Active Information Gathering for Crop Health Monitoring

Kumar Robotics Lab and the Kagan group are collaborating to co-design an optical sensor-detector system for measuring crop health. Novel optical agricultural sensors in development by our collaborators shall be placed on crop surfaces in their early growth stages. These sensors will directly interface crop surfaces, and continually respond to crop conditions (pH, moisture, etc.), changing color with changes in plant conditions.

To extract meaningful information from these sensors, we are developing a new ground robot with on-board RGB camera, hyperspectral camera, IMU, GPS, for localizing these leaf sensors, and extracting actionable information for farmers or mitigation robots.

There are many challenges that come with this multi-disciplinary project that combines material science, computer vision, and robotics. Some of these challenges include detection under variable lighting conditions; detection of translucent leaf sensors of a small form-factor; and (3) precise robot control.

This on-going work is part of the NSF IoT4Ag ERC. This work got featured in an article and front cover of the Penn Engineering magazine: article.