InnoGrow

What it does

Innovative weed-killing device utilising computer vision and lasers for chemical-free weed elimination, offering farmers a sustainable solution that enhances crop yields, promotes environmental health, and supports healthier food resources.

Your inspiration

Growing up in a farming family, I faced the challenges posed by weeds, which consume significant time, resources, and money. After studying at university, I reflected on how to apply my knowledge to benefit my community. I came across a video showcasing similar technology in the U.S. or Europe, but the machine was large, expensive, and unsuitable for Asian agriculture. This inspired me to develop a new idea: creating a smaller, more accessible device tailored for farmers like us, focusing on sustainable and non-toxic weed control. My goal is to provide an effective solution that meets our specific needs.

How it works

The Innogrow Weed Control System effectively manages weeds while protecting crops. Here’s how it works: 1.⁠ ⁠Computer Vision: The system uses our trained machine learning model to analyze field images, accurately distinguishing between crops and weeds. This ensures that only unwanted plants are targeted. 2.⁠ ⁠Laser Activation: Once a weed is identified, a powerful laser is activated. This laser focuses on the weed, disrupting its cellular structure and eliminating it without harming nearby crops or soil. 3.⁠ ⁠Modular Design: Our technique allows this trained model to be integrated into a compact processing board, like a Raspberry Pi. This modular design makes the system versatile—small farms can use it as a handheld tool, while larger farms can mount it on tractors for efficient operation. In summary, Innogrow combines smart imaging and laser technology for effective, sustainable weed control, adaptable to all types of farming.

Design process

Our main goal is to create a laser-based weed-killing machine suitable for the Asian region. Given that our land is not always flat or large like those in Western countries, we needed a compact, adaptable solution for bumpy surfaces that remains affordable for farmers. Initially, we considered a compact rectangular box containing the necessary technology, which could be attached to various agricultural machinery. However, to accommodate the smaller farm sizes in Asia, we opted for a modular design. This allows farmers to connect multiple units, making it versatile for any scale—from gardens to larger farms. During prototype design, we realized a rectangular shape might hinder direct targeting of weeds beneath it. Consequently, we modified the design to resemble an arm extending outward for better precision. To reduce costs, we replaced the expensive computer systems used in larger machines with a Raspberry Pi, making our device more compact and affordable. For the prototype, we successfully developed an algorithm for weed detection. While we couldn’t use a real laser due to safety concerns, our research indicates that it can be easily integrated later, with only a slight increase in weight and battery size.

How it is different

The concept of modularity sets us apart, enhancing adaptability to various geographical conditions and allowing for easy scalability across different farm sizes. Unlike bulky machines, our design is suitable for smaller farms and uneven terrains. Current products are prohibitively expensive—costing millions of USD—making them inaccessible for local farmers. In contrast, we estimate our product will cost only $500. By utilizing a smaller computer instead of costly systems, we significantly reduce the price. Additionally, our proprietary algorithm and self-collected dataset uniquely target weeds found in Asia, particularly Thailand. This dataset is not commercially available, giving us a competitive edge in the region. Focusing specifically on the Asian market, we recognize that existing competitors cater to Western geographies, leaving this region largely untapped. We have consulted with local farmers to ensure our product effectively meets their challenges.

Future plans

Our next steps for the Innogrow Weed Control System focus on scaling up development and enhancing capabilities. We aim to collaborate with local farming communities to gather data, refining our machine learning model with insights into diverse agricultural practices and weed types. Additionally, we plan to expand our modular design to fit various farming equipment and environments. By continuously improving our technology, we hope to make sustainable weed control accessible to more farmers, boosting food productivity and promoting environmentally friendly practices in agriculture.

Awards

- First price in Zheng Da Cup Provincial Round - Second price in Zheng Da Cup National Round - First price in HKAE Pitch Competition (Hong Kong Academy of Engineering) - Winner price in InnowShow Award