Fast Color Straight Line Pattern Recognition in an Object Using High Speed Self Learning Devices

Abstract

Most of the man-made objects are having some straight lines with colors. A very high-speed object recognition method using color straight line patterns is carried out using a novel self-learning device (RKD - Rasiq Krishnakumar Device). Instead of that Artificial Neural Networks (ANN), RKD based networks are used for different steps in this pattern classification. The color and straight-line features are extracted by using high-speed color segmentation and fast efficient straight-line segment extraction methods using the RKD based systems. The training and the testing algorithms of the pattern classification are using RKD-based processing. The fast color features extraction method uses an array of RKD-based devices and the fast efficient straight-line segment extraction method employs an array of processing elements and a main control unit. Some fusion devices are used for a straight line with colors features. The area of interest and the area of line segments of a particular object are other features for improving the accuracy of object classification.