Height Measurement Method For Standing Wood Based On Monocular Vision

In forest resource surveys,tree height is one of the important indicators to assess the quality of standing trees and the growth of forest trees.Hence,how to quickly and conveniently obtain the height information of standing trees is of great importance.Nevertheless,the current height measurement methods for standing wood still face the shortcomings like low degree of automation,time-consuming and labor-intensiveness,as well as complicated measuring equipment.In face of the above problems,in the paper,the author proposes a height measurement method for standing wood based on monocular vision,which is a height measurement method for the standing trees waiting to be measured based on imaging model inside the camera under monocular vision of the smartphone.Main research contents of the paper: firstly,to take the monocular camera and built-in high-precision gyroscope of the mobile phone as tools for collecting data,make use of the close-range photogrammetry and machine vision technology,and segment and locate the smallest external contour of the standing image based on Graph Cut algorithm,to obtain parameters within the camera and non-linear distortion parameters to help get more accurate pixel values of feature points.Secondly,to set up the model to calculate the distance with the target standing tree and calculate the depth information of standing tree image.To study and establish the space geometry model including camera shooting angle,the ordinate pixel value of the object and the actual imaging angle parameter of the object.To calculate the distance between the target standing tree and the camera in different directions according to the principle of stereo imaging system.Thirdly,to establish camera built-in property parameter truth table,and vertical height and depth distance model according to the parameters calibration of the camera.To study the spatial geometry relationship between the tree factor to be measured and the actual image on the premise of camera built-in parameters.To study the conversion relation between the actual scale and pixel unit from world coordinate to camera coordinate and to image coordinate system under different constraints.Fourthly,on the basis of the height measurement method of a single standing tree,the depth information extraction of multiple standing trees under monocular vision is constructed by geometric similarity method,so as to deduce the height measurement method of multiple standing trees.By verification,it’s shown that The relative error range of the measurement method of single standing tree is-0.029 ～0.338 m,and the maximum relative error of the measurement of standing trees is 4.22%,with the minimum relative error 0.30%,among which,as for the height measurement method,only 13.64% of the errors are larger than 6%.As for the measurement method of multiple standing trees,the maximum relative error is 8.33% and minimum relative error is 0.29%,among which maximum relative deviation is 4.36% and minimum relative deviation is 0.93%.The innovative parts of the research: firstly,to set up the vertical height measurement model by using geometric similarity inside the imaging systems of smartphones and cameras;secondly,while obtaining the image information,the geometric similarity method based on monocular vision can effectively reduce the amount of calculation to obtain depth information;thirdly,based on the height measurement method of single standing tree,and by positioning and segmentation of standing tree images,the depth information and height information of multiple trees to be measured can be obtained,thus realizing the measurement of heights of multiple trees in a single image.