Tree height measurement based on image processing with 3-points correction

This paper concerns the problem of tree height measurement based on image processing by 3-points correction. In the traditional method, two marker points were used, one was set on the root of the tree and the other was set one meter high from the root. A tripod was used to fix the camera for taking photos of the tree and the photo was made perpendicular to the ground, which have many restrictions in practice. In this paper, three marker points were used and they were set in a line along the tree trunk with same distance. After the coordinates of the three marker points and the top point of the tree were extracted according to their color features, the image was corrected by three marker points. Then the tree height can be worked out using triangle similarity theory. The experimental results show that the relative measurement error of tree height is much smaller than the traditional method.

I. Introduction

Tree height is one of the most important tree measuring factors in forest inventory. It not only reflects the forest productivity, but also can be used to determine the standing volume and volume growth rate. So tree height is often measured in forest resources survey and other occasions. Currently, some hypsometers are often used in measuring tree height, such as christen hypsometer, Blume-Lessis hypsometer, and so on. With the development of technology, the electronic theodolite, GPS and total station have been introduced to the forest survey [1].