Wide area traffic jam detection method based on unmanned plane airborne platform

Abstract

The invention discloses a wide area traffic jam detection method based on an unmanned plane airborne platform and belongs to the technical field of video images. The method comprises steps that 1, the GPS technology is utilized to acquire height of an unmanned plane to the ground, whether jam detection can be carried out is determined, if not, the spatial position of the unmanned plane is adjusted; 2, positive and negative samples of infrared vehicle images are acquired in an offline mode, HOG characteristics of the samples are extracted, and an SVM model is trained; 3, an infrared image signal is acquired by utilizing an infrared camera; 4, sliding window sampling of the infrared images is carried out, HOG characteristics of a sliding window zone are extracted, a SVM model is utilized to carry out vehicle detection, statistics of a total number N of vehicles in a present vision field is carried out; and 5, according to the height H of the unmanned plane and the total number N of the vehicles, a jam index C is calculated, and the detection result is returned. Through the method, detection on a traffic jam state under the condition of bad weather and low visibility can be accomplished by utilizing the infrared camera, moreover, detection zones can be dynamically selected by the unmanned plane airborne platform, and a detection system is made to be more flexible.

Description

technical field [0001] The invention belongs to the field of video image technology and intelligent transportation technology, and relates to a wide-area traffic congestion detection method based on an unmanned aerial vehicle airborne platform. Background technique [0002] With the economic development, the acceleration of the urbanization process and the continuous expansion of the city scale, the number of motor vehicles continues to increase. According to statistics from the Traffic Management Bureau of the Ministry of Public Security, as of the end of 2015, the number of motor vehicles in the country reached 279 million, including 172 million cars and 583,200 new energy vehicles; Each household owns 31 vehicles. In addition, the number of motor vehicle drivers has reached 327 million, of which more than 280 million are car drivers. The sharp increase in motor vehicle ownership and road traffic flow has made urban traffic congestion increasingly serious. [0003] Traf...

AI Technical Summary

Problems solved by technology

Among them, the magnetic induction (electromagnetic, geomagnetic) detection method is not affected by weather, light, etc., and has stable performance and is widely used. However, it usually needs to be buried on a fixed ground, which is prone to misjudgment of abnormal driving of the vehicle and has a failure rate. High, difficult to maintain and other deficiencies; the wave frequency (ultrasonic, microwave) detection method will attenuate with the propagation distance during the propagation process, so the echo signal is relatively weak, easy to be submerged in the noise, and there are also complicated installation, occlusion , inconvenient maintenance and many other shortcomings; in comparison, the video detection method has gradually become the mainstream method of current traffic jam detection because of its advantages such as wide application range, simple installation process, and high accuracy.

[0006] However, the current video detection methods are mostly based on the color image principle for traffic detection, which is prone to missed vehicle detection at night, in bad weather conditions, and low visibility, and the congestion detection accuracy is extremely low

In addition, the camera equipment used to collect traffic data is usually fixed and lacks flexibility

Images