What fruit looks like on the outside does not always say something about its taste, we all know from experience. An apple or kiwi may look attractive, but its bite and taste can be quite disappointing. Naturally, growers have every interest in meeting their customers’ taste expectations. Wageningen University & Research (WUR) has now developed a device that allows the grower to determine whether the taste of their produce is indeed perfect, without having to destroy fruit to analyse the juice or pulp.
In recent years, WUR’s Agro Food Robotics programme has been working on the development of the so-called all-in-one spectral cabinet. This is a simple looking device, equipped with light, a near-infrared camera, and software. This enables them to map out how much water and sugar the fruit contains and how firm and sweet it is, for example.
Lights, a near-infrared camera, and software
Researcher Puneet Mishra explains how this works. “You put a tray of fruit in the device, the infrared light shines on the fruit, the camera measures the amount of light the fruit reflects and captures this image. The software then translates this into meaningful data and voila, you know how much sugar or water your fruit contains.”
Until now, the sugar content of fruit was determined with a refractometer. Mishra: “But for this you need juice, so you have to destroy the fruit. With this all-in-one spectral cabinet, that is no longer necessary. In this way you save the fruit, you have less loss and waste, and in addition, you save time.”
Breakthrough in fruit cultivation
Other advantages of the device: you can measure an entire batch of fruit at once. Growers like to deliver a homogeneous product, where all fruits are equally ripe, firm, and sweet. In this way they can measure whether that is indeed the case. And the growers can measure it themselves; the device is so user-friendly that anyone throughout the supply chain can operate it. A smaller version is also available to analyse the composition of the fruit while it is still on the tree. “This is really a breakthrough in fruit cultivation,” says Mishra.
One model per fruit type
The device was completed this year, but the researchers are still working on fine-tuning the artificial intelligence behind it. Mishra: “Because fruit species differ enormously from one another in terms of sugar content or physical structure, you need a model for each species in order to be able to draw conclusions about the taste.” In any case, what does not need to be repeatedly reset are the camera’s focus, exposure time, and frame rate. These have now been optimised.
The researchers have so far focused mainly on fruit, but the device lends itself really well to other products as well. For example, growers are very interested in the moisture content of certain plant cuttings, because this says something about their quality.
To Fruit Logistica
In recent months, Mishra and his colleagues have been to three Dutch trade fairs with their all-in-one spectral cabinet. Mishra: “We are also in dialogue with partners who can build the machines and with representatives to sell them. Next February we will probably go to Fruit Logistica, a well-known fruit trade fair in Berlin, to get more people excited about the invention.”