Abstract:
Aflatoxins are harmful secondary metabolites produced by the fungus called Aspergillus flavus, and they represent a severe threat to the safety and cause contamination in agricultural products, particularly peanuts. The presence of these toxins in peanuts does not only endanger human health, but also threaten global food security and economic stability, especially in developing regions where peanuts and other agricultural products are a staple food and cash crops. Their prevalence is exacerbated under specific circumstances, such as dry weather conditions near crop maturity, high moisture levels during harvest, and insufficient drying and poor storage of crops. These toxins can cause liver cancer, stunted growth in children and rejected exports. The Food and Agriculture Organization (FAO) estimates that 25% of global food crops are contaminated with mycotoxins including Aflatoxins. The gravity of this issue is further underscored by the fact that globally, 28.8% of liver cancer cases are Aflatoxin-induced, with 40% of these cases occurring in Africa. Additionally, nearly 5 billion people in developing countries are exposed to Aflatoxins due to a lack of Aflatoxin testing facilities. To address this critical issue, this practical research project designs and implements an affordable system, combining Internet of Things (IoT) and Convolution Neural Network (CNN)-based system to detect Aflatoxins in peanuts and monitor environmental conditions in peanuts storage room. It employs a simple CNN architecture and three pretrained models for training, namely MobileNetV2, VGG16, and ResNet50 using TensorFlow in google colab. MobileNet achieved a 98% accuracy rate, while VGG16 attained 81% and ResNet50 attained 75%. The IoT component enables real-time monitoring of environmental conditions during peanut storage using sensors, issuing alerts as needed, while the CNN analyses peanut images for early Aflatoxin detection. The system also features a user-friendly web-based application that enables stakeholders to remotely monitor and assess storage conditions. The developed prototype of this system improves the prevention of the growth and consumption of these harmful substances.
