frame to pixel stream block

Hello everyone, i want to use frame to pixel stream block in order realize a stereo vision system based on fpga, the the demo i got is using a video as a source, in my case i want to use a pair of images. my question is whether this block (frame to pixelstream, and pixelstream to frame) would work on images or not because it is not working with me. **frame : what is the meaning of this term in this block (a single image from a video stream or something else...) thank you in advance.

Answer 1

[Walter Roberson]

Yes, frame to pixelstream will work on images that have already been read in.

"frame" is a 2D or 3D array representing a single still image. As you go through a video, each time you read one frame (an image). A pixelstream is mostly just reshaping the array into a vector. However, pixel streams might also include information about the structure of the frame, that would allow the video to be recreated afterwards -- information about pixel timing, "backporch" and so on.

[khalil khalil]

thanks a lot

Answer 2

[Bharath Venkataraman]

Please take a look at the examples in vision HDL Toolbox. For example, the Edge Detection and Image Overlay example reads in a video which is then passed to the Frame To Pixels block.

For stereo vision, you can either

1. combine the two pixels into one stream (so two uint8 values could go together as a uint16 value)
2. combine the two streams into a vector of 2 (so input would be 2 channels, 1x2 uint8)
3. use two Frame To Pixels blocks, each getting the left/right side

[khalil khalil]

many thanks to you.