Question on indoor navigation and face tracking - RPLidar or RGB-D camera or both
My requirement is to do indoor navigation/obstacle avoidance and also face detection and tracking using a ground robot (not UAV). I know that I can use RPLidar for creating occupancy grid maps but for face detection, Iidars will not work and I will need to use an RGB-D camera for detection and tracking. Now, I do not want to buy both a Lidar (for SLAM and navigation) and a depth sensing camera (for face detection and tracking) since it will be expensive to buy both. My question is - can I create 2D maps using an RGB-D camera and will it be fine to use it for indoor purposes in natural light and darkness also (for face detection). Or a Lidar will outperform an RGB-D camera when it comes to navigation.
Asked by ritesh on 2019-06-18 07:41:02 UTC
Answers
you could use the depth image to laserscan node http://wiki.ros.org/depthimage_to_laserscan to generate lidar data for 2d mapping
Asked by nickw on 2019-06-18 14:11:15 UTC
Comments
If you want to choose a single sensor, using an RGBD camera is the way to go. You can check turtlebot2 family, for example, that uses this config.
The RGBD will allow for 3D perception, but you'll need more CPU consuming algorithms to map and localize with that sensor. A LIDAR in the other hand is usually cheaper, has a greater range, precision and framerate, but you'll only see a "slice" of the world.
It boils down to which kind of application you have in mind.
Asked by Procópio on 2020-02-18 03:33:58 UTC
Comments