Researchers at the Geometric Image Processing (GIP) lab at the Taub Faculty of Computer Sciences at the Technion developed innovative technology that can automatically monitor stress in agricultural produce. Technion researchers Alon Zvirin (research assistant); Prof. Ron Kimmel, head of the GIP lab; and lab engineer Yaron Honen developed smart technology for monitoring and predicting stress in agricultural produce and for leaf segmentation.
Detecting drought stress helps save plants, diagnose diseases, and predict crop quantities, which is critical information for farmers.
The researchers used color and thermal photos and deep learning to successfully predict stress conditions and the development of new leaves. In their studies on leaf segmentation, the researchers achieved unprecedented results in the identification of Arabidopsis leaves and tobacco through deep learning. In order to teach the system, the group developed an enormous database of photos of artificial leaves and then tested their technology on other crops as well, including avocado, bananas, cucumbers, and corn.
Other young scholars involved in developing this technology were Dmitri Kuznetsov, an MA student supervised by Prof. Irad Yavne and Prof. Ron Kimmel; and Sagi Lebanon, a graduate of the Psagot excellence program who recently began studying for his MA. Their article on stress detection was published at the European Conference on Computer Vision (ECCV), and their article about segmentation was published at the CVPR conference.
Use of robots for construction and architectural production is a vision that is close to becoming a reality, and is considered one of the main trends of the next construction revolution. However, despite the developments in this field, architectural projects use construction methods that require human intervention for production and for the calculation and design of the various products.
In the past few years, robotics devices have been gradually filling the gap between the required level of sophistication and the level of performance at the site.
One of the teams working on bridging this gap comprises researchers from the Taub Faculty of Computer Science at the Technion – Prof. Miri Ben Chen, Dr. Kacper Pluta, and Michal Edelstein, and their colleague Prof. Amir Vaxman from Utrecht University. They developed an algorithm for architects that can find automatic solutions for robotic production of complex surfaces. They created a computational framework that worked well under laboratory conditions into which any complex, design can be input in order to receive a series of fragments that are ready for production, including double curvature surfaces. Their achievement was published in ACM Transactions on Graphics.
“It’s important to realize that industrial robotic production is not a technological caprice”, explains Prof. Ben Chen. “It has many advantages for sustainability as it conserves material, shortens production time, and reduces the environmental harm caused by construction. The algorithm we developed can take complex surfaces and divide them into small hexagonal segments in a manner that increases the mechanical advantages of the surface. The algorithm generates a three-dimensional plan for robotic production. Continued development of the computational system can yield optimal solutions”.