Automatic Defect Segmentation of ‘ Jonagold ’ 1 Apples on Multi - Spectral Images : A 2 Comparative Study

10 In this work, several thresholding and classification-based techniques were employed 11 for pixel-wise segmentation of surface defects of ‘Jonagold’ apples. Observations 12 showed that segmentation by supervised classifiers was more accurate than the 13 rest. Also, average of class-specific recognition errors was more reliable than error 14 measures based on defect size or global recognition. Segmentation accuracy im15 proved when pixels were represented as a neighborhood. Effect of down-sampling 16 on segmentation accuracy and computation times showed that multi-layer percep17 trons were the best. Russet was the most difficult defect to segment, whereas flesh 18 damage the least. The proposed method was much more precise on healthy fruit. 19