Low cost camera allows to know the inner state of fruits

26/10/2015

It is always difficult to select a fruit in the supermarket, on the outside it looks in perfect condition, but inside it may already be ripening. For this reason, the University of Washington (UW) together with Microsoft Research, developed a camera that allows us to see the interior of the product in greater detail.

This is HyperCam, a low-cost hyperspectral camera that uses both visible and near-infrared light to "see" under surfaces and capture details that are not seen.

This type of camera is normally used in industrial applications and can cost from several thousand to tens of thousands of dollars, reports the University of Washington.

In a document presented at the UbiComp 2015 conference, the team detailed that the hardware solution costs about USD 800, or potentially as little as USD 50 if only the essentials are manufactured and added to a mobile phone camera.

The researchers also developed smart software that easily finds the "hidden" differences between what the hyperspectral camera captures and what can be seen with the naked eye.

As part of a preliminary investigation of the utility of HyperCam as a biometric tool, in a test with 25 users the system was able to differentiate between images of users' hands with 99% accuracy.

In another test, the team also took hyperspectral 10 fruit images, including strawberries, mangoes and avocados (avocados) over the course of a week. The HyperCam images predicted the relative ripeness of fruits with 94% accuracy, compared to the 62% of a typical camera.

“With this type of camera you can go to the grocery store and find out which fruit to choose by looking under the skin and seeing if there is something wrong inside. It's like having a food safety app in your pocket. ", said Shwetak Patel, professor of computer science and electronic engineering at the Washington Research Foundation at UW.

Today hyperspectral imagery is used for everything from satellite imagery and energy monitoring to infrastructure and food safety inspections, but the high cost of the technology has limited its use for industrial or commercial purposes. Given this, the team at the University of Washington and Microsoft Research wanted to see if it could make a relatively simple and affordable hyperspectral camera for consumer use.

As reported, the system still has to overcome some challenges. For example, the camera does not work particularly well in bright light and the team is looking for ways to make it smaller to include it in a mobile phone.

Source: Fruit Portal