A new motion-estimation algorithm based on morphological correlation

A novel algorithm is proposed for efficient estimation of two-dimensional planar motion. This approach is based on the idea of morphological correlation first introduced for shape representation. The advantage of this algorithm is that it can be implemented in real time, using high-speed morphological filters and memory. Most importantly, this technique is capable of estimating the speed of relatively fast-moving objects, the maximum speed being dependent on the size of the object. Promising results have been obtained through simulated experiments. This approach is inherently more efficient than other feature-based approaches if frame-to-frame correlation is maintained. This is the situation in most cases except for extremely fast-moving objects