Image-based vehicle detection method and system

Abstract

The invention relates to an image-based vehicle detection method and system, and relates to the field of vehicle detection. The method comprises the steps of S1, acquiring n images of a road; S2, constructing a database according to the n images; S3, calling m images in the database to perform preprocessing, thereby obtaining a training set and a test set, wherein m is smaller than or equal to n;S4, acquiring a deep neural network model, training the deep neural network model according to the training set and the test set to obtain a C model, and executing the step S5 when various parametersof the C model meet a first preset condition; and S5, detecting a vehicle in a new image by utilizing the C model. According to the scheme, the technical problem of how to improve the stability of vehicle identification is solved; and the image-based vehicle detection method and system is suitable for traffic state detection.

Description

technical field [0001] The invention relates to the field of vehicle detection, in particular to an image-based vehicle detection method and system. Background technique [0002] The current vehicle detection method usually uses UAVs for information collection. Compared with fixed information collection points, UAVs can quickly move to places not covered by information collection points for supplementary information collection. At present, about The use of UAVs for vehicle detection has a certain research foundation, such as the use of image registration and change detection based on feature points to achieve real-time detection and tracking of vehicles, or to detect vehicles based on the median background split method, Or perform vehicle detection based on saliency information and Haar-like local shape features. [0003] However, using the above methods, it is necessary to define an algorithm, manually extract features, and then combine features to realize vehicle recognit...

AI Technical Summary

Problems solved by technology

[0003] However, using the above methods, it is necessary to define algorithms, manually extract features, and then combine features to realize vehicle recognition. Manual feature extraction depends on the experience of the extraction personnel, resulting in unstable vehicle recognition results.

Images