Machine Vision Approach to Assessing Postharvest Quality of Cucumbers During Storage

A machine vision system was used to monitor color changes and presence of any damage in stored cucumber. Images of cucumber were acquired in a "soft box," which provided a highly diffuse lighting scene, ideal for observing visual changes in the skin of cucumber. A cucumber center pixel accumulation (CCPA) algorithm was used to select center pixels from grayscale images. All the center pixels from 400 images (each obtained by 0.9° rotation) were accumulated to obtain an image of 1280*400-pixel size, which corresponds to a whole cucumber surface. To monitor damage that progressed over time, absolute differential damage progression (ADDP) plots were made from accumulated grayscale images. For the ADDP plot, the blue (B) channel in RGB color space was found to be optimal in terms of interpreting the damage progression from the plot and the corresponding 3-D histograms. Acquired RGB images were transformed into L*a*b* and HSV color spaces. The color space that was the most sensitive overall, i.e., could capture most of the information about the day-to-day color changes of cucumber, was identified through a principal component analysis (PCA). According to the PCA, all individual components in the RGB color space were found to be suitable to obtain information about the external changes of cucumber. Overall, the machine vision approach was found suited as a non-destructive technique for monitoring the external quality of cucumber during storage.