The HAWKwood Database

Abstract

We present a database consisting of wood pile images, which can be used as a benchmark to evaluate the performance of wood pile detection and surveying algorithms. We distinguish six database cate- gories which can be used for different types of algorithms. Images of real and synthetic scenes are provided, which consist of 7655 images divided into 354 data sets. Depending on the category the data sets either include ground truth data or forestry specific measurements with which algorithms may be compared.