|
|
 |
|
|
|
 |
| Project: | Robotic Mass Removal of Citrus Fruits | | Client: | United States Department of Agriculture |
Energid Technologies is developing an automated citrus harvesting system that will reduce cost and improve production. It addresses the single largest expense in citrus farming using many low-cost flexible members to remove fruit from trees in grove rows such as those shown below.
Currently, almost all oranges and grapefruit are picked by hand - over 50,000 people are needed to pick the over 50 billion oranges representing the US crop each year. That's more than 1 trillion oranges in a generation, all removed from trees by hand, outdoors, often from ladders, often in adverse conditions.
Energid has developed a unique and ingenious way to harvest automatically using low-cost robotic mechanisms, with guidance through color cameras and computer vision algorithms. The approach leverages over $10M in technology development for robotics and machine vision undertaken at Energid. It is only through this advanced (patented and protected) technology that our robotic approach to citrus harvesting is feasible.
With Energid's method, a machine driven by one person can pick over 10 tons of fruit per hour, reducing cost and improving harvesting predictability and performance. Energid has developed and tested a first prototype, and is refining a second prototype now. One key component is the identification of the location of citrus using multiple cameras. An example composite image of a scene using four machine -vision cameras is shown below.
Our system segments the multiple images to identify the citrus fruits in each, as shown in the image below, then correlates these locations to determine the 3D location of the fruit for harvesting.
|
|
|
|
