Intelligent edge detector based on multiple edge maps

Abstract

An intelligent edge detection method is proposed. The method is based on the use of pattern recognition and machine learning techniques to combine the outputs of multiple edge detection algorithms. In this way, the limitations of the individual edge detectors can be overcome and performance enhancement is achieved. Two widely used classification algorithms, the Naive Bayes Classifier and the Multi-layer Perceptron, were selected for the learning task. The proposed system was evaluated on artificial and real images. A simple class labeling system based on the output of all edge detectors is suggested to provide controllability between detection sensitivity and noise resistance. Principal Component Analysis was performed to reduce computational burden and improve detection accuracy. The method is shown to achieve a practical compromise between detection sensitivity, computational complexity, and noise immunity.

Removal of Gaussian noise on the image edges using the Prewitt operator and threshold function technical

“…It is defined as the process of locating the boundaries of objects or textures showed on the image. The positions of these boundaries are defined critical in the process of image enhancement, recognition, restoration, and compression [3,4].…”