Passive measurement method of tree diameter at breast height using a smartphone

Abstract

The accurate measurement of diameter at breast height (DBH) is essential to forest resources inventory, management, and carbon cycle modeling. Monocular vision passive measurement method based on a smartphone is a technique that allows for automatic, portable, and mobile measurement of DBH. Here, we propose a method for measuring the DBH of multiple trees from a single image taken by a smartphone camera, using machine vision and close-range photogrammetry technology. First, we present a visual segmentation approach based on an improved frequency-tuned saliency algorithm, which allows us to extract the trunk contour. Then, we establish a depth extraction model to calculate the depth of each tree. To calculate the DBH of the trees in the image, we establish an adaptive feature coordinate system for DBH measurement of multiple trees, which helps us to study the conversion relationship between the coordinate systems and establish the tree DBH measurement model. Furthermore, the tree DBH measurement model is established based on the rigid motion law between each space coordinate systems; we first establishes the mapping relationship between the pixel coordinate system and the object space coordinate system. Then, according to the depth information, we can obtain the coordinate of the DBH measurement position in the image plane coordinate system, and then derive the diameter at breast height from the image. Experiment results showed that the calculated DBH have a strong relationship with the truth values (RMSE = 0.217 cm). The method yields an average relative error of 2.32% at distances of 2–10 m. For trees with DBH values less than 20 cm, the absolute error is <0.237 cm; for larger trees, the relative error is less than 1.272%. These errors fall within the accuracy requirements of the continuous inventory of Chinese national forest resources.