Maintaining hygiene in kitchen environments is one of those problems that sounds straightforward but is surprisingly hard to automate reliably. Manual inspection is inconsistent, time-consuming, and simply not practical in busy commercial settings. This paper describes a Smart Kitchen Hygiene Monitoring System (SKHMS) that was built to address this gap using a combination of computer vision and low-cost sensor hardware. The system had two parts working together: a software module that used a laptop webcam, OpenCV preprocessing, and a TensorFlow-based deep learning model to classify kitchen hygiene conditions in real time; and a hardware module built around an Arduino Uno that read data from gas, temperature, and air quality sensors. The deep learning model covered three areas vegetable freshness, countertop cleanliness, and personal hygiene of kitchen staff and achieved an overall classification accuracy of 94.8% with precision, recall, and F1-score values of 94.1%, 94.5%, and 94.3% respectively. The sensor module reliably flagged LPG leaks, overheating, and smoke within acceptable response times. Tested under realistic kitchen conditions, the system showed that combining vision-based and sensor-based monitoring into one platform is both feasible and practical for everyday deployment.
Sonawane et al. (2026) studied this question.