Abstract
A procedure for measuring the point-to-point local variance of a spatial random scene is discussed. The relations that are required between the second-order statistics of the random process and the parameters of the optical configuration to furnish accurate and stable variance estimates are established. The procedure is experimentally verified by using calibrated samples of the random checkerboard and the overlapping circular grain models. A modification of the procedure enables the generation of spatial masks that are useful in automatic image processing. The spatial masks are used as binary spatial gates to isolate certain classes of objects within a scene and reject all others. Typical masks are synthesized and their applications illustrated.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
