Vehicle image optimization method and system based on adversarial learning

Abstract

The invention discloses a vehicle image optimization method and system based on adversarial learning. The vehicle image optimization method comprises the steps: collecting vehicle images photographedat different angles, and dividing the vehicle images into a standard scene image and a non-standard scene image; carrying out image preprocessing on the non-standard image to obtain a low-quality dataset; constructing a vehicle image optimization model based on the generative adversarial network, wherein the model is composed of a generator, a discriminator and a feature extractor; training a vehicle image optimization model based on the generative adversarial network, setting a loss function, calculating a network weight gradient by adopting back propagation, and updating parameters of the vehicle image optimization model; and after the vehicle image optimization model is trained, reserving the generator as a final vehicle image optimization model, inputting multi-scene vehicle images, and outputting optimized standard scene images. According to the invention, migration from complex scene vehicle images to standard scene vehicle images is realized, and the purpose of optimizing the image quality is achieved, and the vehicle detection and recognition accuracy is improved.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a vehicle image optimization method and system based on adversarial learning. Background technique [0002] In the field of public transportation, vehicle detection and recognition technology is widely used in checkpoint systems, intelligent transportation systems, automatic driving systems, electronic police systems and other sub-fields. However, there is a common problem in current vehicle detection and recognition technologies. The vehicle recognition algorithm has very strict requirements on the input data, such as centered viewing angle, clear picture, appropriate brightness, etc., otherwise the accuracy rate will drop significantly. Therefore, in real complex monitoring scenarios, such as the vehicle shooting angle is too large, outdoor lighting Changes, unclear imaging, occlusions, etc., the application of existing algorithms has great limitations, and it is prone ...

AI Technical Summary

Problems solved by technology

[0002] In the field of public transportation, vehicle detection and recognition technology is widely used in checkpoint systems, intelligent transportation systems, automatic driving systems, electronic police systems and other sub-fields. However, there is a common problem in current vehicle detection and recognition technologies. The vehicle recognition algorithm has very strict requirements on the input data, such as centered viewing angle, clear picture, appropriate brightness, etc., otherwise the accuracy rate will drop significantly. Therefore, in real complex monitoring scenarios, such as the vehicle shooting angle is too large, outdoor lighting Changes, unclear imaging, occlusions, etc., there are great limitations in the application of existing algorithms, and the problem of high misidentification rate is prone to occur

[0003] As for the vehicle image optimization technology, currently more is the image optimization of a single scene, such as vehicle angle correction. The existing solution is to extract the key points of the vehicle first, and perform matrix calculations based on the key point positions to achieve vehicle alignment. Based on traditional machine learning, Manual extraction of features is required, the degree of automation is low, and it is difficult to adapt to complex scenes. In addition, this method can only be used for vehicle angle correction, and other factors that cause the decline in recognition accuracy are not taken into consideration, so there are certain limitations. sex

Images