Technion researchers have developed innovative technology for automated stress monitoring in agricultural crops. Early detection of water and heat stress is crucial for agricultural producers, as the reduction in humidity is reflected in limited stomatal conductance, resulting in reduced growth and, ultimately, premature plant death. The development was led by the folks at GIP, the Geometric Image Processing Laboratory, at the Henry and Marilyn Taub College of Computer Science. 

Technion researchers Research Assistant Alon Zvirin, GIP Lab Head Professor Ron Kimmel and Chief Engineer Yaron Honen have developed smart technology for monitoring and predicting crop stress and crop segmentation. leaves. In the context of the former, Zvirin explains: “The detection of drought stress allows saving the plant, allows the identification of diseases and the prediction of crop yield quantities, all of which are crucial information for the grower.” Using color photography, thermal imaging, and deep learning, the researchers were able to predict stress and new leaf development with great success; in a trial of the technology on banana seedlings, an impressive level of prediction of over 90% accuracy was achieved. In the context of the latter, leaf segmentation, the researchers achieved unprecedented results in the identification of Arabidopsis and tobacco leaves by applying deep learning. To train the system on a large number of samples, the research team developed a vast database containing images of artificial leaves, and then also tested the technology on other crops: avocado, banana, cucumber, and maize.

Young researchers

According to Zvirin, “We included young researchers who were just starting out in the process of technological development. They came up with great ideas and did a great job. Two of them are also listed as lead authors of the articles: Dmitri Kuznichov, who will soon complete his master's degree under the supervision of Prof. Irad Yavneh and Prof. Ron Kimmel, and Sagi Levanon, a graduate of the Psagot Excellence Program, who has started to study his second degree at the Faculty ”. The article on stress detection was published in the European Conference on Computer Vision, ECCV, and the article on segmentation was published in the Conference on Computer Vision and Pattern Recognition, CVPR.