RADAR, LIDAR AND CAMERA CALIBRATION
THE IMPORTANCE OF CAMERA AND LIDAR CALIBRATION
The quest towards fully autonomous driving has made significant strides over the past decade. As technology becomes more and more cutting edge, it is crucial to be familiar with some of this new equipment. If you drive any of the high-end new model cars, you may have noticed some unfamiliar terms on the car’s description panel. These include Electronic Power Suspension, Braking Assistance, Cruise Control, Lane Departure, and Traffic Sign Detection etc.
All these are referred to as advanced driver assistance systems whose main focus is to improve safety. The latest technology to hit the proverbial streets is the use of Camera-radar fusion to track objects. You can track to see if you’re in the proper lane, detect obstacles when your view is blurred, or even track moving object that may pose a collision risk.
The main problem here is most of these systems use multiple sensors located on different parts of your car. This means that although they may seem functional, on paper, they may be off by some degree if not properly calibrated.
WHY IS CALIBRATION NECESSARY?
You have to note that you are dealing with several different modalities. Let’s put this in perspective. The process of installing a front facing camera is pretty straight forward, but with Camera calibration, you are essentially dealing with equipment that maps out your surroundings as you drive, reverse, switch lanes, stop in traffic, park, and even braking.
This means you have multiple modalities working together simultaneously to ensure you’re driving as safely as possible.
These modalities include: Light Detection Ranging (LiDARs), multiple strategically placed cameras, RADARs, GPS systems, IMU and many more. For you to accurately utilise these instruments, they need to be properly fitted. Also their positioning and orientation needs to be known for the sensors to work in extrinsic unity.
SIMPLIFIED EXAMPLE OF THE CALIBRATION PROCEDURE
Camera-Radar calibration comes down to two main functions. Stationary and dynamic calibration. The multiple sensors are scanned with a specified diagnostic tool and then tested on a road with clear lane markings. The on-road test only lasts about 30 mins, at which point the designated scan tool indicates the success of the calibration process.
WHAT YOU GET?
Cameras were some of the first inclusions in the quest for autonomous driving. They provided high resolutions and high frame rates which in-turn improved driver vision during reversing, as well as obstacle identification. Fast forward to 2019, now you have the value of RADAR and Light Detection and Ranging Technologies that utilise infrared light to help map out your surroundings more effectively.
All these sensors put together help eliminate visual shortcomings that may ultimately result in collisions, and when used jointly, they become significantly more efficient.
BOTTOM LINE
Note that the cameras, the Radar and the LiDAR equipment all perform different tasks. The LiDARs provide a 360-degree sparse point cloud, while the cameras give you HD, High-resolution images. Put together, you get top-of-the-line object detection, recognition and more precise positional information that gives you a safer estimate when driving. It is for this reason that calibration becomes crucial.
