July 20, 2022
Rising inflation and a volatile workforce environment have led many companies to turn to advanced technologies such as robots and machine vision to automate processes and increase productivity, while also cutting costs and driving revenue. For one major sweet potato exporter in Israel, this meant finding a system capable of automatically picking and placing sweet potatoes.
Initially, the company had a machine that sorted sweet potatoes by size on a conveyor. It brought the potatoes to different output stations, where at least four employees at each station would pick potatoes and place them into boxes. The company hired Assatec Robotics to automate the process, but picking sweet potatoes — even manually — presents challenges because of the variety of shapes and sizes. After Assatec encountered issues early on, when testing different robotic grippers, it turned to Soft Robotics’ mGrip solution — an IP69K-rated, food-safe gripper that uses an advanced material science–based approach.
The mGrip’s patented elastomeric plastic grippers feature hollow fingers with flexible holds that are pneumatically controllable at low pressure, allowing robots to gently grasp objects such as sweet potatoes without adjustments to the variables of the overall system. In the new setup, sweet potatoes move down the conveyor belt and pass under an NSIX CVK5 3D camera positioned above the input conveyor. The camera captures 3D images of the potatoes at a maximum depth resolution of 1280 x 720 at up to 30 fps. Running on an industrial PC with a NVIDIA GeForce GTX 1650 GPU, NSIX Vision Keys software guides a FANUC M-3iA/6S delta robot equipped with soft grippers to pick the potatoes. To optimize where the robot places the potatoes, Assatec developed a custom volume-fitting algorithm within Vision Keys software.
Two additional 3D cameras positioned above the boxing area capture images of each box being packed, and the software analyzes the images and decides where to place the next potato. Once the robot places a potato, the system starts the process of picking another potato and finding a spot for it in the box. When a box is fully packed, the system takes another 3D image, ensuring that no potatoes are above the top edge of the box. If any potatoes are sticking out, a vibratory mechanism under the box spreads the products out to even them, another job previously done by employees. The new system can do the work of up to four people, with an entire cycle of guided pick-and-place of each sweet potato taking just 1.2 seconds.
In addition to handling sweet potatoes, the new system can handle nearly any other type of fruit or vegetable, including oranges and pineapples, opening new possibilities in automation.