Using machine learning for food quality and safety assurance

The latest research published by eResearch Office’s Dr Frederic Isingizwe on detecting defects in fresh agri-food products dealt with detecting soft damage to apple fruit while they are still invisible to the naked eye.

Damage to fresh agri-food products due to brute impact or compression force can occur during handling and transport, can be invisible at an early stage but becomes more pronounced with time, either in the consumer’s hands or on a retailer’s shelf. Such damage to fresh produce accelerates the deterioration of fruit and vegetables and can facilitate infections by micro-organisms, which makes products unsafe to consume.

The research was conducted to aid with sorting and grading fresh products, either at an industrial or smaller scale. We demonstrated that these invisible defects can be detected using shortwave hyperspectral imaging techniques and by using machine learning algorithms, we established the degree to which the differentiation of defective from sound apple fruits is feasible.

Read more about this work here.

X-ray Observations Supporting MIGHTEE MeerKAT Science Project Unveiled

New X-ray map reveals the growing supermassive black holes in next-generation MeerKAT survey fields 

One of the largest X-ray surveys using the European Space Agency’s XMM-Newton space observatory has mapped nearly 12,000 X-ray sources across three large, prime regions of the sky. The X-ray sources represent active galactic nuclei and galaxy clusters, and the survey captures the growth of the supermassive black holes at the cores of these galaxies. This X-ray survey complements previous X-ray surveys, allowing the researchers to map active galactic nuclei in a wide range of cosmic environments.

The XMM-SERVS survey lays key groundwork for studying the cosmic history and physical  properties of active galaxies 

Figure 1: XMM-Newton image of the 4.6-square-degree W-CDF-S field reveals the wide, sensitive view of the X-ray sky provided by XMM-SERVS. The detected sources, most of which are growing supermassive black holes, are color coded according to the energies of the X rays detected (with red having the lowest energies and blue the highest). The white outline indicates the area of the Chandra Deep Field-South, a well-known ultradeep pencil-beam X ray survey. The image highlights how XMM-SERVS has now provided sensitive panoramic X ray imaging around this survey. The XMM-Newton image covers an area about 20 times larger than the apparent size of the full moon, shown to scale at upper left.
Figure 2: XMM-Newton image of the 3.2-square-degree ELAIS-S1 field, which is about 15 times larger than the apparent size of the full moon (shown to scale at lower right). XMM SERVS provides a wide, sensitive X-ray view of this region.

These X-ray observations will be invaluable to study the active galactic nuclei (i.e. black holes) and galaxy clusters (the largest cosmic structures bound together by gravity) detected by the MIGHTEE MeerKAT Large Survey Project (led by UWC Visiting Professor Matt Jarvis and UWC Research Chair Russ Taylor) in its ongoing mission to study the faint radio sky.

Qingling Ni and W. Niel Brandt from Penn State presented the results of the XMM-Spitzer  Extragalactic Representative Volume Survey (XMM-SERVS) at a press briefing during the 238th meeting of the American Astronomical  Society on 7 June. A paper describing the survey, authored by an international team of astronomers including UWC’s eResearch Director Prof Mattia Vaccari, has been accepted for publication in The Astrophysical Journal Supplement. A pre-print is also available on arxiv.org.

“X-ray surveys are the best way to find growing supermassive black holes, which are located at  the cores of many large galaxies,” said Ni, a graduate student at Penn State and lead author of  the paper. “With this massive new survey, we can access population data about growing  supermassive black holes to better understand their physical properties and evolution over cosmic history.” 

“This survey represents key foundational work upon which, I suspect, hundreds of studies will  be built over the next decade or two,” said Brandt, Verne M. Willaman Professor of Astronomy  and Astrophysics and professor of physics at Penn State, and one of the leaders of the study.  “XMM-Newton was the best mission to gather these data, and we needed to invest a lot of  observation time for this study—with a total combined exposure of nearly 60 days—because it  will be so important for active galaxy studies, galaxy cluster studies, and for understanding  large-scale structures in the universe. It required a multiyear, multinational effort and it’s  incredibly gratifying to get it done. We are most grateful to the European Space Agency and  NASA for their long-term support of this work.” 

Caption from featured image: XMM-Newton image of the 5.3-square-degree XMM-LSS field, which is about 25 times  larger than the apparent size of the full moon (shown to scale at lower right). XMM-LSS was the  first XMM-SERVS field to have been observed by XMM-Newton. Chien-Ting Chen, a former postdoctoral researcher at Penn State who is now an astronomer at USRA, led the work for this  field (see Chen et al. 2018, Mon. Not. Roy. Ast. Soc.). XMM-SERVS provides a wide, sensitive X ray view of this region