Mold Testing

Li, Pao, et al. Microchemical Journal 193 (2023): 109203.

Citrus fruits, essential in global agriculture and trade, are vulnerable to mold infections, which can occur during production, storage, and transport, leading to significant postharvest losses. Detecting these infections, particularly in their early stages, is challenging, as mold often manifests internally, without visible symptoms. Traditional methods fail to identify mold hidden beneath the citrus peel, especially before sporulation.

A promising solution involves portable Near Infrared Diffuse Reflectance Spectroscopy (NIRDRS) combined with chemometric techniques. This study investigates the penetrability of NIR light into the citrus peel and evaluates its potential for detecting hidden mold. The results reveal that short-wave near infrared (SWNIR) light penetrates more effectively than long-wave near infrared (LWNIR). Identification models based on LWNIR significantly outperformed those based on SWNIR, achieving 100% accuracy in mold detection using advanced pattern recognition methods like Soft Independent Pattern Classification (SIMCA), Support Vector Machine (SVM), and Partial Least Squares Discriminant Analysis (PLS-DA). These models were further validated with an external data set collected one month later, ensuring robust performance.

This rapid, non-destructive approach offers a reliable method for detecting mold infection in citrus, safeguarding consumer rights and enhancing postharvest quality control.

Farrugia, Jessica, et al. Current Research in Food Science 4 (2021): 18-27.

The detection of mould contamination in food products is crucial for ensuring food safety and quality. In recent years, non-destructive process analytical technologies, such as hyperspectral imaging, have garnered significant attention in the food science industry for their ability to detect contaminants without damaging the product. This study explores the use of hyperspectral imaging combined with Principal Component Analysis (PCA) to detect mould growth on milk agar and cheeselet samples.

The results demonstrated that PCA effectively localized and visualized mycelial growth on the surface of contaminated samples. The first three principal components (PCs) accounted for over 99% of the total variance, with the second PC being particularly useful in highlighting the presence of mould on cheeselets. By applying PCA loadings obtained from a set of training samples, it was possible to accurately detect contamination on new test samples, confirming the method's robustness.

This approach offers a rapid, non-contact solution for detecting mould contamination in food production, potentially improving quality control in cheese production and other food industries. The study also suggests that this methodology can be adapted to detect contaminants in a variety of solid and semi-solid food products, paving the way for its widespread use in food safety monitoring.

Ebrahimi, Parvaneh, et al. Food Control 55 (2015): 82-89.

Accurate quantification of mold growth is essential for microbiological studies, particularly in the context of biopreservation and predictive microbiology. However, traditional methods often fail to provide precise, non-destructive measurements of mold colony dynamics. This study introduces a new approach for quantifying mold growth, employing multispectral imaging combined with k-means clustering to assess the size and composition of mold colonies.

The method was successfully applied to measure the growth of "Penicillium" mold on different media, including transparent and opaque surfaces like milk and CDIM (complex defined industrial media). The approach demonstrated exceptional sensitivity, allowing for the quantification of small colony size differences that would otherwise go undetected by visual inspection. Notably, the k-means algorithm effectively distinguished between the white and green segments of the colonies, enabling precise quantification of mold growth.

This non-destructive technique proved particularly valuable for studying how environmental factors, such as pH and medium composition, influence mold growth. The method was also successful in tracking the inhibition effects of antifungal treatments, revealing differences in growth patterns that would be difficult to discern through visual inspection alone.

The PCluster algorithm offers a robust, objective, and less labor-intensive solution for monitoring mold growth, providing a valuable tool for both academic and industrial applications in mold control and food biopreservation.