Computer vision aims to build a computer that can see like humans. Humans can immediately recognize and define an object after seeing and recording an object. This is different from computer visual systems, where camera recordings cannot be directly translated, defined, and recognized by computers, therefore digital image processing is needed first. In this study, computer vision technology was used to detect fruit based on color, size, and shape in real time. The fruit is placed on a conveyor belt, then the fruit object is captured by a webcam using object tracking. Computer vision algorithms and programs can detect fruit objects and recognize ripe and unripe fruits by converting RGB (Red, Green, Blue) colors into HSV (Hue, Saturation, Value) for the color segmentation process. After the detection process, a selector is placed at the end of the conveyor which is used to sort the fruit into 2 categories, namely ripe and unripe. In addition, this study also determines the size and shape of the fruit. From design, realization, and testing, it was found that the success rate of detecting ripe fruit was 97.33% and unripe fruit was 93.33%. To get maximum results, it needs to be supported by room lighting settings that are kept constant.