Kinect-based call making and answering monitoring method of driver

Abstract

The invention provides a Kinect-based call making and answering monitoring method of a driver. The method comprises the following steps: adopting Kinect somatosensory equipment to carry out image acquisition on a driving area, and carrying out face detection on an acquired image through a face strong classifier which finishes training; if a face is in the presence, adopting the Kinect somatosensory equipment to capture the limb movement of the driver, extracting the depth information of the driver, and constructing a human body limb structure bone diagram according to the extracted depth information; according to the human body limb structure bone diagram, independently calculating a relative distance between a head node and a left hand node and the relative distance between the head node and a right hand node; if the relative distance between the head node and the left hand node or the relative distance between the head node and the right hand node is smaller than a limiting value, judging as an abnormal driving state; and if the duration of the abnormal driving state exceeds a preset threshold value, giving an alarm. The Kinect-based call making and answering monitoring method has the advantages of being high in stability and accurate and reliable in detection results and improves driving safety.

Description

technical field [0001] The invention relates to the technical field of safe driving monitoring, in particular to a Kinect-based method for monitoring a driver's calls. Background technique [0002] With the improvement of people's living standards this year, cars have gradually entered people's lives as a family necessity. The rapid growth of car ownership has brought great convenience to people, but at the same time it has also led to frequent occurrence of various traffic accidents, and people's lives and property have suffered great losses. According to statistics, the probability of accidents while driving while looking at mobile phones is 23 times that of ordinary driving, and the probability of accidents while driving and talking on the phone is 2.8 times that of ordinary driving. [0003] How to effectively prevent the driver from answering and calling while driving, many scholars have proposed some good ideas. For example, Chinese patent application CN104573659A dis...

AI Technical Summary

Problems solved by technology

[0003] How to effectively prevent the driver from answering and calling while driving, many scholars have proposed some good ideas. For example, Chinese patent application CN104573659A discloses "a method for monitoring the driver's answering and calling based on svm", and Chinese patent application CN104156717A discloses "A Method for Recognition of Driver Violation of Phone Calls Based on Image Processing Technology", these two patent applications monitor the driver in real time through the camera, and judge whether the driver is answering the phone through the recognition of the hands and head. In the actual driving environment, the image is greatly affected by the light, and it is difficult to detect it accurately; Chinese patent application CN104486485A discloses "a method for actively preventing the driver from answering the phone". The Hall switch detects the magnetic flux at high speed and prevents the driver from answering and calling. In this way, the signal of the mobile phone is detected through the mobile phone, and it cannot be determined whether it is the mobile phone signal of the driver or the passenger, and the purpose of monitoring the driver's answering and calling cannot be achieved.

Images