Abstract
The use of Machine vision systems has become more widespread in manufacturing processes for the purposes of quality control inspection, and product identification and sorting. Typical Machine Vision applications need to run in real time (30 frames per second), and as a result most of the existing systems are built from hardware to meet this speed requirement. There is currently no single processor that is reasonably priced and fast enough to provide real time performance on Machine Vision applications. This paper describes a Transputer based system that employs different architectures and algorithms to achieve real time processing speeds for some Machine Vision applications. The paper discusses the differences between sequential and parallel architectures, and the way the unique abilities of the Transputers are utilized to create a flexible system that provides the best performance for a variety of applications. The areas of Machine Vision discussed are Image Acquisition, Image Enhancement, Feature Extraction and Image Interpretation. Image Acquisition and interpretation are discussed briefly, with an in depth discussion of the algorithms and architecture needed to optimize Image Enhancement and Feature Extraction on a Transputer based system.
Enhancement of ARFI-VTI Elastography Images in Order to Preliminary Rapid Screening of Benign and Malignant Breast Tumors Using Multilayer Fractional-Order Machine Vision Classifier