10.46243/jst.2021.v6.i04.pp144-151 registered

We propose a computer vision device that utilizes a multi-spectral imaging sensor to detect external defects on orange citrus fruits. To begin, the proposed algorithm segments the orange fruit solely using the NIR portion of the captured Near-Infrared (NIR) and RGB images. Second, segmented RGB and NIR orange fruit images are pre- processed using certain adaptive pre-processing techniques. As a result, a thresholding technique is used to detect defects in the orange fruit's seven distinct color components. Finally, voters vote on whether or not the citrus fruit image is flawed based on the seven threshold color variable images. Utilities and parcel distribution companies have increased their performance over the last year. Online shopping provides many benefits to the postal and distribution industries. The seller's items are packaged in box-shaped cardboard boxes or wooden boxes in a variety of sizes. A contour-based shape representation algorithm is used to detect contoured objects. Contour is made up of fragments of edge or curve that represent geometric concepts. The object's size must be determined in order to determine its surface area. Dimensions are sometimes used to refer to an item's length, width, and height. The volume of the parcel box will be calculated using a multiplication programmed based on its length, width, and height. As a result, in a computer vision-based automated sorting system, contour-based object detection could be used to determine the volume of an object. The most critical characteristics for correct citrus classification and sorting are color and size.