TimberSeg（TimberSeg 1.0）

在各种环境和条件下，成堆的木材中的单个原木的分割掩模。图像是从操作员的视角拍摄的。
TimberSeg 1.0数据集由220图像组成，显示加拿大各种环境和条件下的原木。图像用每个日志实例的分割掩码以及相应的边界框和类标签进行密集注释。该数据集旨在实现自主林业代理，因此它包含从操作员的角度来看的近2500个原木实例。图像是在森林，路边附近，伐木场和装满木材的拖车上方拍摄的。考虑到抓握的角度，对原木进行了注释，这意味着仅对桩上方且可访问的原木进行了分段。

Segmentation masks for individual logs within timber piles under various environments and conditions. Images are captured from the operator's perspective.

The TimberSeg 1.0 dataset consists of 220 images showing logs in diverse environments and conditions across Canada. The images are densely annotated with segmentation masks for each log instance, along with corresponding bounding boxes and class labels. This dataset is designed for autonomous forestry AI agents, thus containing nearly 2500 log instances from the operator's viewpoint. The images were taken in forests, near roadsides, at lumber yards, and above timber-loaded trailers. Logs were annotated considering the gripping angle, meaning that only logs situated above the pile and accessible are segmented.

搜集汇总

背景与挑战

背景概述

TimberSeg 1.0是一个包含220张图像的数据集，主要用于在多种环境和条件下对木材堆中的单个原木进行分割掩模标注，支持自主林业代理的开发。数据集包含近2500个原木实例，标注了分割掩码、边界框和类标签，图像从操作员视角拍摄。