According to data from the Central Statistics Agency (BPS), throughout 2021 Indonesia will be able to produce 8.74 million tonnes of bananas. Production rose 6.82% from the previous year which amounted to 8.18 million tons. Many Indonesians like bananas, but not everyone can easily distinguish ripe bananas. Many studies have been conducted on the classification of banana ripeness levels, but only using 1 camera. In this research, an image processing system will be created with a little manipulation in the form of an additional OV760 camera and light intensity parameters. Data acquisition was carried out using a 0.3 MP OV7670 camera and a 12 MP cellphone camera in condition 1, namely with a light intensity of 21 lux, 1% white light, and 50% brightness and in condition 2 with a light intensity of 43 lux, 100% white light, and brightness 100%. The bananas studied were divided into 3 classifications, namely raw, ripe and overripe. Feature extraction uses the HSV color model and image classification method uses KNN. The data in this study are 336 images, consisting of 312 training images and 24 test data from each camera. The accuracy results of system testing using the OV7670 camera in condition 1 and condition 2 were 41.66% and 75%. Meanwhile, the accuracy results using a cellphone camera in condition I and condition 2 were 33.33% and 95.833%, proving that there was quite a significant influence. As a whole, the system using a cellphone camera produces an accuracy of 64.583%, a precision of 100%, and a recall of 64.583%, while testing using an OV7670 camera produces an accuracy of 58.333%, a precision of 100%, and a recall of 58.333%. The benefit of this research is that it will be known exactly how much influence the light intensity and camera specifications used have on the banana maturity classification system and its accuracy value so that it will make it easier for ordinary people to recognize which bananas are raw, ripe, and overripe.
Key words: classification, banana maturity, knn, OV7670 camera, cellphone camera