Vehicle card monitoring video deblurring method based on common camera acquisition condition
A monitoring video and camera technology, applied in the field of computer vision, can solve problems such as unstable model training, and achieve the effects of improving stability, improving quality, and improving convergence speed
Pending Publication Date: 2022-02-18
NORTHWESTERN POLYTECHNICAL UNIV
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
This method can effectively preprocess the input data and optimize the network training process by using the auxiliary information of optical flow, which can overcome the problem of unstable training of the existing model and significantly improve the image quality
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0021] The present invention will be further described below in conjunction with the accompanying drawings and embodiments, and the present invention includes but not limited to the following embodiments.
[0022] Such as figure 1 As shown, the present invention provides a method for deblurring surveillance video of trucks based on common camera acquisition conditions, and its specific implementation process is as follows:
[0023] 1. Video data acquisition and preprocessing
[0024] (1) Data collection
[0025] Two cameras are used to shoot the scene of the truck truck with the same angle of view on the same road section at the same time, and the video data consisting of several frames of blurred and clear image pairs are obtained.
[0026] Before data collection, some adjustments can be made to the two cameras, including: use a tripod and a quick release plate to set up two cameras of different specifications side by side at the same position, keep the distance between the...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention provides a vehicle card monitoring video deblurring method based on a common camera acquisition condition. The method comprises the following steps: firstly, carrying out pixel-level feature matching on continuous frames acquired by a vehicle card camera; then, acquiring motion optical flow information between frames by using a PWC-Net network, and adding auxiliary features for a network learning process; thirdly, designing an end-to-end neural network for multi-frame deblurring, and performing end-to-end network training by taking the aligned continuous frames and optical flow data as input; and finally, processing by using the trained network to obtain a deblurring result image. According to the method, the input data is effectively preprocessed, and the network training process is optimized by using the optical flow auxiliary information, so that the problem that the existing model is unstable in training can be overcome, and the image quality is remarkably improved. Meanwhile, due to the fact that the real fuzzy video data set shot by the two video imaging devices is utilized, the method is higher in practicability and pertinence.
Description
technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a method for deblurring monitoring video of trucks based on common camera acquisition conditions. Background technique [0002] Video deblurring is a research topic that has attracted much attention in the field of computer vision in recent years. The process of video deblurring can be modeled as a potential clear frame based on the information of several consecutive video frames at a certain moment and processed by the designed network. Among them, the key lies in the structure and processing methods and capabilities of the network. [0003] Video deblurring algorithms are mainly divided into three types. One is based on the priori traditional algorithm, which makes a priori assumption between the blurred image and the clear image, and then estimates the optical flow and the corresponding potential clear image. In , the prior information is often manually designed. ...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06T5/00G06T3/40G06N3/08G06N3/04G06V10/46G06V10/82
CPCG06T3/4007G06N3/08G06T2207/20081G06T2207/30232G06T2207/10016G06N3/043G06N3/045G06T5/73
Inventor 王琦马欣李学龙
Owner NORTHWESTERN POLYTECHNICAL UNIV