Method for detecting position, posture, and three-dimensional profile of vehicle from traffic images

Abstract

The invention discloses a method for detecting the position, posture, and three-dimensional profile of a vehicle from traffic images, which comprises the following steps: carrying out image acquisition on traffic scenes by a digital camera; detecting the position, posture, and three-dimensional profile of a vehicle by a symbolic activity unit model; processing possible occluded phenomenon of the vehicle; and finally, calculating the actual size of the three-dimensional profile of the vehicle according to the calibration results of the digital camera and the identified profile of the vehicle. The invention has good adaptability to various common shooting angles in a traffic monitoring system, and by using the method, the three-dimensional profile of the vehicle can be observed and detectedfrom different angles, therefore, the invention can be used in vehicle recognition, plate-number identification, and vehicle behavior judgment.

Description

technical field [0001] The invention relates to the field of intelligent traffic monitoring, the method belongs to the technical field of image processing and pattern recognition, and is a method for detecting vehicle position, attitude and three-dimensional outline from traffic images. Background technique [0002] In the field of intelligent traffic monitoring, accurate acquisition of vehicle position, attitude, outline and other information has a wide range of applications, such as statistics of traffic flow, license plate recognition, vehicle classification and other applications. With the advancement of science and technology, there have been many kinds of traffic video surveillance systems that can detect the location and size of vehicles, but there are still two problems: first, the existing vehicle identification technology cannot accurately obtain vehicle information. The information of the three-dimensional outline of the vehicle, which brings great difficulties to...

AI Technical Summary

Problems solved by technology

With the advancement of science and technology, there have been many kinds of traffic video surveillance systems that can detect the location and size of vehicles, but there are still two problems: first, the existing vehicle identification technology cannot accurately obtain vehicle information. The information of the three-dimensional outline of the vehicle, which brings great difficulties to further applications, leading to the use of approximate calculation results in the existing system applications (such as vehicle classification according to the length of the vehicle), with large errors

Second, a practical problem is that as urban traffic becomes more and more congested, vehicles in traffic video surveillance images usually have occlusions, which makes it difficult to identify vehicles, and occlusions usually degrade the performance of vehicle recognition systems

If the blocked vehicle cannot be identified, it will not be able to meet the requirements of the traffic video surveillance system in a congested and busy traffic environment

Images