Bruise Detection Using Complexity for Oorin Apples

The complexity of binary image for apples as normal, bruised around centerbruised around edge, and bruised around both edge and center were analyzed. Amulti-threshold method (thresholds determined by area rates 0.3,0.4,...0.7 for complexity rates analysis) was used successfully for bruise classification. Leaving out apples with bruising around the edge, apples were classified 85.7% correctly.