Video vehicle detection method of traffic jam scene based on hidden Markov model

Abstract

The invention relates to a video vehicle detection method of a traffic jam scene based on a hidden Markov model (HMM), belonging to the field of intelligent traffic vehicle detection based on video. The invention solves the problem of searching video vehicles under the jam scene by utilizing a method and the HMM based on a virtual coil and the inherent characteristics of a vehicle flow. The method comprises the following steps of: dividing images in a region into three states of vehicle heads, vehicle bodies and road surfaces; extracting the characteristics by utilizing Haar wavelet transform, main component analysis (PCA), multiple decision analysis (MDA) and the like; training the model; revising a state sequence to improve the detection efficiency of the method; and analyzing the obtained state sequence to achieve the purpose of vehicle detection by searching a section from the vehicle head to the vehicle body in the sequence. The invention describes the characteristics of the vehicle flow by utilizing the HMM, solves the problem of vehicle video detection under the jam condition, has the advantages of easy realization and simple structure and is suitable for real-time application.

Description

technical field [0001] The video vehicle detection method of the congested traffic scene based on the hidden Markov model involves setting a suitable virtual coil in the video of the traffic scene, and then using the method to analyze the image changes in the virtual coil to achieve the purpose of vehicle detection, which belongs to video-based intelligence. The field of traffic vehicle detection. Background technique [0002] Video vehicle detection techniques can be divided into two categories: region detection-based methods and tracking-based methods. The virtual coil method is a typical method based on area detection. This type of method realizes vehicle detection by detecting image changes in a specific area of ​​the video. Since human prior knowledge is added when selecting the detection area, and The actual processing area is much smaller than the entire video area, the efficiency of the method is very high, and it is suitable for real-time application, but the stabi...

AI Technical Summary

Problems solved by technology

[0008] Aiming at the problems of high computational complexity and low detection rate in congested traffic scenes faced by most current video vehicle detection technologies, the present invention proposes an HMM-based video vehicle detection method under congested conditions. On the one hand, the method Taking advantage of the low computational complexity of the virtual coil-based method, on the other hand, HMM is used to model the traffic flow to improve the detection rate of vehicles in congested traffic scenarios.

Images