CropWorld 2010: Precision mapping system for blackgrass shows promise

First year trials with a camera-based grassweed automated mapping system have picked up just under half of the blackgrass heads in mapped fields.

But importantly it had never failed to detect blackgrass in images where it was present, said Reading Universitys Alistair Murdoch, who led the DEFRA-funded one year project.

"We have some work to do on the programme to improve the detection rate, but at the end of the project we have proven the concept that you can identify blackgrass in a field of wheat."

The system used generally available image sensor technology mounted on to a spray boom or tractor cab, meaning the cost of the system shouldnt be prohibitive, he noted. "The image analysis is based on red, green, blue sensors that are converted to hue, saturation and value to help distinguish different shades of green.

"We can then use different algorithms (calculations) to take care of differences caused by conditions changing during the day."

The ultimate aim was to allow growers to map fields in one year and then use precision spray application techniques the next, he told delegates at CropWorld.

"Initial studies suggest that there is a correlation between uncontrolled weeds in one season and what comes up the following year."

The technology could then be used to help guide pre-emergence sprays accurately, he suggested.

A new four-year Technology Strategy Board project would attempt to develop the technology further, including into other crops and with other weeds at different growth stages, and also explore the possibility of mapping and spraying in the same season.

"Our projection for 2019 is for 6% of the arable land to be mapped in this way," he concluded.