CN-121305868-B - Video processing device and system for road traffic safety monitoring

CN121305868BCN 121305868 BCN121305868 BCN 121305868BCN-121305868-B

Abstract

The invention relates to video image communication, in particular to a video processing device and a system for road traffic safety monitoring, which comprise the steps of acquiring road traffic video image data through a video acquisition module, utilizing a video compression preprocessing module, acquiring importance degrees of all partial areas in a video image and a lossy compression threshold value through analyzing LBP values and light transmittance in the video image data, preprocessing the video image data before compression, and realizing efficient compression and data transmission as well as accurate analysis and early warning of the video image of road traffic through a video coding compression module, a video image communication module, a video decompression module, a video storage module, an alarm module and a visualization module. The invention can preprocess the video image while protecting important data so as to improve the compression efficiency of the video image, ensure the credibility of the analysis result of the subsequent video data and obtain accurate and reliable warning signals.

Inventors

WU JINFENG
LIU JIAMING
DAI QIANWEN
Zhang Guokan
YUAN DONGXU

Assignees

深圳市汇宝翔科技有限公司

Dates

Publication Date: 20260512
Application Date: 20251015

Claims (9)

1. A video processing system for monitoring highway traffic safety, comprising the following modules: the video acquisition module is used for acquiring road traffic video data; the video compression preprocessing module is used for acquiring the preprocessed road traffic video data and comprises the following steps: Obtaining a sky area and a scenery area of a foggy day gray image and a foggy day road image, obtaining light transmittance and LBP values of pixel points in the scenery area, obtaining the number of the pixel points corresponding to each type of LBP values in the scenery area, the number of types of all light transmittance and the number of types of light transmittance of the pixel points corresponding to each type of LBP values, and obtaining a fourth characteristic quantity according to the difference of the light transmittance of the pixel points corresponding to each type of LBP values; Obtaining a first correction result by using the type number of the light transmittance of the pixel points corresponding to each type of LBP value in the scene area and the fourth characteristic quantity, and obtaining a first importance degree of the pixel points corresponding to each type of LBP value according to the type number of the pixel points corresponding to each type of LBP value in the scene area and the first correction result; Obtaining a second important degree factor according to the number of the types of the gradient mean values of the pixel points corresponding to each type of LBP value and the number of the types of the gray feature mean values, obtaining a second important degree according to the second important degree factor, setting the second important degree of all the pixel points in the sky area to be 0, and jointly calling the second important degree of the pixel points in the sky area and the second important degree in the scene area as a third important degree of the pixel points; Dividing the region according to the third importance degree to obtain a plurality of divided regions, marking the average value of the third importance degree of all pixel points of each divided region as a fourth importance degree, and obtaining the compression threshold value of each region according to the fourth importance degree; The video coding compression module is used for compressing the road traffic video data by using a compression threshold value; The video image communication module is used for rapidly transmitting the encoded and compressed road traffic video data; The video storage module is used for storing the compressed or communication-transmitted road traffic video data; The video decompression module is used for decompressing the road traffic video data transmitted by the video image communication module or stored by the video storage module; the alarm module is used for decompressing the video image analysis of the module and reminding people and vehicles on the road to pay attention to safety; the visualization module is used for displaying the decompressed video data and the warning information; the first importance degree is obtained by the following steps: Wherein the elements of the third feature set The number of pixel points on the jth LBP value in the scene area in the foggy road image is represented, the first characteristic quantity N represents the number of pixel points in the scene area in the foggy road image, and the third characteristic quantity Representing the number of types of light transmittance values corresponding to pixel points on the jth LBP value in the scene area in the foggy-day road image, and a second characteristic quantity A fourth feature quantity representing the number of kinds of light transmittance values of pixel points in the scene area And the difference value of the maximum light transmittance and the minimum light transmittance in the light transmittance corresponding to the pixel point on the jth LBP value in the scene area in the foggy-day road image is represented.
2. The video processing system for monitoring traffic safety of a highway according to claim 1, wherein the sky area and the scene area are obtained by the following steps: And using the part, in the dark channel image of the foggy weather image, with the gray value of the pixel points larger than the threshold value as a sky area and using the rest pixel points as scenery areas.
3. The video processing system for monitoring highway traffic safety according to claim 2, wherein the optimal segmentation threshold is obtained by the following method: and taking a dark channel image of the foggy day image as input of an Ojin algorithm, and marking an algorithm output result as an optimal segmentation threshold.
4. The video processing system for monitoring highway traffic safety according to claim 1, wherein the method for obtaining the fourth feature value according to the difference of the light transmittance of the pixel point corresponding to each type of LBP value comprises the following steps: and calculating the difference between the maximum light transmittance and the minimum light transmittance in the light transmittance corresponding to the pixel points on various LBP values in the scene area, and recording the difference between the maximum light transmittance and the minimum light transmittance as a fourth characteristic quantity.
5. The video processing system for monitoring highway traffic safety according to claim 1, wherein the gradient mean value category number and the gray feature mean value category number are obtained by the following steps: Counting the number of pixels corresponding to the same light transmittance in a scene area, obtaining gradient average values of the pixels under the same light transmittance according to the gradient values of the pixels and the number of the pixels corresponding to the same light transmittance, and taking the gradient average values of the pixels under all light transmittance in the scene area as a gradient average value set; Counting the number of gray value types of all pixel points in a scene area in an 8 neighborhood range including the pixel points as gray features, obtaining gray feature average values of the pixel points under the same light transmittance according to the gray features of the pixel points and the number of the pixel points corresponding to the same light transmittance, and taking the gray feature average values of the pixel points under all light transmittance of each type of LBP value in the scene area as a gray feature average value set.
6. The video processing system for monitoring highway traffic safety according to claim 1, wherein the second importance level is obtained by the following method: taking the normalized product result of Euclidean norms of the gradient mean value variety quantity and the gray feature mean value variety quantity and the first importance degree as a second importance degree, and taking the second importance degree of all pixel points in the scene area as a second importance degree set.
7. The video processing system for monitoring highway traffic safety according to claim 1, wherein the compression threshold is obtained by the following method: According to the third importance degree of each pixel point in the foggy-day road image, carrying out region division on the foggy-day gray level image by utilizing a region growing algorithm, and taking the average value of the third importance degree of each divided region as the fourth importance degree of each divided region; the maximum compression threshold is reduced by a multiple of the fourth degree of importance, and the reduced result is used as the compression threshold of the divided area.
8. The video processing system for monitoring highway traffic safety according to claim 7, wherein the maximum compression threshold is obtained by the following method: and taking the average value of the gray level difference absolute values between all pixel points in the divided area and the transversely adjacent pixel points as the maximum compression threshold value.
9. A video processing device for road traffic safety monitoring, which adopts the video processing system for road traffic safety monitoring according to any one of claims 1-8, and is characterized in that the device comprises a monitoring camera in a video acquisition module, a processor in a video compression preprocessing module, a video coding compression module and a video decompression module, 5G communication network equipment in a video image communication module, a memory in a video storage module, a display of a visualization module, a processor in an alarm module and an audible and visual alarm device.

Description

Video processing device and system for road traffic safety monitoring Technical Field The invention relates to the technical field of image communication, in particular to a video processing device and a system for road traffic safety monitoring. Background The road traffic accident warning signal equipment and the system collect video data in the peripheral range of each road section in a foggy environment through the monitoring cameras of the fully covered road video monitoring network, analyze, process and compress the video data carrying real-time condition information, transmit the collected video data to a data processing center in batches, in a component and in a time-sharing mode through a 5G transmission network, upload the collected video data to a cloud storage or a local storage, finally process and scientifically and reasonably analyze the mastered video data, give out a professional processing scheme and command information for controlling a warning signal device, realize the transmission of the command information to each warning signal device, complete the intelligent switching of the intelligent lamp mark working mode and help drivers to better acquire road traffic conditions. In order to ensure the accuracy of the instruction information sent to each warning signal device through the central console in order to ensure the analysis result of the road traffic video data passing through the video data compression processor, the warning system is required to collect a large amount of video data in the peripheral range of each road section in the foggy environment, so that the reliability of the subsequent analysis of the video data is ensured, but the excessive video data volume can increase the load when the network transmits the data, and the video data transmission delay can occur, so that the subsequent warning signal delay transmission according to the analysis result of the road traffic condition can cause more serious traffic accidents. The existing video data compression is usually lossless or lossy compression on the whole data, the complete lossless compression can transmit partial video data with little meaning, the load effect on reducing the network transmission video data is not obvious, and the lossy compression can lose the video data for analyzing the road traffic condition, so that the analysis on the road traffic condition is inaccurate. Therefore, the method and the device have the advantages that partial data with little significance for analyzing the road traffic condition in the video data are subjected to lossy compression through selective compression, the partial data with significance for analyzing the road traffic condition are reserved, the compression efficiency of the video data and the real-time and high-efficiency transmission of the data are improved, the reliability of the transmitted video data is ensured, and the central console can rapidly and accurately send instruction information to each warning signal device. Disclosure of Invention The invention provides a video processing device and a system for road traffic safety monitoring, which aim to solve the existing problems. The invention relates to a video processing device and a system for road traffic safety monitoring, which adopts the following technical scheme: one embodiment of the present invention provides a video processing system for monitoring highway traffic safety, which comprises the following modules: the video acquisition module is used for acquiring road traffic video data; the video compression preprocessing module is used for acquiring the preprocessed road traffic video data and comprises the following steps: Obtaining a sky area and a scenery area of a foggy day gray image and a foggy day road image, obtaining light transmittance and LBP values of pixel points in the scenery area, obtaining the number of the pixel points corresponding to each type of LBP values in the scenery area, the number of types of all light transmittance and the number of types of light transmittance of the pixel points corresponding to each type of LBP values, and obtaining a fourth characteristic quantity according to the difference of the light transmittance of the pixel points corresponding to each type of LBP values; Obtaining a first correction result by using the type number of the light transmittance of the pixel points corresponding to each type of LBP value in the scene area and the fourth characteristic quantity, and obtaining a first importance degree of the pixel points corresponding to each type of LBP value according to the type number of the pixel points corresponding to each type of LBP value in the scene area and the first correction result; Obtaining a second important degree factor according to the number of the types of the gradient mean values of the pixel points corresponding to each type of LBP value and the number of the types of the gray feature mean values, obtaining a second important