Development of an Autonomous Ground Robot for Field High Throughput Phenotyping

Abstract

High throughput phenotyping holds promise of accelerating crop breeding programs and has gained great attention in the past several years. Although many platforms for data collection, including ground and aerial vehicles, have been developed to adapt to different crops and field setups, fully autonomous ground robots have not been readily available and widely used so far. In this study, we developed a ground robot that is capable of autonomously navigating in the field. We designed a sensor mounting frame that can be modified to mount different sensors for different data collection tasks. The robot was tested in a cotton breeding field and color images of the plants were collected. We developed an algorithm to count the cotton bolls using the color images. The preliminary results demonstrated that the system was capable of collecting data autonomously without human intervention and the collected data were useful to extract phenotypic traits.