New Technology Offers To See Behind Corners

This post is also available in: עברית (Hebrew)

A new camera system for seeing objects behind obstacles may be just around the corner. Traditionally, such systems relied on radar technology, which are usually big, expensive, low-resolution, and limited to short distances. In recent years a switch to LIDAR (laser illuminated detection and ranging) and microwave technology has occurred. These systems are not without their problems. The images they produce may be satisfactory, but the time it takes to produce them – an hour or more – is not. “That’s not going to work if you want to know whether a car is coming around the corner,” says Genevieve Gariepy, a physics doctoral student at Heriot-Watt University in Edinburgh, UK.

Gariepy and colleagues have now designed a system that works around these limitations. It is similar in function to LIDAR, but works much faster. It can detect an object precisely and quickly, “so we can track the object’s motion in real time,” Gariepy says.

Gariepy and her team’s innovation is using a high-frequency firing laser and a highly sensitive camera. The laser fires bursts of pulses at the ground immediately behind the corner or obstacle. When the light hits the ground it scatters, some of it returning to the original target area after bouncing off from the object being tracked. A highly sensitive, super fast, SPAD (single-photon avalanche diode) camera captures the returning light. The camera can capture images at a rate of 20 billion frames per second, and each of the camera’s 32 pixels is sensitive enough to detect a single photon. The data is then analysed to create a reliable image of the goings on behind the obstacle.

In testing, the team used a 30 cm humanoid model moving at 2.8 cm a second. “We used that in this proof-of-concept demonstration just because it’s easier to do things on a small scale,” Gariepy says. “But of course we want to go to a bigger scale and detect real people or cars. So we’ve started to use me as an object to track.”

Currently, it takes about three seconds for the SPAD camera to acquire the necessary data, and another two for the calculations. The team hopes to bring the time needed down to one second, as well as scale up the application to be able to detect objects as large as a car.