Vehicle illegal parking detection method based on convolutional neural network

Abstract

The invention provides a vehicle illegal parking detection method based on a convolutional neural network. The method comprises the steps that (1) a no-parking region in a video frame image, a presetviolation early-warning threshold value and a preset violation threshold value are set respectively; (2) target vehicles in the set no-parking region are detected through a convolutional neural network model, and vehicle information is recorded; (3) whether a current vehicle set A in the step (2) and a historical vehicle set B contain vehicles which can be matched is judged; (4) whether delay timeof a current target vehicle Ai is greater than the preset violation early-warning threshold value in the step (1) is judged; (5) whether the early-warned delay time of the current target vehicle Ai is greater than the preset violation threshold value in the step (1) is judged; and (6) the vehicle information in the historical vehicle set B is updated, a next video frame image is input, the step (2) is entered to start vehicle violation detection on the next video frame image, and the operation from the step (2) to the step (5) is repeated.

Description

technical field [0001] The invention relates to the technical fields of intelligent transportation systems, deep learning and image recognition, and in particular to a method for detecting illegal parking of vehicles based on a convolutional neural network. Background technique [0002] With the progress and development of society, the number of urban motor vehicles is increasing, and the economic losses and casualties caused by road traffic accidents are also increasing. Controlling traffic accidents has become a problem that traffic management departments pay more and more attention to, and the primary cause of traffic accidents is The reason is automobile violations. As a common violation, illegal parking will not only cause traffic jams and traffic paralysis, but may even cause serious traffic accidents. At present, the detection method for illegal parking is mainly by manual detection. However, this requires arranging personnel to manually monitor all places where park...

AI Technical Summary

Problems solved by technology

In the actual road scene application, there are often the following problems: 1) Outdoor lighting conditions change significantly over time, and the video features of vehicles under different lighting conditions during the day and night are very different; Factors and vehicles have similar motion characteristics, making it difficult to distinguish; 3) Under severe weather conditions such as rain, snow, fog and lightning, the video quality drops seriously

These methods fail to perform effective vehicle detection in the foreground area because they believe that the foreground area is where the vehicle appears, which makes it difficult for such methods to obtain the detection accuracy required for practical applications.

Images