At the recent 2023 Agricultural and Food Tech Startup Expo, a number of cutting-edge agricultural machines utilizing artificial intelligence technology were showcased.

△Various applications supporting agricultural fields, such as growth monitoring, pest monitoring, and automatic harvesting, were being developed based on vision recognition and were preparing to enter the market. Although they had in common that they were all based on artificial intelligence, the direction and focus of each company's use of it to develop products seemed slightly different.
In product development, technological innovation is important, but developing with a focus on the needs of consumers and the effectiveness and cost-effectiveness when applied in actual fields can be a factor in accelerating the product's market entry and market share.

Gogofarm CEO Ko Won-seok said, “Robots should not be viewed solely in terms of ‘function’ but also in terms of ‘price.’” He expressed his opinion that if the unit price of a product is set high due to technological advancements and excessive multi-functionality, it may be difficult to conquer the market, saying, “We should optimize the product specifications by considering the purchasing power of farmers to purchase agricultural machinery.”

Gogofarm's autonomous harvesting robot 'Robongi' has reduced its functions by only equipping it with a stopping function for obstacles, excluding the avoidance function that requires complex algorithms for autonomous driving along a designated path.

In addition, in terms of harvesting method, the method of directly holding the fruit and harvesting it can damage the surface of the fruit, so a method of holding the stem, cutting it, and then loading it was adopted. This is a development direction that reflects the position of the consumers and focuses on the process of maximizing the marketability of the fruit.

Metafarmers, which develops intelligent agricultural robots, is a new startup company founded by master's and doctoral researchers from the Department of Mechanical Engineering at Seoul National University, and took its first steps into the industry in 2022.

If we are to talk about the weaknesses of AI-based solutions, it is necessary to fine-tune the learned model to fit the on-site environment in order to provide customized solutions for each site, and AI personnel will inevitably be required in this customization process.

MetaFarmers has developed an AI-based farming software called 'tapFarmers' to provide more user-friendly fine-tuning functions. The vision sensor mounted on the harvesting robot precisely distinguishes the maturity of the crop, and the user decides which crop to harvest by clicking on it through the controller.

The data on crops harvested at this time will be relearned by Tapfarmers and optimized for the farm and crops within two weeks using adaptive artificial intelligence technology, a Metafarmers official explained. Semi-automated harvesting and fine-tuning are performed simultaneously, which also increases the functional accuracy of automated harvesting operations, resulting in a triple benefit.

Based on a 5,000-pyeong farm, two robots replace one worker to support harvest automation, and CTO Wonjae Yoon said, “We are receiving commercialization support through the Rural Development Administration’s 1 and 1 Sprout Business Program, and are conducting field verification in collaboration with the Iksan Haneulchae Agricultural Cooperative.”

There are many companies developing AI-based cutting-edge agricultural machinery, but there are not many technology companies that have implemented large-scale field applications and verifications and accumulated field operation data.

Iocrops operates a 12,000-pyeong smart farm in Miryang, Gyeongnam, and has developed its own AI-based smart farm robot, ‘Herma Scout,’ which is being used in actual fields.

Herma Scout walks around the farm, taking images and using AI vision analysis to analyze the maturity of crops and efficiently manage shipping times, etc. △It can monitor various crop growth information in real time, such as fruit size, quantity, maturity, and leaf area, and collect data by autonomously moving 24 hours a day on rails inside the greenhouse.

“We are currently developing a product that will also detect pests and diseases, automate harvesting, and even transport in the future,” said Yoo Jang-seon, marketing manager for Iocrops. “As of July, a new product has been released, and the first product with growth investigation functions has been delivered to a client.”