Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

How to address a hole detection "problem"

Hi. How to solve following: 1. Display a live stream video on a RPI with a 5.8 Ghz reciever connected to the USB port. In a PC the reciever is recognized as a USB Camera and there you can use VLC. When using the VLC in the RPI it does not recognize the reciever as an USB connected camera. Suggestions? 2. Assuming that issue 1 is solved. How to a) show the live stream on a display and b) save an image set to an time interval, lets say every 3 sec. This saved image will then be analysed for changes in comparisson with the first image taken. If no change is found then this image will be discarded and the first one is still "template". If a change is found then that image will be saved as a new "template" for next image until a new changed is detetcted then that image will be the new template for next image and so on. Number of images to be saved will be defined from operation to operation. 3. Assuming that 2. works then how to proceed to analyse the change between 2 images? One idea I have is to make the image in grayscale and then rasterize the image in lets say 256 lines horizontally and then 256 lines vertically and use gray scale distribution to analyse the change. Or is there simpler ways to detect changes. The expected changes between the saved images will be "holes" in different shapes and diameter whereas the non changed areas will be intact. I have tried the blobb script but that works only if you have very clear and defined "holes". 4. Assuming that 3 is solved then - how to highlight the changes with a circle or similar with one color for each change and another color for previous change.

As you already have figured out I am a very very newbe at this but I am slowly learning and thought that I might try first to see if the "solution" is already out there :)

I am of cource open for other solutions that my "raster" idea.