Fruitylicious: Mobile application for fruit ripeness determination based on fruit image

Development of local fruit industry in Indonesia is very high, but less competitive than imported fruits. Produced fruit kinds in this country is very diverse, but the use of technology to support the production and distribution is still not widely applied. This makes the local fruit market less competitive than imported fruits that has largely been applying technology production support. In this research, we analyse the relationship between fruit digital image and sweetness level of it. Image processing method carried out to prepare a digital image that is ready to be processed in the matching stage. K-Nearest Neighbor method is used to match fruit digital image with its sweetness levels. Sweetness levels were measured using Brix degrees' units. Matching results would be useful to predict fruit ripeness based on digital image, so that the conventional measurement methods that should spoil the fruit can be handled. Experiments were performed in several types of fruits include: bananas, apples and melons. Each of these fruit image is recorded from several difference angles. We also measured levels of sweetness using a sweetness measuring instrument, called refractometer. Both of these will become the training materials for classification system and then performed using kNN. The resulting method of this research is implemented in the application, named Fruitylicious.