See Through the Eyes of a Robot with FPGA

Hello FPGA lovers! Today’s project focuses on a key aspect of artificial intelligence and robotics. Robotic Vision is something different from human perception and understanding this and coding for Vision with FPGA can be a great step towards building projects which depend on robotic perception and vision using FPGA.

The materials needed, apart from your FPGA board, are a 2.8” TFT screen, an OV7670 camera board with no FIFO and DuPont cables. The author has added links to all these components in step 1 from where you can buy them directly. Though the interconnections have not been shown in the form of an image by the author, the pin assignments (you may have to adapt this to your FPGA) of the FPGA board in interfacing it with the TFT screen and Camera have been mentioned in step3. The UCF (.ucf) file which contains more detailed interconnections can be obtained from the zip file in step 1.

The code has been broken down into modules and VHDL (.vhd) is the language used. The VHDL files can be found when you open the zip file and each module has a description given on the header. A detailed explanation and working of the code is given in step 4 along with a Top level block diagram.

Once you are done with the interconnections of the camera and TFT with the FPGA, you can upload the BIN file. The code has been designed to detect threshold values of primary Red, Blue and Green colors from an image scanned by the camera. If these threshold values are detected, the screen shows white spots in the corresponding areas and remains black in the other areas. This perspective can help you implementing advanced DSP modules when you build a robot which relies on vision.

Sayonara baby!

By Dennisv15