Effect of Frame Processing Frequency on Object Identification Using MobileNetV2 Neural Network for a Mobile Robot

This article is devoted to the analysis of the relationship between frame processing frequency, identification speed, recognition accuracy and system resources. The article presents a mathematical description of the dependencies, which allows for a quantitative assessment of the effect of changing the processing frequency on these parameters. A program code based on the Python language was developed, which integrates the MobileNetV2 model for practical testing and optimization. The conducted experiments allow us to identify optimal settings that provide a balance between recognition accuracy and processing speed, taking into account system resources. The results of the study provide useful recommendations for improving computer vision systems in mobile robots