The majority of grocery store visitors assume little else about how to get a small sweet potato into a shopping basket from the ground. Customers have come to expect a certain consistency from the slightly special red root fruit, which can be found in a variety of sizes and flavor profiles. Sweet potatoes are subjected to lengthy and laborious quality checks to identify problematic batches that are either too strong, not sweet enough, or often deemed improbable to sell in order to meet that market demand. A new research suggests that spectral cams and AI could speed up the process, which is now being performed meticulously by humans in a lab.
Researchers from the University of Illinois conducted a study that was published this week in Computers and Electronics in Agriculture to examine whether the results of a spectral diagnostic camera’s surveillance may help identify specific corn characteristics that are generally determined by regular investigators and testing. Large amounts of data are collected by spectroscopic monitors across the electromagnetic spectrum, which are frequently used to study the chemical makeup of components. In this case, the researchers wanted to see if they could examine data from the potato images to accurately determine a spud’s firmness, liquid content, and clean matter content—three important attributes that contribute to the vegetable’s general taste and market appeal. Normally, this method requires tedious, often inefficient testing that can include leaving test potatoes heated in a 103-degree Celsius oven for 24 hours.
“Traditionally, quality assessment is done using laboratory analytical methods,” University of Illinois College of Agricultural, Consumer, and Environmental Sciences assistant professor Mohammed Kamruzzaman said in a statement. “You need to wait for the results and you need different instruments to measure different attributes in the lab.”
The researchers photographed 141 defect-free sweet potatoes from a variety of angles. The abundance of data that hyperspectral imaging produces can be a blessing and a curse for researchers looking for particular variables. The researchers used an AI model to help filter down the noisy data into various wavelengths to address that issue. Then, they were able to connect those wavelengths to the particular desirable sweet potato characteristics they were looking for.
“With hyperspectral imaging, you can measure several parameters simultaneously. You can assess every potato in a batch, not just a few samples,” Kamruzzaman added.
Vegetable inspection could be accelerated by AI and hyperspectral cameras.
Farmers and food inspectors could use their combination of artificial intelligence and hyperspectral imaging to precisely and affordably screen sweet potatoes for important attributes while reducing food waste created as a byproduct of conventional testing, according to the researchers. And while this particular study focused on sweet potatoes, it’s possible to find desired characteristics in a wide range of other vegetables and fruits as well. Kamruzzaman claims that in the end, he and his coworkers want to quickly and simply make batches of scan dough. The researchers envision developing an app for the consumer market that can scan a potato and look up its specific characteristics one day. Such an app, in theory, could cut down on patrons awkwardly fondling their produce.
Doctoral student and study lead author Toukir Ahmed wrote,” We think this is a novel application of this method for sweet potato assessment.” This cutting-edge research has the potential to help with applications in a wide range of other agricultural and biological research fields as well.
In an effort to increase efficiency and combat growing farm labor shortages, the agriculture industry is increasingly turning to AI solutions. Farmers around the world are hoping these new innovations can eventually lower food prices and increase their own profitability at the same time as autonomous Tulip-inspecting machines in Holland or self-driving John Deere tractors. How exactly that will all play out, however, remains to be seen. AI-based agricultural advancements may take longer to realize in economically developing nations, where some farming is still done by hand.
