An Effective Method for Extracting Capsule by Color Image Processing

In order to extract capsule defects and reduce the process time in the real-time capsule defects identification system, a novel method using color image processing has been presented. We take RGB color space as the image processing environment and build an algebra relationship among the R, G and B components to directly segment the capsule defects instead of transforming these components to HSI color space. The first step of this method is to remove the background and effectively extract the capsule from the CCD-captured image. Based on the color variation relationship between the defect regions and the surrounding ones, the capsule defects are further segmented. Experimental results show that the method can successfully extract capsules better than in HSI color space, meanwhile, it can greatly accelerate the process and finish extracting within 3.5 milliseconds in contrast with 29.3 milliseconds in HSI space. Using this method, we have effectively extracted the capsules as well as their defects. The method would be practically useful in capsule defect identification system.