Today's crop breeders are trying to boost yields while preparing plants to withstand severe weather and changing climates. To succeed, they must locate the genes for high-yielding, hardy traits in crop plants’ DNA, a task that is needed to accelerate breeding to meet global food demand.
Crop breeders run massive experiments comparing thousands of different cultivars, or varieties, of crops over hundreds of acres and measure key traits, like plant emergence or height, by hand. The task is expensive, time-consuming, inaccurate, and ultimately impossible — a team can only manually measure a fraction of plants in a field.
The lack of automation for measuring plant traits is a bottleneck to progress, but it remains difficult to make robotic systems that can count plants autonomously: the fields are vast, the data can be noisy (unlike benchmark datasets), and the robot has to stay within the tight rows in the challenging under-canopy environment.
A 13”-wide, 24-pound, four-wheeled robot was developed that is transportable, compact, and autonomous. The robot, called TerraSentia, captures each plant from top to bottom using a suite of sensors (cameras), algorithms, and deep learning. The robot transmits the data in real time to the operator's phone or laptop.
It is lightweight enough to traverse a field without seriously damaging crops, and monitors plant health by looking at things like growth rate and coloration. Its sensors are also designed to be flexible and customizable to suit the needs of the breeder and growers.
One challenge is that plants aren't equally spaced, so just assuming that a single plant is in the camera frame is not good enough. A method was developed that uses the camera motion to adjust to varying inter-plant spacing, which has led to the robust system for counting plants in different fields, with different and varying spacing, and at different speeds. Using a transfer learning method, TerraSentia was taught to count corn plants with just 300 images.
Currently, TerraSentia can cover an 80-acre field in a day; its “scale neutral” technology enables the use of multiple robots (“swarms”) to cover more land.