Hyperspectral Imaging for Ripe Tomatoes

Why are there always only green bananas in the supermarket? And why is the avocado never really ripe when you want to eat it? The reason for this is the botanical process called climacteric. So-called climacteric fruits such as bananas, avocados and tomatoes continue to ripen even after the harvest. However, to ensure that the fruits are ready for consumption despite long transport routes, they are harvested while still unripe. The Brix or sugar content can be determined with hyperspectral imaging and determines the perfect time for harvesting.

 

Spectroscopy Detects the »Fingerprint« of Materials 

With the help of spectroscopy, we investigate how a sample changes the spectrum of the incident light. Hyperspectral sensors provide data from more than 200 spectral bands, because the reflection curve of each spectral band is measured for each image pixel. This enables the exact identification of the material and its chemical composition, as each material has its own spectral signature, such as an individual »fingerprint«. Hyperspectral imaging solutions play a particularly important role in food inspection, as this method offers many advantages:

• high sensitivity of the sensors
• good linearity in low light
• low readout noise
• high process speed

A hyperspectral sensor consists of an objective, a light source, a spectrograph, which breaks down the reflected light into its components, and a detector chip (FPA), which records this data. This data is stored as a hyperspectral data cube. It comprises all frames, each of which contains the complete spectral data for each pixel. To process the large amounts of data, spectral algorithms with artificial intelligence are used, which are trained for various characteristics such as size, degree of rottenness, colour or shape.

In addition to defective products and foreign materials, this technology can also be used to measure the ripeness of fruit and vegetables non-destructively. As these mainly consist of water and various sugars, the density and thus the Brix value also approximately determines the sugar content.

The same technology used here for food inspection can also be used for other inline applications. Plastic and metal recycling are two examples where hyperspectral imaging is used in waste separation.