05. Vision and Self-Driving Cars

Sensors

Traditionally, RADAR is used for long-range detections, while cameras are used for rich sensory inputs.

When it comes to sensor configurations for SDCs, we have to note the details therein; LiDARs and RADARs are what are known as active sensors. That is, they sense the environment based on transmissions of energy.

Cameras on the other hand, are passive sensors; they can only sense the environment based on energy (in this case, photons) already in the scene.

These sensor details have serious repercussions with respect to the types of algorithms we end up using to analyze this data. Computer vision has many powerful tools, but part of good design had to do not only with knowing what to do, but also what not to do when we use sensor data from multiple sources.

Now, as we mentioned, computer vision is a very powerful tool, but it should not necessarily be associated with only camera images. It's possible to also construct LiDAR images with a LiDAR sensor, thereby giving you measured-depth alongside well as classified pixels. Next you'll see some LiDAR sensor output and resulting data!