Night vision obstacle detection and avoidance based on Bio-Inspired Vision Sensors

The rise of autonomous vehicles relies on collision avoidance systems. Usually, they employ vision sensors. Yet, detecting obstacles during the night is a challenging task with traditional cameras. A recent study on suggests using event-based cameras for this task. Besides other advantages, they can perform in all lighting conditions.

Image courtesy of Pexels.

A proposed obstacle detection algorithm consists of four parts. Firstly, background activity noise is removed. Then, objects are detected using a slicing algorithm based on event accumulation. An event corner detection algorithm lets to estimate the depth of 2D features.

Finally, an asynchronous adaptive collision avoidance algorithm plans the trajectory depending on observed objects. The evaluation shows that the suggested method surpasses traditional cameras. It can, for instance, detect a running person during night’s time while the conventional methods cannot.

Moving towards autonomy, unmanned vehicles rely heavily on state-of-the-art collision avoidance systems (CAS). However, the detection of obstacles especially during night-time is still a challenging task since the lighting conditions are not sufficient for traditional cameras to function properly. Therefore, we exploit the powerful attributes of event-based cameras to perform obstacle detection in low lighting conditions. Event cameras trigger events asynchronously at high output temporal rate with high dynamic range of up to 120 dB. The algorithm filters background activity noise and extracts objects using robust Hough transform technique. The depth of each detected object is computed by triangulating 2D features extracted utilising LC-Harris. Finally, asynchronous adaptive collision avoidance (AACA) algorithm is applied for effective avoidance. Qualitative evaluation is compared using event-camera and traditional camera.