calibration points to be studied was separated and calculated with the pixel center coordinate replacing the calibration points.And at last,the two tree images were compared to calculate the spatial information points on the trees in a certain time interval,so the position change of the time could be determined.The traditional measuring method mainly employs the measuring tools such as meter rule and caliber rule to realize the measuring of the distance between marked information points,and the data processing is relatively simple.But because the center of the calibration points is to be determined artificially,the deviation is not small.Besides,the method has been greatly influenced by the environment,so it is difficult to realize precise measuring.The horizontal increment and vertical increment of this method are very small,which is not consistent with trees growth pattern of the height grow fast and roughness length grow slowly.From the results,the visual measuring method which combines the image processing and the binocular vision can be more precise and accurate in data processing,and the experimental results measured are more consistent with the actual growth laws of tree growth than the traditional one.In addition,it perfectly achieves nondestructive measurement for tree growth,and judge the change of growth process of trees.