Features in Automotive

[sangar]

Hi All,

I have looked into some of the Automotive features that are available in the car...They are

1. Collision warning that helps the driver to keep a safe distance from the front vehicle...

2. Object detection, specifically, whether it is a pedestrian crossing or Animal crossing

3. Median/divider detection

4. Traffic sign board detection

Other than collision warning which needs LIDAR technology to tell the object distance, all the above require an image sensor with a camera to capture an image and do some processing in ISP...But, I wonder that how the system has the knowledge to tell the things...I mean, for example, in the object detection feature, how it tells the object is a pedestrian is crossing (or) an animal crossing...

Is it a kind of machine learning?... Please suggest


Thanks,
Sangar

[Scrts]

Correct, Computer Vision & Machine Learning. Lots of it. Petabytes of data.

[ale500]

Object detection is also done with RADARs for automatic cruise control and for lane change assistance, blind stop detection, cross traffic detection.

[sangar]

Hi,

I found another interesting feature, Pothole detection...But I am not sure how it works...Is it same capturing an image of the road ahead and analyse (again machine learning) it to see potholes are there and tells the type of potholes (Or) any specific sensor which tells potholes...

[sangar]

Hi,

I have the following question,

Traffic sign detection, Humb detection, Pothole detection and signboard detection: all these features exploit an image captured by a single camera which sits on the front mirror...(Or)multiple cameras with different fps have to be in place for the above features...Please let me know...And suggest any source to know the hardware being used in Automotive...

Thanks,
Sangar