A Preliminary Method for Tracking In-Season Grapevine Cluster Closure Using Image Segmentation and Image Thresholding

Figure 3

An example of Pinot gris clusters for histogram shifting and image thresholding representing the difference in Otsu’s thresholding values between original and contrast images; (a–d) are images from June 29 to July 19 (E-L 27–32) where no difference was observed; (e) is an image of July 28 (E-L 33) where Otsu’s threshold value for original image captures a larger portion of pixel values (>80), which can be the berry pixels, whereas for contrast images, the threshold value mainly captures the tail of pixel values (<120) representing gaps between the berries.