Smart De-weeding system using Computer Vision & AI | AgTech Startup Ideas


This 1001 startup idea from YoStartups is to develop a Smart De-weeding system using Computer Vision & AI  for weed control in agriculture. By making the process more efficient, the solution will select the most effective and optimum quantity of herbicide limiting the over the use of herbicides that adversely affect our land, environment, and human health.


Market Definition for 1001 Startup Ideas – Smart De-weeding system using Computer Vision & AI

A study by experts with the Weed Science Society of America showed that the corn and soybean yields would be decreased by about 50 percent if weed control measures were not implemented in the United States and Canada. The seven-year study, led by Kansas State University professor Anita Dille, showed a total potential economic loss of $43 billion a year.

Talking about Australia, it is estimated that Australian farmers incur around $1.5 billion a year in weed control activities and a further $2.5 billion in lost production per year. The long-term damage of weeds to the environment is difficult to quantify, but in monetary terms, it is expected to be almost of the same scale as that of total damage to the agriculture industry.


Competition Analysis for 1001 Startup Ideas – Smart De-weeding system using Computer Vision & AI

Founded by Stanford alumni in 2011, Blue River Technology is using computer vision and robotics to devise ‘See & Spray’ machines that precisely observe and treat each plant in a field. The process dramatically reduces the quantity of herbicides and pesticide to be sprayed across the field by limiting the application of chemicals only to the targeted plants. The company has devised a modular robotic arm that can be attached to a farm tractor to perform the action.

Blue River Technology has raised over $30 million in total funding in 4 rounds from multiple investors. In Sep 2017, John Deere acquired the company at a valuation of $305 million.

Another similar project is QUT’s AgBot II prototype that is equipped with cameras, sensors, and software and is designed to work in an autonomous environment to navigate, detect, classify weeds and manage them either chemically or mechanically as well as apply fertilizer for site-specific crop management. AgBot II is currently being tested in the field to improve and refine its application in the agricultural fields.


Pain Point & Target Audience for 1001 Startup Ideas – Smart De-weeding system using Computer Vision & AI


Usage of pesticides and herbicides has always been an issue of debate among researchers and agriculture veterans. As per Pesticide Literature Review published by Ontario College of Family Physicians, it urges to limit the usage of pesticides as the study finds consistent links between pesticide exposure and serious illnesses such as cancer, reproductive problems, and neurological diseases. Various other reports, including one by the World Health Organisation, have suggested that some of the herbicides are carcinogenic to humans.

Today the practice is to spray the entire agricultural field and the plants with the same herbicide for weed control. This process is inefficient as there is a lot of wastage of pesticide which not only it increases the input costs of farming but also causes irreparable damage to the environment. The constant exposure to a large amount of pesticides also increases the plant’s resistance to specific herbicides.

So there is a need for more scientific and result oriented system which targets plants on their current health indicators and uses pesticides only in the required quantities and the areas that are affected. The target audience for this startup would be large-scale farm owners, food processing, and agriculture produce based companies managing their farmlands.


Value Proposition for 1001 Startup Ideas – Smart De-weeding system using Computer Vision & AI


To address the mentioned pain points in the previous section “Pain Point & Target Audience,” we propose to build a computer vision analytics and artificial intelligence based smart solution for weed control. Through computer vision, the system will identify the exact location and type of the weed, and then using artificial intelligence it will determine the most effective herbicide for treatment. The process will ensure the optimal use of herbicide at the targeted plants.

The startup can also take help of spectrometry cameras to predict the overall health of the crop/plants to provide early warning signals to the farm owners.


Business Model for 1001 Startup Ideas – Smart De-weeding system using Computer Vision & AI


The startup offering will include two major components, firstly the one-time installation of computer vision hardware, spectrometry cameras, and herbicide spray system including PID control systems.

Secondly, the startup will provide cloud-based analytic services on a monthly subscription model. It will include a centralized dashboard for reporting on the health of the crops and the different type of weeds.

All the computer vision data will be uploaded in batches to a centralized big data and machine-learning platform, which will learn from the historical trends of different weeds and the effectiveness of herbicides. Based on the outcomes of these learning, the system will synchronize with the customer’s hardware and software to deliver the right kind and quantity of herbicides for the identified weed.


Way to Market for 1001 Startup Ideas – Smart De-weeding system using Computer Vision & AI


All the hardware needed for the proposed solution is readily available in the market. The startup should partner with some of the vendors for providing and installing all the hardware.

The key differentiator and the proprietary tool for the startup would be its ability to handle a large amount of data on different kind weeds and its near perfect ability to identify the weed among the produce from the images. The startup would also need to take help of machine learning tools to learn and respond to the herbicide tolerance levels in the weeds and correspondingly identify the right type of chemical and its dosage.

Once the basic platform is developed, the startup should feed it with all the available online images of different kinds of weeds and their established herbicides. Once the system can identify the commonly available weeds and their respective herbicides, the startup should start trialing with 2-3 large farms in their country. It will not only help them in testing their weed identification process but also fine-tune the dosage and concentration of herbicides needed for different weeds. The start-up can also use services of the farmers to verify their findings and use their experience in the learning process.


Key Milestones for 1001 Startup Ideas – Smart De-weeding system using Computer Vision & AI


The startup should develop the data collection and machine-learning platform in first two months. Once the platform is ready, the startup should start feeding it with the images of different kind of weeds available online along with its preferred herbicide and dosage/concentration levels. The startup must complete feeding this information over next one month. The start-up should involve farmers at all levels of domain knowledge till the system reaches a near perfect ability to distinguish weeds and regulate the use of pesticides.

Post that the startup should tie up with the vendors for providing hardware and one-time installation, and run a beta trial of their solution with at least six large farms in the first year. During the period, the startup will near perfect its algorithm to predict any weed and its effective treatment.

Investment Needed for 1001 Startup Ideas – Smart De-weeding system using Computer Vision & AI

The startup won’t need huge upfront investments, as the hardware part will be completely outsourced to the partner vendors. The startup should look to raise about $100k from Agtech focused accelerators and angel investors for platform and machine learning algorithm development.


Team Capability for 1001 Startup Ideas – Smart De-weeding system using Computer Vision & AI


The team must include a plant scientist to lead the product development. The team must also include members with experience in image processing and machine learning tools. The team would need data scientists to design the algorithms which can train and supervise the learning process to improve the accuracy of the solution. The start-up should also engage and deploy various farmer groups to increase the accuracy of the solution in various geographies for different crops.


Investor/Expert Take for 1001 Startup Ideas – Smart De-weeding system using Computer Vision & AI


The proposed startup doesn’t necessarily need to develop any new technology, but to diligently combine the already existing technologies from computer vision analytics, AI, Instrumentation, IOT and cloud storage to formulate a unique solution to solve one of the biggest problems for the farmers.

The Agtech industry has been recently seeing strong interest from even the generalist accelerators or those from other industries have stepped in to bridge the early stage-funding gap. During the first half of this year, the self-dubbed “Accelerator VC” SOSV invested in 26 agtech startups through its IndieBio and Food-X accelerators in San Francisco, Ireland and New York.


This 1001 startup idea article is from Yostartups, YoStartups is a Pre Accelerator and it empowers entrepreneurs to propel their business ideas into successful ventures. Yostartups’ core mission is to take the message of entrepreneurship to 1 billion people globally by 2020.

In case you are looking at scaling or launching a venture on similar lines, you can apply for our virtual acceleration program Excelrate on our website,, Yostartups’ Excelrate program will help you in streamlining and structuring your startup idea. We have limited scholarships and discount packs for deserving startups, depending on the strength of your application, you may qualify for a discount.

*Any information given in the article does not constitute any warranty of marketability or fitness for a particular business idea for any market or industry segment. It is the readers’ sole responsibility to further research and assess the market opportunity.


Hey There!

If you have any question, send us an email and we'll get back to you, soon.

Start typing and press Enter to search