Segmentação e contagem de troncos de madeira utilizando deep learning e processamento de imagens
This work addresses a domain-specific problem for forestry or logistics by providing an incremental improvement in wood log counting using existing deep learning techniques.
The paper tackled the problem of counting wood logs in images by segmenting them from the background using a CGAN-based method and then counting with connected components, achieving over 89% segmentation accuracy and over 97% counting accuracy.
Counting objects in images is a pattern recognition problem that focuses on identifying an element to determine its incidence and is approached in the literature as Visual Object Counting (VOC). In this work, we propose a methodology to count wood logs. First, wood logs are segmented from the image background. This first segmentation step is obtained using the Pix2Pix framework that implements Conditional Generative Adversarial Networks (CGANs). Second, the clusters are counted using Connected Components. The average accuracy of the segmentation exceeds 89% while the average amount of wood logs identified based on total accounted is over 97%.