Determination of Sugar Apple Ripeness via Image Processing Using Convolutional Neural Network
Rhys B. Sanchez, Jose Angelo C. Esteves, Noel B. Linsangan
Abstract
One type of fruit that is seasonally available in the Philippines is the sugar apple which is known as "Atis." Specifically, no technological advancements regarding sugar apples ripeness classification are created. Sugar apples are manually separated based on their ripeness when harvested. This research focuses on using image processing through CNN to determine the ripeness of sugar apples, which will benefit the sugar apple fruit farmers and harvesters. The researchers created a machine prototype which is able to capture the image of a sugar apple and determine the ripeness classification in the image detected. The researchers found that the use of image processing in determining the ripeness of the sugar apple is adequate and accurate based on the datasets that the machine is trained to recognize. Looking at the gathered results of the images when compared to the manual inspection of the harvesters in each image taken, the researchers were able to achieve an accuracy of 86.84% in determining the ripeness level of sugar apples using convolutional neural network.