SVM algorithm-based pedestrian road crossing decision modeling method

Abstract

The invention relates to the technical field of traffic safety, and particularly discloses a SVM algorithm-based pedestrian road crossing decision modeling method, and the method comprises the steps of collecting actual pedestrian data, and collecting the facial feature data of a plurality of actual pedestrians in the road crossing process through a face recognition device; extracting the number of times of checking direction switching and the duration of keeping the side-looking state of the pedestrians who violate crossing the sidewalk within a period of time when the pedestrians start to pass through the sidewalk, and obtaining an original data set; dividing the original data set into a traversal data set and a waiting data set, and respectively adding a traversal decision type label and a waiting decision type label; when the pedestrian road crossing decision model based on the SVM algorithm is trained, taking the number of times of viewing direction switching and the duration of keeping a side view state of pedestrians as independent variables, and taking decision type labels as dependent variables; and testing the model. The invention is low in modeling cost and convenient in data acquisition, and the obtained model has the advantages of small calculation amount, high operation speed and accurate prediction.

Description

technical field [0001] The invention relates to the technical field of traffic safety, in particular to a decision modeling method for pedestrians crossing a road based on an SVM algorithm. Background technique [0002] In real life, the situation of pedestrians crossing the road illegally occurs from time to time, including the situation of pedestrians crossing the road at intersections, sidewalks and traffic lights. Illegal crossing of pedestrians poses a threat that cannot be ignored to vehicles in motion and pedestrians themselves. With the development of assisted driving technology, more and more assisted driving technologies are gradually reducing the driver's driving burden. With the rapid development of image recognition technology, especially the development of face recognition technology and computer computing power, it is possible to predict the decision-making intention of pedestrians through face recognition technology. In order to solve the problem of accurat...

AI Technical Summary

Problems solved by technology

[0004] The Chinese patent with the application number CN201910066829.4 and titled "A System and Method for Detection of Non-motor Vehicle or Pedestrian Crossing Intent" proposes to use the lateral position data of the non-motor vehicle, the number of times the cyclist looks back or the pedestrian's The number of times of turning heads, the lateral position and moving speed of pedestrians on the sidewalk to judge the intention of pedestrians or motor vehicles to cross the street, but it does not involve the method of modeling based on the data collected when multiple actual pedestrians cross the road, which predicts pedestrians crossing the street The accuracy of intent is not reliable

Images