PERBANDINGAN METODE SEGMENTASI OTOMATIS REGION OF INTEREST DAN K-MEAN CLUSTERING PADA APLIKASI DETEKSI KUALITAS DAGING SAPI BERBASIS MOBILE

The importance of the beef quality detection plays an essential role in coping with one of problems related to the distribution of poor beef quality. The beef quality is highly determined by a number of parameters such as size, texture, colour features and odor. Today, to determine the beef quality is conducted by observing the colour or form. As a consequence, this method still has many weaknesses such as in the human assessment that might be still subjective and inconsistent. This research aims making an application to detect the beef quality by using a number of phases of image processing in detail and by comparing the accuracy of two segmentation methods being used. The application was made using the operational system of Android integrated with SDK Android, library OpenCV and Eclipse. Image processing phase consisted of the pre-image processing of greyish level, histogram equalization, segmentation of Region of Interest, segmentation of k-mean clustering, and analysis on the texture by extracting the features of each digital beef images. The final phase of this research was to measure the higher level of accuracy from two segmentation methods. The proposed method could provide the accuracy of k-mean segmentation around 80%; while, the segmentation of ROI provided the accuracy at 90%