Dolotron uses AI vision to safely zap garden weeds

20 May 2019382 Views

Share on FacebookTweet about this on TwitterShare on LinkedInShare on Google+Pin on PinterestShare on RedditEmail this to someone

Dubravko Sršan. Image: Dolotron

Share on FacebookTweet about this on TwitterShare on LinkedInShare on Google+Pin on PinterestShare on RedditEmail this to someone

Our start-up of the week is Dolotron, a Croatian tech firm taking part in the Intel Movidius Edge AI Incubator at Talent Garden in Dublin.

“Dolotron uses state-of-the-art engineering technologies from the traditional machine-vision market to deliver vision solutions to new markets and business fields,” explained Dolotron CEO Dubravko Sršan.

Dolotron is an engineering start-up from Croatia that specialises in software design, prototyping and quality assurance for the latest vision hardware.

‘Our approach is to use AI vision system to localise weeds and use very precise spraying of non-toxic, non-selective chemicals directly on weed leaves’
– DUBRAVKO SRŠAN

One of 10 start-ups to present at the recent Intel Movidius Edge AI Incubator open day at Talent Garden in Dublin, Dolotron is working with the Intel Movidius Myriad platform to drive better product performance with lower power consumption, while enabling a smaller overall product footprint. Its neural compute engine also enables a dedicated hardware accelerator for deep neural network interfaces.

“Expertise in building digital cameras, LED illumination, embedded processing units and vision algorithms offer us capabilities to design and build vision products that would be otherwise built together from different standard vision components. The advantages of this approach are low-cost, low-power, small size and tight integration of different functional blocks,” Sršan said.

The market

“Our goal is to deliver affordable technology to solve daily problems,” Sršan explained.

“Beside lawn mowing, killing lawn weed is one of the most labour-intensive lawn care methods. For most, spraying special chemicals is not an option. Those chemicals are toxic to humans and animals. Nobody wants his child or pet to get in touch with these chemicals.”

The founder

Sršan received his MSc in computer science from the University of Zagreb’s Faculty of Electrical Engineering and Computing in 1999. His career started in Germany as a computer system designer and analyst for QUISS GmbH, Munich.

In 2000, he founded the firm Smartek Vision to build industrial digital cameras and LED controllers. The company was sold in 2012. Since then, he has been devoted to Dolotron and contributes his more than 18 years of experience and knowledge in electronics, 3D modelling, firmware and software programming, and machine-vision algorithms to the team.

“I have always had a passion for robots and other applications where real objects are manipulated in the real world,” he said.

The technology

“Our approach is to use AI vision system to localise weeds and use very precise spraying of non-toxic, non-selective chemicals directly on weed leaves.

“The solution is low-cost, low-power, AI-based, embedded vision system based on the Intel Movidius VPU (visual processor unit). Specially designed LED illumination with carefully designed optical set-up delivers images that make it possible to distinguish weed from lawn in any outdoor lighting condition.

“The ultimate goal is to reduce or completely abandon the use of toxic chemicals in agriculture weed-killing applications,” Sršan explained.

Seeds of growth

Sršan said that Dolotron’s bread and butter business is the OEM machine-vision components market.

“We are a fully self-financed company. Revenues from the components market allows us to make long and costly development of products we would need external financing for.”

He has a global ambition for Dolotron. “After having completed the development of weed-killing products, we aim to go global, but we will definitely need help. We do not have global sales and marketing capabilities. In addition, the production capabilities have to be further increased.”

The biggest challenge therefore is getting the right people together and motivating them.

“Financing the business and putting a lot of personal engagement into the company is also a big challenge. The founder’s family has to be the biggest supporter of every start-up, otherwise it is difficult to manage family and business together,” said Sršan.

Being from Croatia, Dolotron’s insight into the Irish start-up scene is based more on recent experience. “Irish start-ups are often on the cutting edge of technology. That makes it possible to get high-quality funding and have higher added-value products,” he said.

His advice to fellow founders comes as follows: “Be patient and persistent because self-starting is all about putting a lot of personal effort in the company without any guarantee of success.

“Be open for networking with other companies. Good business contacts are most important for start-ups.

“Do not try to make everything perfect. Try to do everything as good as possible.

“Make products that you and your friends would like to buy.”

Want stories like this and more direct to your inbox? Sign up for Tech Trends, Silicon Republic’s weekly digest of need-to-know tech news.

John Kennedy is an award-winning technology journalist who served as editor of Siliconrepublic.com for 17 years.

editorial@siliconrepublic.com