We are using Halcon to do image analysis (IDS USB3 camera with telecentric lens). However, the Halcon keys we need to buy for every machines is a burden, so we want to move to OpenCV.
What we need to have at the end of the task
1. a simple C++ project (visual studio 2015, 2013 or 2010) which can work standalone (so we can feed it with camera screenshots to see how functions interact) - more info down.
2. a dll we can integrate in our project (visual studio 2015) so all functions can be called from our source code.
3. We need to have the source code of the project
- detection of the position of the notch (in pixels towards the bottom of the full picture) - see pic1
- detection of a reference dot (1mm radius) (in pixels towards the bottom of the full picture) - see pic2
- readout of a code - pic3
- we must have the possibility to have a second camera form open which shows a closeup of the region where the mouse is.
An online demo of the current capabilities can be given via [Removed by Freelancer.com Admin - please see Section 13 of our Terms and Conditions]. An overload of pictures can be given.
Questions regarding OpenCV.
- Can it analyze parts of camera screenshots instead of full images so analysis is speeded up? - Range of Interest can be set during startup of the product.
- Do I need a display window (like I'm using a Halcon window right now)? I must be able to draw crosses, lines, add text to it...
Not in this scope - If the project above is finished, it might be extended with additional features
- finding the borders of the stone at a give height (picture similar as pic1)
- finding the correct camera focus height by analyzing 4 to 6 images which all have a 100u difference
- taking pictures (similar as pic 1) every degree to make a 3D model of the stone. In this 3D model, points must be able to placed and lines must be able to drawn
Any comments to make the SW better are very welcome.