Measuring the Ripeness of Fruit with Hyperspectral Imaging and Deep Learning

TL;DR

This paper introduces a system combining hyperspectral imaging and deep learning to accurately assess fruit ripeness, validated on avocados and kiwis, with a focus on data collection, model performance, and visualization of ripening.

Contribution

It presents a novel deep learning approach using hyperspectral data for fruit ripeness measurement, including a publicly available dataset and visualization techniques.

Findings

Deep neural network outperforms baseline models in ripeness prediction.

A new dataset of ripening avocados and kiwis is made publicly available.

A visualization method for the ripening process is introduced.

Abstract

We present a system to measure the ripeness of fruit with a hyperspectral camera and a suitable deep neural network architecture. This architecture did outperform competitive baseline models on the prediction of the ripeness state of fruit. For this, we recorded a data set of ripening avocados and kiwis, which we make public. We also describe the process of data collection in a manner that the adaption for other fruit is easy. The trained network is validated empirically, and we investigate the trained features. Furthermore, a technique is introduced to visualize the ripening process.