Monitoring pistachio health using data fusion of machine vision and electronic nose (E-nose)

Abstract

Pistachios, often referred to as “green gold” due to their high economic value, are vulnerable to various pests, with aflatoxin contamination being a particularly critical issue. The cracks in pistachio shells create an ideal environment for fungal growth and the insects that spread them. Contamination by toxic molds and subsequent aflatoxin production poses a significant threat to pistachio exports, making accurate detection essential. Current detection methods primarily rely on chemical analysis, which can be time-consuming, labor-intensive, and expensive. In this study, we developed a cost-effective and reliable approach for detecting fungal contamination in pistachios by combining an electronic nose (E-nose) equipped with eight metal oxide semiconductor sensors and color imaging technology. Experimental treatments were prepared using three spore concentrations: 10², 10⁴, and 10⁶ spores/mL. A medium-level data fusion strategy was employed and compared with Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM) models. Results showed that different fungal concentrations could be effectively distinguished by the third day post-inoculation. These findings demonstrate that integrating color imaging with E-nose technology offers a powerful solution for intelligent, in situ detection of fungal contamination, ensuring food safety and quality control in the pistachio industry.