Machine Vision

I use standard algorithms for  image processing such as optimal thresholding and connected component labeling to implement  fast and robust computer vision systems based on digital cameras.

In this example a car's engine block is separated from the background and a minimal enclosing rectangle is computed to give a coarse estimation of  the objects location. A better estimate of the location is achieved by segmenting the engines valve seatings and precisely determining the centers. Using information from the CAD model one can establish the exact coordinates system of the part. Click on the images to see them at full resolution.

Home

Email