Research On Pills Counting Algorithm Based On Image Processing

With the development of the information technology, Image processing technology is becoming more and more important and powerful. Nowadays in the medical field, the existing pills counter based on the mechanical principles can not meet the demand of the medical staff in the hospital. In this thesis, we propose an pills counting algorithm based on scanned image processing. This kind of non-touching counting mode is quiet, will not damage the pills, which will have a bright future.The pills counting algorithm based on scanned image processing include 3 steps: image segmentation, post-processing and couting. There are many methods for color image segmentation. The current color image segmentation algorithms have discussed the color space and the mainstream segmenting algorithms, which contains Thresholding, Edge Detection, Region growing and FCM. Then, we compare these segmentation algorithms. At last, we point out a image segmentation algorithm based on the statistical classification theory. With this segmentation algorithm, we can get better segmentation results for the experimental image features. The image post-processing algorithms based on erode algorithm optimizes the edge of the image, which can be regarded as a filter. The basic computing of mathematical morphology is also introduced in this part. Pills counting is based on connected region labeling with the pixel labeling algorithm. When the algorithm runs to complete, the numbers of the pills has also been computed.According to the experiment, the proposed algorithm is robust and has a good result even if the background of the drug board is reflective and shaded. The precision of the algorithm is very high.