A method, device and camera for detecting illegal driving behavior of animal-drawn vehicle targets

Abstract

Embodiments of the present invention provide a method, device, and camera for detecting illegal driving behavior of an animal-drawn vehicle target, wherein the method for detecting illegal driving behavior of an animal-drawn vehicle target includes: acquiring an image to be detected in a specified scene; using a preset target detection algorithm, treating Detect images for non-motor vehicle target detection, judge whether there are non-motor vehicle targets in the image to be detected, and determine the area images where each non-motor vehicle target is located; use the preset target recognition algorithm to perform target recognition on each area image, and judge Whether each non-motor vehicle target is an animal-drawn vehicle target; use the preset behavior detection algorithm to detect the behavior of each animal-drawn vehicle target, judge whether each animal-drawn vehicle target has illegal driving behavior, and determine the type of illegal driving behavior. Through this scheme, the illegal driving behavior of the animal-drawn vehicle target can be detected correctly.

Description

technical field [0001] The invention relates to the technical field of intelligent transportation, in particular to a method, device and camera for detecting illegal driving behavior of an animal-drawn vehicle target. Background technique [0002] Animal-drawn carts refer to vehicles that rely on animal power (such as cattle, horses, donkeys, mules, etc.) to drive on the road. At present, animal-drawn vehicles appear less frequently in urban traffic scenes, but in some remote suburbs, the use of animal-drawn vehicles is still relatively common. Animal-drawn vehicles have a greater risk factor during use and pose greater safety hazards. Moreover, compared with the huge traffic flow, the frequency of illegal driving behavior of animal-drawn vehicles is low. If the traditional manual screening method is used to screen out the illegal driving behavior of animal-drawn vehicles from a large number of surveillance video frames, since most of the Surveillance video frame is invalid...

AI Technical Summary

Problems solved by technology

[0004] Since the animal-drawn vehicle target is composed of three parts: driver, livestock and man-made means of transportation, the attribute characteristics of the animal-drawn vehicle target are relatively complex. If the above-mentioned vehicle target detection method is directly applied to the animal-drawn vehicle target detection, it has a high false detection rate, and , due to the complex and diverse behaviors of animal-drawn vehicle targets, the above-mentioned vehicle target detection method cannot correctly detect whether the animal-drawn vehicle target is illegally driving

Images