Lasers can help blind to detect hazards

Thursday, 14 July 2016
Elaine Wong, associate professor and ARC Future Fellow at the University of Melbourne, is looking to change that with technology that uses lasers to identify non-protruding hazards.

In Australia, around 357,000 people are blind or have low vision, according to Vision Australia.

For those affected, living independently can be a challenge, particularly when it comes to detecting hazards when outside the home.

Elaine Wong, associate professor and ARC Future Fellow at the University of Melbourne, is looking to change that with technology that uses lasers to identify non-protruding hazards.

Wong has firsthand experience with how blindness can affect individuals and their families – her 8-year-old son was born with congenital blindness.

Following his birth, she was in contact with Vision Australia and Guide Dogs, two of the biggest organisations for the blind and people with low sight in Australia.

Wong was communicating with the organisations about a capstone project to detect obstacles for final-year students in the Masters of Engineering. Vision Australia told her it needed a device to detect non-protruding hazards, such as dips in the road and pavement, descending stairs and curbs, and potholes, which can be difficult for people to detect with the white cane.

“The situation is worse for people who are also dependent on ambulatory aids, for example, wheelchairs or require walking frames, so these people can’t efficiently use a cane,” Wong said.

Wong spent time trying to work out how to create a device that could be hands-free, with the team, which includes Aravinda Rao, Jayavardhana Gubbi, Marimuthu Palaniswami, and both Vision Australia and Guide Dogs Australia, eventually coming up with an optical laser-based solution to ‘decipher’ the ground in front of a person.

“It’s a structured laser with a pattern that you cast in front of the wheelchair or the walking frame, depending on what sort of ambulatory aid that you use,” Wong said.

“Depending on the deviation of this pattern, you can tell whether there is a depression, or a pothole, or descending stair or curb in front of you.”

Laser work

The laser works by scanning the ground around 1m in front of the user, and capturing a video frame to obtain a position.

It was one of the most challenging aspects of developing the device – the team was limited to two off-the-shelf lasers in order to keep costs low – a grid pattern with horizontal and vertical lines and a cross-hair pattern.

During tests, the team went through the normal motions for someone in a wheelchair and captured hazards using a GoPro and laser. They then took the information back to the lab to carry out post processing, with the team developing their own image processing using machine learning and position algorithms.

The team also visually recorded hazards their device captured to compare what they saw with what the device said.

During tests, the team discovered that the grid pattern achieved much more accurate results due to a higher resolution.

“Basically the more lines there are – the more intersections – then the higher the deviation is from the original pattern, so you’re able to detect hazards more efficiently,” Wong said.

Future work

The team has several aims for the device – to make it as low cost, user-friendly and small as possible. Ideally, it could be placed on any ambulatory aid, such as a wheelchair, walking aid or even a white cane.

The team is now working on miniaturising the device, integrating the laser and a 2D video camera, and refining the type of laser the device uses.

“I’m keen to try different types of structured patterns to see which will be the most accurate,” Wong said.

“The laser we’re using now is actually visible light – red light – and it works very well at night. What we’re trying to do is to replicate this function for the day.”

This will involve looking at different types of wavelengths.

The team is also working on how users should be alerted about possible hazards. For example, whether by auditory alert telling them there is a hazard or through vibrations.

This will be decided after further discussions with Vision Australia and Guide Dogs Australia to find out what would be most user friendly option.

“We do know from talking to the client that people don’t like tactile feedback – vibrations – because it takes time to learn when it comes to how to interpret what vibrations mean. So maybe the best way to go is to just announce that there is a hazard,” Wong said.

The product has been developed with the help of a $50,000 seed grant from the Ian Potter Foundation in around 2014. But Wong said it has been hard to get traction with further funding.

“Disability engineering is actually quite hard to get funding for,” she said.

But Wong has made it her personal mission to increase awareness around those who are blind or have low vision because she’s seen firsthand how the stigma around people who have disabilities can affect them.

“Kids with disabilities are basically written off from the get-go because they are a ‘burden’ to society,” she said.

“Collectively, I think we should keep reminding people and have success stories. To me it’s really important to have success stories. We need to remind people that vision impairment is just one aspect. They can contribute to society.”