Vision-Based Pipe Monitoring Robot for Crack Detection Using Canny Edge Detection Method as an Image Processing Technique

摘要

Piping setup is very important to ensure the safety and eligibility of the piping system before applied in industry. One of the techniques to facilitate perfect piping setup is by employing pipe monitoring robot. Pipe monitoring robot is designed in this research to monitor cracks or any other defects occur inside a pipe. This automatic monitoring is conducted by the application of image processing with canny edge detection. Canny edge detection method detects the edges or lines of any cracks inside the pipe and processes them to create differences in image, therefore only the cracks can be shown and finally, those cracks can be well analyzed. Canny edge detection has 5 processing techniques that are smoothing, finding gradients, non-maximum suppression, double thresholding, and edge tracking by hysteresis. In this research, the experiment was conducted by letting a robot monitoring new pipe and detecting cracks. Two cracks samples were taken and analyzed. The results show that the best value for smoothing is 10 and 5 for thresholding in getting not too blurred or to sharp result.