A small North Vancouver high technology firm is banking on increasing demands for productivity gains to drive sales of its innovative vision systems hardware and software solutions.

Established in 1994, BrainTech Inc. designs and develops adaptive pattern recognition technologies and products. Until 1997, it was primarily involved in research and development. To date, the company, which employs 20 people and shares support services with another small high tech firm, has only been earning about $100,000 annually. But it’s built up the software side of its business and is expecting its Odysee System to jump start sales into the $1-million range by the end of this year. “By 2000, it will be 10 times that. We want to have thousands of units out there,” says Charles Hooge, the company’s vice-president of research and development.

BrainTech, now publicly traded, has an e-commerce web site (www.bnti.com) that provides secure transactions, product demos, web services and live customer service.

Its expertise is in machine vision systems, a combination of sensors, lighting, computer hardware and software that captures and analyzes images that are compared to a defined standard. For example, if basketballs are moving down a conveyor and suddenly a tennis ball appears, the vision system would spot it and a quality control guy could bounce that ball to the Wimbledon shipment.

“The system is based on fuzzy logic. It’s not a new technology, but it’s new to the marketplace,” says Hooge. Similar to neural networks in its ability to learn, it recognizes objects based on physical characteristics rather than measurements, it learns faster, it doesn’t take as long to be trained and it can edit on the fly.

BrainTech is particularly excited about its Odysee Development Studio, version 3.0, which was introduced July 1. The component-based application, available for about $2,200, allows machine vision integrators, scientists and engineers to design systems without having to resort to typed in programming commands. They just connect hardware and software in a visual drag and drop environment. Features include a graphical user interface builder, on-line debugging and optimization, and Odyseewrap, which is a tool for integrating C/C++ programming language. Functions are represented by boxes the user connects to develop the application. This shortens the time between concept to prototype by about two-thirds.

BrainTech has partnered with many suppliers of development tool libraries, which are being incorporated into its products. Now there are more than 3,500 different functions available to the user.

Interestingly, Odysee is already creating exciting possibilities, thanks to a cooperative project with the National Research Council of Canada’s Integrated Manufacturing Technologies Institute (IMTI) in Vancouver. Funded by government, IMTI’s role is to undertake R&D that’s too risky for most companies to undertake and too applied for universities.

The idea behind the joint project was to go beyond passive sensing and create a robotic sorting system using BrainTech’s Development Studio. Kevin Stanley, an electrical engineer and technical officer at the Vancouver IMTI, already had a vision-guided robot thanks to a previous research project. BrainTech’s sorting application was added to the mix and they came up with a Robotic Object Classification System (ROCS) in a remarkably short six weeks.

“A sensor can tell you something’s not right about a piece on the assembly line, but it can’t do anything about it. An operator has to remove it. We wanted to add an automated arm to eliminate the need for an operator,” says Stanley.

In May, they took ROCS to the International Robots and Vision Show in Detroit, more to demonstrate BrainTech’s capabilities than to solicit commercial applications for the system. They had it sorting different shapes of plastic plumbing parts, which attracted some attention at the show.

The system set-up included a camera, connected to a PC, which was connected to the robot. BrainTech’s classifier took the plastic part’s signature, determined what kind of part it was, and told the robot which bin to put it in. It’s capable of operating at a speed of 10 parts per second, which Hooge contends is fast enough for 99% of applications.

Indeed, now IMTI and BrainTech are working on a major project for a Vancouver-based recycler that has a great deal of commercial potential. They’ll be putting together a prototype system for sorting plastics they’re calling a Sorting Precision Quality Recycler (SPQR).

“Ninety per cent of the plastic that gets collected in blue boxes doesn’t get recycled because the sorting machines are too expensive and human inspectors often can’t tell the difference between the different types,” says Hooge.

This represents a huge problem and a big opportunity. It turns out BrainTech’s BrainTron classifier is the only device capable of accurately identifying those six different plastics. For the recycler, it will mean doing the job for a third of the price.

Hooge estimates the potential market for this kind solution to be worth $1 billion. He says the prototype should be ready in about six months. But BrainTech can also take that application and adapt it to other industries. For a firm that’s in the recognition technologies business, having that kind of vision is a good thing.