Vehicle attribute identification method based on RDSNet

Abstract

The invention relates to a vehicle attribute identification method based on RDSNet, and the method comprises the following steps: S1, collecting a vehicle image, processing the vehicle image, and dividing the vehicle image into a vehicle data set and a vehicle attribute data set; S2, constructing a network model based on RDSNet, and performing training according to the vehicle data set to obtain avehicle detection model; S3, constructing a vehicle attribute classification network model based on fine-grained classification, and performing training according to the vehicle attribute data set toobtain a vehicle attribute classification model; S4, enabling the to-be-detected complex scene image to pass through the vehicle detection model, and obtaining an accurate boundary frame of each vehicle of the to-be-detected complex scene image; S5, inputting the image processed in the step S4 into a vehicle attribute classification model to obtain vehicle attribute information; and S6, marking the obtained accurate boundary frame and vehicle attribute information of the vehicle in a to-be-detected complex scene image. The accuracy degree of vehicle attribute identification is effectively improved.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to an RDSNet-based vehicle attribute recognition method. Background technique [0002] In recent years, with the development of science and technology, vehicle detection and recognition has become a research field of great interest to people, and has been applied in many mature application scenarios. Vehicle attribute recognition is one of the hot research fields. Most of the traditional vehicle attribute recognition methods are the superposition of object detection and fine-grained classification. However, the vehicle detection model may not be well suited to the task of fine-grained classification, and the classification of vehicle attributes depends on the vehicle frame detected by the vehicle detection method. This is especially true in complex scenes. [0003] In order to identify the vehicle attributes, we must first detect the vehicle, which can be achieved by target detecti...

AI Technical Summary

Problems solved by technology

For example, the detection frame cannot surround the vehicle well, resulting in the loss of some feature points, which affects the identification of subsequent vehicle attributes.

Images