I, as part of a project, have decided to implement a small scale version of automated collision avoidance system. I chose to go with Image based object detection. After initial research, I figured out that Arduino can't handle IP (alone), so RPI was the way to go. Having fixed this, I stumbled upon OpenCV. Now my plan of action is something like this.
I plan to buy a 2 wheel drive motor car (link given at the end) and use it as the building block. Use the RPI for controlling the vehicle as well as for Image Processing (along with the RPI Camera Accessory) to have a real time operation.For an object, I hope to put a black colour thing in the car's path and then hope to program it to "start" taking action when at a distance of 50 cm, with beeps/LED flash and then eventually brake the motors, bringing the car to a stop. My end goal would be to stop the car somewhere around 10 cm to 15 cm from the object.
I have been reading up on this and figured that RPI with OpenCV should do the job, however being not so experienced with IP and RPI, I am not sure if a simple Black object detection is "easy" for a RPI or not or how complex it is. Based on the first answer given, I am assuming that one RPI should do both the jobs - controlling the car and IP. My question is that Can this system (one RPI) deal/handle with a general object detection - something not black - any shape or colour? OR other way round - Given one RPI and the car, what Type of images can I use for IP, considering its capability? I am guessing Black object would help keep things simple, however not sure. Need advice. (No Machine learning algos are to be employed).
This question is preventing me from going full-throttle. I would genuinely appreciate any help.
Link for the Car Chassis - https://www.amazon.de/gp/product/B072ZWRM2B/ref=s9u_simh_gw_i1?ie=UTF8&pd_rd_i=B072ZWRM2B&pd_rd_r=4b652f7f-eb46-11e7-ad04-0b97a8a2142a&pd_rd_w=EJbrw&pd_rd_wg=Bnw5r&pf_rd_m=A3JWKAKR8XB7XF&pf_rd_s=&pf_rd_r=EQ2MK485GXAZKDYYYG1N&pf_rd_t=36701&pf_rd_p=c210947d-c955-4398-98aa-d1dc27e614f1&pf_rd_i=desktop