Research On 3D Reconstruction Technology Of Trunk Internal Defects Based On Radar Scanning

Trunk defects(void,decay,looseness)are common problems in old trees and big trees,especially in old trees.The internal defects of the trunk will affect the stability of the tree,resulting in the decline of wind resistance,flood resistance and lodging resistance,which is easy to cause safety risks.When the defect is serious,it will threaten the safety of life and property and cause heavy economic losses.For ancient trees,because of their own non-replicability and important value,this kind of loss will be more huge.Therefore,it is very important to detect the internal defects of the tree trunk.Because the interior of the trunk is relatively hidden,the traditional naked eye observation method can not directly find the potential defects inside the trunk,so it is necessary to use non-destructive testing technology.At present,the result of defect detection using TRU tree radar nondestructive testing technology is generally two-dimensional plane image.In analysis and observation,the information contained in two-dimensional sectional images is very limited for people to understand the shape and location of internal defects in the trunk,so it can not directly reflect the information of internal defects in the trunk.In view of this,this study takes the cedar(Cedrus deodara)about 40 years old as the research object,focusing on the two-dimensional image of defects in the trunk,defect image information extraction and three-dimensional model construction and other aspects of research,and finally verified by an example.The realization of three-dimensional reconstruction of defects can overcome the limitations of two-dimensional images and reflect the shape and location information of internal defects of tree trunks stereoscopically and intuitively,so as to understand the severity and development trend of tree trunk defects and master the law of tree trunk defects.it is of great significance for further tree risk assessment,tree stability protection and scientific maintenance management.The main conclusions of this study are as follows:1.The average method,the maximum method and the weighted average method are used to convert the gray level of the two-dimensional image of color defects,the results show that the image converted by the weighted average method has suitable light and shade and clearer outline;When 3 × 3,5 × 5 and 7 × 7 windows are used for Gaussian and median filtering,the filtering effects of the two filtering methods become worse with the increase of the window,and the filtering effect of 3 × 3 window is the best.In the 3 × 3window,the mean square error of Gaussian filtering and median filtering algorithm is 16.92 and 7.11 respectively,the peak signal-to-noise ratio is 35.84 and 39.61,and the signal-to-noise ratio is 33.36 and 37.13 respectively.The results show that the effect of median filter is better than that of Gaussian filter;Using linear interpolation and matching interpolation to carry out the second interpolation on the basis of the first interpolation can effectively reduce the gap between sectional images and avoid the "ladder phenomenon" in the subsequent reconstruction process.In the evaluation of interpolation effect,the running time of linear interpolation and matching interpolation is 2.34 s and 3.94 s respectively,and the gray inequality points are 36686.2 and 38898 respectively.The mean square errors are204.17 and 216.01 respectively.The evaluation results show that the linear interpolation method is effective;Finally,the linear grayscale transformation is used to enhance the image,the results show that the overall contrast between the defective part and the benign part is obvious when k > 1,d=0.2.Among the five edge detection operators(Roberts,Prewitt,Sobel,Lo G,Canny)used in the edge contour extraction of defect two-dimensional image,the trunk contour and defect contour extracted by Canny operator are the most complete,only a small amount of contour is missing,and the effect is better than other operators.The results of image erosion and expansion processing show that the combination of erosion and expansion closed operation processing on the basis of Canny edge detection can effectively fill the missing contours.3.The surface rendering algorithm runs fast,but it still has limitations in retaining the internal defect information of the tree trunk,and the physical model can only be restored to a certain extent;the data processed by volume rendering is the whole data set,and the reconstruction speed is slow,but it is three-dimensional and intuitive,which enhances the spatial sense of defects.At the same time,the introduction of transparency in the process of volume rendering can greatly improve the visualization level of internal defects without further cutting.The verification of an example based on volume rendering also shows the reliability of this method.