Visualization of Geometrical Defects of Radiators using Edge Detection Techniques

Abstract: Quality of products is needed to be inspected before they are used in the industries. A radiator is a heat exchanger composed of layers of tubes and fins bonded together known as the core of the heat exchanger. Any damages to the fins of the radiator, reduces the heat dissipation capacity of the core. The traditional inspection system of the products classify defects only based on their gray value, which is far from satisfactory because of its low productivity, low reliability and poor economy. This paper proposes an efficient approach for the detection of defects using Edge detection methods based on canny edge detector. From the pre-processed image, defect areas are located by specifying the region of interest. The total matched percentage is calculated with the number of white and black points by comparing the images.

Keywords: Pre-processing technique, Image enhancement, edge sharpening, edge detection filters.