Automated papaya harvests possible with ripeness detecting device
Image credit: Dreamstime
Brazilian researchers have created a non-invasive device which identifies the different stages of fruit ripening with 95 per cent accuracy. This could make the mechanisation of papaya harvesting a reality.
As papayas ripen, they soften and their skins usually turn from deep green to a warm yellow-amber colour. While human pickers can identify ripe fruit from their appearance and softness, this is a complex real-world challenge for computers.
“Yellow peel doesn’t always mean the fruit is ripe in the sense that its compounds have been completely converted into sugar, so that it’s sweet and soft to the palate,” said Douglas Fernandes Barbin, a researcher at the school of food engineering at the University of Campinas, Brazil. “Sometimes the peel is partly yellow and partly green.”
To open up the possibility of a reliable automated method for picking this fruit, it is necessary to develop a system capable of detecting which fruit is ripe without destroying the fruit in the process. Supported by the São Paulo Research Foundation, Barbin is leading a project to do just this using algorithms and computer-vision techniques.
According to Barbin, Brazilian farmers are keen to mechanise the papaya harvest. This could allow, for instance, papayas to be picked such that they reach the ideal stage of ripeness as they arrive at markets in other parts of the country.
“The idea is to automate harvesting using non-invasive imaging technologies such as computer-aided analysis of digital images in visible light as well as infrared,” he said.
Barbin and his team have been investigating the use of a portable sensor that illuminates the fruit for spectral analysis. While looking at visible light reflected from papayas (looking at their colour) may not guarantee ripeness, infrared light reflected from the fruit contains extra important information about the biochemical stage of ripeness. Supplementing data collected from digital images of the fruit with this information enhances the precision of the system.
To test the non-invasive system, the researchers bought, measured and weighed a selection of golden papayas. They determined the exact colour of the peel using a calorimeter and also analysed properties such as firmness, pH and ascorbic acid content. Based on these mechanical measurements, the researchers were able to categorise the papayas into three stages of ripeness.
Next, the researchers used their device to estimate the ripeness of the fruit. Two colour images of each papaya were processed to separate colour channels, then evaluated. The characteristics of the papayas were fed into an algorithm to model the categorisation of fruit ripeness. These predictions were 94.7 per cent accurate compared with their painstakingly obtained values for ripeness.
Barbin and his colleagues are working to transform the sensor into a portable manual device that can be used practically in plantations. Eventually, this system could be used as a critical component of fruit-picking robots like those proposed for cucumber farms and under development at the University of Plymouth to replace seasonal European farm labourers once the UK exits the EU.