Missing child detection method for school bus safety

Abstract

The invention relates to a school bus safety-oriented detection method for missing children, which combines infrared, video and audio sensor signals, and two advanced computer processing technologies of video analysis and audio analysis to reliably detect whether there are children left on the school bus. At the same time, the ARM processing board is used for reasonable power control, which reduces system power consumption and can be widely used in the field of school bus safety. Combining infrared, video and audio signals, it can reliably detect whether there are children left behind on the school bus, and it has low power consumption, convenient installation, and low cost, and can be widely used in the field of school bus safety.

Description

technical field [0001] The invention relates to a method for detecting missing children and belongs to the technical field of video surveillance. Background technique [0002] Missing children is a major hidden danger of school bus safety. In recent years, there have been many child deaths caused by missing children on school buses, which has aroused strong social concern. At present, it is mainly through strengthening ideological education and formulating more reasonable safety measures to avoid missing children as much as possible. For example, let the teacher who follows the bus go to the rear of the school bus to trigger the safety warning device after the students get off the bus, so that the teacher can carefully check whether there are missing children. However, manual inspection may cause accidents due to human negligence or carelessness, and there are potential safety hazards. In order to minimize the tragedy caused by missing children on the school bus, it is urg...

AI Technical Summary

Problems solved by technology

For example, in the field of human detection, there are currently many methods that use video analysis technology to detect human bodies. For example, Jing Shao et al. (Deeply Learned Attributes for Crowded Scene Understanding, CVPR 2015) use deep learning methods to detect human targets in crowded scenes, but this method The effect of human detection on vehicles with occlusion is not good; Seung-Wook Kim et al. (Sensor Fusion-Based People Counting System Using the Active AppearanceModels, ICCE 2013) use the active shape model to detect the head area, but this method is mainly for shopping malls, museums In large places such as large places, the camera is installed far away from the people, so that the head area is consistent and the interference is small; however, due to the limited space of the vehicle, it is difficult to apply this technology to the vehicle

Juan Chen et al. (Automatic Head Detection for Passenger Flow Analysis in Bus Surveillance Videos, CISP 2012) proposed a people counting method based on video analysis, which counts the number of people entering and leaving by installing a camera at the door, but cannot detect the people left in the car

In general, relying solely on a certain sensor will inevitably lead to false detection or missed detection.

Images