A Real-Time Multisensory Image Segmentation Algorithm with an Application to Visual and X-Ray Inspection

A new multisensory image segmentation algorithm is presented. In this algorithm, the images from different sensors are segmented in a sequential manner using curve evolution methods. There are no fusion rules involved and no controlling weights to adjust. It is effective in eliminating errors in single modality segmentation, and is fast enough for segmentation in real-time applications. The algorithm is applied to real-time fan bone detection in deboned poultry meat based on visual and x-ray images. Results show that the fusion-based inspection algorithm is efficient, accurate, and robust to registration errors.