Evaluation of germination rate of tomato seeds with autonomous image processing and artificial neural networks system
Standard seed tests are designed to evaluate germination rate under controlled conditions by manual counting with specially trained technicians, which is time-consuming and labour-intensive. Potential enhancement of seed testing methods can be achieved by applying visual techniques to collect sample data. This paper describes a computer vision system based on image acquisition with RGB camera, image processing and machine learning techniques, which was implemented for automatic assessment of germination rate of the tomato seeds (Solanum lycopersicum L.). The entire system was built using the open-source applications ImageJ, WEKA and their public Java classes and was linked by a specially developed code, so no expensive commercial software was required. After object detection proceeds by Image J, which outcomes 11 object features, artificial neural networks (ANN) were implemented and directly compared to manual counting on a sample of 700 seeds germinated in 28 Petri dishes (90×98×18 mm). The results indicated that the automated system was able to classify 95.4% of germinated tomato seeds correctly.
Stajnko, D., Rozman, Č. and Škrubej, U. (2021). Evaluation of germination rate of tomato seeds with autonomous image processing and artificial neural networks system. Acta Hortic. 1326, 303-310
Solanum lycopersicum L., ANN, WEKA, machine learning