Semi-Automatic Detection and Tracking of Growing Mushrooms on Image Sequences

Simon, Gyula and Vakulya, Gergely and Tarsoly, Sándor and Galambos, Péter (2022) Semi-Automatic Detection and Tracking of Growing Mushrooms on Image Sequences. In: AIS 2022 - 17th International Symposium on Applied Informatics and Related Areas - Proceedings. Obuda University, Székesfehérvár, pp. 17-21. ISBN 9789634493020

Preview

Text gomba_AIS_2022.pdf Download (978kB) | Preview

Abstract

The efficient management of autonomous mushroom production plants requires the model of growth rate of mushrooms. Photos of the plants are used as input for the growth models, which then predict the development of individual mushrooms. Recently machine learning techniques have been successfully applied to create such models. For the machine learning systems, however, large number of training samples are required. The training samples include photos of the plant and also ground truth markers indicating the position and size of the mushrooms on the photo. In this paper an image processing system is introduced, which is able to create good quality ground truth from sequences of images of the plant. The proposed system can automatically detect the mushroom positions and sizes on each of the pictures, but also allows user intervention to minimize the number of detection errors.

Actions (login required)

Edit Item