Lukas, the robot that removes weeds

A mobile agricultural robot named Lukas weeds fields in Sweden. With its infrared cameras and a computer running some specialized software, Lukas can recognize the difference between crops and weeds and remove automatically the weeds.

A mobile agricultural robot named Lukas weeds fields in Sweden, according to innovations-report, Germany. With its infrared cameras and a computer running some specialized software, Lukas can recognize the difference between crops and weeds and remove automatically the weeds. It could be used for all vegetables that are grown in rows, like lettuce, cauliflower, and carrots, or sugar beets, for which it has originally been designed. This robot is not ready yet to enter the market, but it should greatly benefit ecological growers by reducing costs and by eliminating the need to find people willing to do this work.

Let's look at the problem that eco-friendly farmers are facing today.

In ecological cultivation, weeding is performed manually, entailing not only economic burdens for many growers but also logistic ones -it’s hard to find people willing to do this work, and many growers therefore have to limit their production.

This is why Björn Åstrand, from Halmstad University in Sweden, has been working since 1998 on a robot using a vision-based system to automatically make its way around a field and remove weeds both between cultivation rows and in the rows, between plants.

Below is a picture of Lukas following rows in a sugar-beet field (Credit: Halmstad University). This image has been extracted from this video (1 minute and 18 seconds, 17.8 MB).

Lukas following rows in a sugar-beet field

And on this one, you can see Lukas removing weeds in a sugar-beet field (Credit: Halmstad University). This picture comes from this other video (27 seconds, 6.31 MB).

Lukas removing weeds

How does Lukas work?

The robot functions with the aid of computerized image processing. An infrared camera is installed on the robot to read the rows. The images are processed using a specially developed computer program that in turn steers the robot’s wheels and weeding tool.
Within the rows, the robot distinguishes between crop and weeds with the aid of another camera, which takes color images, and a program that analyzes the color and form of the plants. This method works extremely well under certain conditions, but the system is susceptible to differences in the appearance of crops. The appearance can differ considerably within a field and can be influenced by factors such as rain, wind, and diseases.

With these constraints, Lukas is not completely ready to be marketed. But it could help farmers in a few years. The number of jobs in manual weed control would be reduced, but higher educational level jobs would be created to handle the robots -- and to build them.

This research was the basis for the PhD thesis of Björn Åstrand at Chalmers University of Technology in Göteborg. This thesis was presented on September 8, 2005 under the name "Vision-based Perception for Mechatronic Weed Control." Here is a link to the abstract which does a good summary of this agricultural mobile robot.

For those of you who read Swedish, Lukas was also mentioned in a news release from Halmstad University on September 9, 2005, "Robot som rensar ogräs."

Finally, you can find earlier mentions of this eco-friendly robot in two previous reports. The first one is named "The Best Ideas Environmental Innovation 2002" (PDF format, 20 pages, 843 KB). One of these best ideas was named "Robot with eyes performs weed control mechanically" and appears on page 7 of this report.

The other report has been published by PricewaterhouseCoopers in and is called "Examples of eco efficient innovations" (PDF format, 52 pages, 1.29 MB). Lukas is described on pages 21 and 22.

Sources: innovations-report, Germany, September 16, 2005; and various web sites

You'll find related stories by following the links below.

Newsletters

You have been successfully signed up. To sign up for more newsletters or to manage your account, visit the Newsletter Subscription Center.
See All
See All