A research team led by Washington State University has developed a robotic harvesting system capable of detecting and picking strawberries concealed beneath foliage, a challenge that has hindered previous automation efforts. The system integrates an artificial intelligence-based vision system, soft silicone grippers, and a fan mechanism designed to gently move leaves, thereby improving access to hidden fruit.
Laboratory and field tests showed that the robot identified strawberries with 80% accuracy and classified whether the fruit was obscured 93% of the time. The inclusion of the fan system significantly enhanced performance; without it, the robot picked 58% of visible strawberries, while the addition of the fan increased the success rate to 74%. However, the process remains relatively slow, requiring about 20 seconds per fruit on average.
The research, published in Computers and Electronics in Agriculture in July, was led by Zixuan He as part of his doctoral work at WSU’s Biological Systems Engineering Department. He is now a postdoctoral researcher at Aarhus University in Denmark. Co-authors include Manoj Karkee, a former WSU professor now at Cornell University, and Qin Zhang, professor emeritus and former director of WSU’s Center for Precision & Automated Agricultural Systems.
The project targets commercial viability in open-field strawberry production, which has proven difficult for automation due to the fruit’s delicate nature and frequent concealment under leaves. Previous robotic harvesting systems have typically focused on greenhouse or elevated “table-top” growing environments, where visibility is less obstructed.
The WSU-developed robot operates as a four-wheeled platform that traverses strawberry rows, using a 3D camera system to capture color and depth images. These are processed using a machine learning model to identify the location of each fruit. The robot then calculates how to deploy its grippers for harvesting, using the fan to displace leaves where necessary.
The global strawberry market, valued at approximately $20 billion, is expected to grow by 6% annually over the next decade. Labor shortages in the sector have intensified interest in automation, although researchers involved in the project caution that robotic systems are unlikely to replace human labor entirely in the near term. Instead, such technology may be used to augment manual harvesting, particularly in regions facing seasonal labor constraints. The researchers said that the approach developed for strawberries could potentially be adapted for other crops with similar harvesting challenges, such as grapes.
Photo credit Zixuan He
