Quality assurance for fresh bananas is vital for consumers and the fruit industry. Quick and precise classification of a banana’s ripeness stage is key to this process, necessitating the use of image processing tools. The ripeness of a banana significantly influences its taste, market price, and overall quality. To address this, we proposed an automatic computer vision system to identify the ripening stages of bananas. The process of fruit maturity detection is divided into various stages, depending on the type of fruit detection used.

A Raspberry Pi 5MP camera, coupled with image processing methods, is employed for this purpose. The primary stages involved in this process include pre-processing, detection using the camera, and image processing.

We use an artificial neural network-based framework that utilizes color, the development of brown spots, and Tamura statistical texture features to classify and grade the ripening stage of banana fruit. Spectroscopy has significantly advanced image processing techniques.

In the final stages, the results are classified or clustered according to specific requirements. This approach ensures a more efficient and accurate assessment of banana ripeness, ultimately enhancing the quality control measures in place within the fruit industry.

A screenshot of the output using an example is given below, the [2] indicates that the banana is over-ripe. An unripe banana would indicate [0] and ripe one would result in an output as [1].