MonoDBH: Monocular Tree DBH Estimation Dataset

MonoDBH dataset is a monocular computer-vision approach designed to estimate a tree’s Diameter at Breast Height (DBH) from a single RGB image acquired using a standard camera. In this framework, the image is captured at approximately breast height, conventionally 1.3 m above ground level, with a human hand placed in contact with the tree stem to provide an in-scene physical scale reference of known or calibratable dimensions. The method uses the visual information contained in the two-dimensional image to identify the trunk boundaries at the measurement height and infer the apparent stem width in pixel space. This pixel-domain width is then transformed into a real-world diameter estimate by referencing the scale object, thereby enabling metric reconstruction without requiring stereo imagery, LiDAR, or additional range-sensing hardware. By relying on a single monocular RGB image and a simple anthropometric reference, MonoDBH provides a low-cost and operationally practical solution for rapid field-based DBH estimation, especially in resource-constrained forest inventory settings.