Canadian innovation field testing for robotic broccoli harvesting performed in Quebec
The future is now, as Industry 4.0 in agriculture melds with robot broccoli pickers.
Developed by Lapalme Groupe Conception Mécanique of Varennes, Quebec, the Sami 4.0 is an intelligent multifunctional agri system—a robot—that can perform the work of traditional hand pickers of vegetables utilizing a vision system and robotic arms linked to artificial intelligence. By adding a mobile app, farmers will even be able to manage a harvest on their smart phone.
It can work at the same tempo as a human harvester, except cut more because of the multiple robotic harvesting arms.
Founded by company president Éric Lapalme, he conceived the idea after a visit to his family farm in the summer of 2018, while watching workers harvest broccoli.
“I began to look at the pickers with my eyes as an engineer and to dissect the eyes, intelligence and arms process necessary to carry out the act of picking,” explained Lapalme, an engineer by trade. “At that time, I was working on the automation of a factory and the idea came to me to bring robots to the fields.”
Simon Belanger, the Business Development Director for Lapalme Agtech added: “On different projects, he had already developed advanced vision mechanics and robotic arms and realized he could develop a technology to fill the labour gap.”
Lapalme was surprised to discover, after checking with the patent office, that he was the first to conceive of such a concept.
“In the summer of 2020,” noted Belanger, “we wanted to de-risk the technology, so we created a very preliminary prototype to test the vision and the robotic arm. The results were conclusive, and we decided to invest our money and go for it.”
Belanger explained that the company received $1.1-million from the Quebec government as well as $500,000 from the NRC (Canada), and created its Lapalme Agtech subsidiary. “We then worked hard to have our RPC 4.0 demonstrator at the Expo Champs trade show in St-Liboire and were the highlight of the show. We’ve since been in the field finalizing the programming configurations and gaining exceptional results.”
Guided by 2D and 3D camera systems, Sami 4.0 uses its vision system to detect, position and qualify the broccoli. Following the detection, an analysis is made by artificial intelligence, which is sent to the robotic arms, telling it to harvest the crop—or not—based on its maturity/ripeness. A waterjet system using 50,000 psi is applied to cut the broccoli.
Using a tablet, such as an iPad, a farmer can choose the size of the broccoli to harvest. “For example, if a five-, six-, or seven-inch broccoli is opted for, the robotic arms will only harvest that size,” stated Belanger. “On the other hand, the cameras will detect and count the other sizes so that at the end of the day the farmer will have a complete inventory of the field transmitted to their CRM System.
“This will allow the farmer to increase productivity and to have a better planning of the future harvest,” he added.
The vision system has been developed in partnership with both the Quebec-based INO (Institut national d'optique) and the CRVI (Centre de robotique et de vision industrielle).
The Sami 4.0 utilizes artificial intelligence and was taught by engineers to recognize the vegetable broccoli, as well as how to make the required movement to harvest it. To that end, more than 500,000 photographs of vegetables were taken in its environment and in all possible weather conditions. About 50,000 of those images were selected and labeled for the robot to reference—but, said the company, the process has to be repeated for each vegetable type that can be picked by the robotic system, something that was performed per producer requests in 2020.
Initially field tested in 2020 with broccoli, Lapalme was encouraged. “The trial worked very well, and the results are conclusive. Since April (of 2020), our team of engineers has been working on setting up a demonstrator project including four robots.”
In September of 2021, the automated broccoli picking system was officially unveiled in field tests at two farms, in Montérégie and Lanaudière, Quebec.
“For the tests, we used four robotic arms in dynamic with the tractor moving forward,” related Belanger. “The speed of the tractor was the same as with the human harvesters working behind it—but we (Sami 4.0) managed to have a faster tempo.”
Benefits of the SAMI 4.0 robotic system, include:
• Increased productivity;
• Increased crop management;
• Fills the void of the picking labour shortage;
• Performance report provided;
• Improved inventory management for the customer;
• Return on Investment (ROI) in about three to four years.
Because of Covid-19 travel restrictions, farmers have experienced and may continue to experience a manual labour shortage. To counter that, a variable quantity of robotic arms can be affixed to Sami 4.0 allowing a single tractor operator to perform the work of 10 pickers.
“One robotic arm performs the work of one human harvester,” said Belanger, “But we can add many more arms.”
The price of the SAMI 4.0 will be around $1.5-million, depending on the number of robotic arms requested, and can cover up to 40 rows wide.
Right now, SAMI 4.0 has only acquired AI knowledge to harvest broccoli, but Lapalme Agtech is hopeful that future iterations of the robotic system will be able to harvest peppers, tomatoes, and other vegetables, as well as fruits such as strawberries and raspberries.
Belanger pointed out that the tested beta SAMI will visually look different when delivered to a customer, as well as configured differently dependant on the number of robotic arms requested by the farmer.
Lapalme Agtech has said that the design will vary according to the type of culture. For example, to harvest broccoli, cauliflower or cabbage, two long conveyors with a series of robots can be provided. But for fruit trees, it is imagined that a vertical machine with arched arms will be required to pick fruits like peaches, apples, pears or nectarines.
Other future itineration may be to use it for sugar bushes where the automation could be installed in front of an all-terrain vehicle, so the vision system could recognize the maple via cardinal points and by old incisions so it can cut at the correct height on the trunk. Ultimately, it could be used to execute other tasks such as weeding, insect and disease detection.
“I like to say that the SAMI 4.0 is like a cellphone, and to do more than make a phone call, you need apps,” said Belanger. “The SAMI 4.0 is the same principle. We have just developed the broccoli app, and soon we will create other apps for harvesting cabbage, lettuce, asparagus and more, plus one for weeding.”
At this point in time, SAMI 4.0 has a proven track record for picking visible vegetables. “As SAMI’s technology is based on an application principle, the possibilities are endless. The production will be personalized for the needs of the individual farmer,” added Lapalme.
Tested and now deemed ready, Lapalme Agtech has begun to take orders on the Sami robotic harvesting system for delivery in the Spring of 2022.
“Everywhere you look, automation is gaining traction and is boosting the competitiveness of businesses. Today marks the arrival of agriculture into the industry 4.0 era,” summed up Lapalme.
This article was featured in the December 2021 Precision Agriculture Digital Digest — view it here.