Here we introduce a system which detects crack on wall by using image processing. As image is susceptible to noise we used some image preprocessing steps to detect crack more accurately. System works on most image formats. System mostly focuses on intensity value. This is done for sake of accuracy. System removes all undesirable noise. To detect crack, image is binarized and holes are filled so that image is more clearer to detect cracks. All small insignificant blobs are removed. Using blob analysis methodology, we detect number of connected objects. Based on the connected components system detects whether image contains crack or not. System is able to detect deeper as well as minor cracks. System uses many image processing steps to detect the cracks. Once the crack is detected by the system, System applies bounding box technology to display the crack to user. Thus , this is an innovative approach to detect crack on wall. We used image preprocessing steps as well as crack detection method to get accurate result. The proposed system is able to detect deeper cracks with 80% success rate and minor cracks with 50-60% accuracy.
- Involves preprocessing steps as well as crack detection method to get accurate result
- Detects deeper as well as minor cracks.
- Fails to work properly on poor quality images.
- Reduced accuracy in shadowed or poor lighting walls.