CN-121999063-A - Multi-video stream color change intelligent monitoring method and system based on multi-scale time sequence analysis

CN121999063ACN 121999063 ACN121999063 ACN 121999063ACN-121999063-A

Abstract

A multi-video stream color change intelligent monitoring method and system based on multi-scale time sequence analysis belong to the technical field of intelligent video monitoring. Aiming at the problem that abnormal reddening of a water body needs to be monitored in real time in scenes such as sewage treatment, firstly, an independent processing channel is established for each path of video stream, a current frame is converted into an HSV color space from RGB, a double-range HSV red detection algorithm is adopted to accurately identify a red area and calculate the red pixel duty ratio of each frame, time sequence fusion is carried out on a duty ratio sequence through moving average filtering, a reference red duty ratio is dynamically updated on the basis of a weighted sliding window, a multi-scale criterion is established by combining an adaptive absolute threshold value and a relative change rate threshold value to realize abnormal judgment, and once obvious color mutation is detected, an alarm is immediately triggered and information such as a time stamp, the red duty ratio, the change rate and the like is output. The system supports concurrent processing of multiple paths of videos, has high sensitivity, strong robustness and low false alarm rate, and is suitable for intelligent visual monitoring in the fields of water affairs, environmental protection and the like.

Inventors

LIU LONGZHI
CAI YINGTAO
HE JINXIN
YU ZHOU
LIU YU
LU XI
CHEN XIANG
Lv Chenkai
JIANG YUNPENG
LI TONG
WANG JIAN
WEN TAO

Assignees

中国长江三峡集团有限公司
长江生态环保集团有限公司

Dates

Publication Date: 20260508
Application Date: 20260104

Claims (20)

1. The intelligent monitoring method for the color change of the multi-video stream based on the multi-scale time sequence analysis is characterized by comprising the following steps of: s1, initializing a plurality of monitoring threads, and establishing independent processing channels for each video stream; s2, performing RGB to HSV color space conversion on the current frame of each video stream; s3, detecting a red area based on a double-range HSV red detection algorithm, wherein the double-range covers a red area with low tone and high tone in a hue circle; S4, calculating the red duty ratio of each frame; s5, carrying out moving average filtering on the red duty ratio sequence to obtain a fused red duty ratio; S6, dynamically updating the reference red duty ratio based on the weighted sliding window; s7, combining the self-adaptive absolute threshold value and the self-adaptive relative change rate threshold value, and executing a multi-scale color change criterion to judge whether abnormality occurs; and S8, when the judgment is abnormal, triggering an alarm mechanism and outputting abnormal information.
2. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 1, wherein in the step S2, the mathematical expression of RGB to HSV color space conversion is: ; Wherein, the Is an input RGB image which is displayed as a picture, Is an HSV image of the output, The coordinates of the pixels are represented and, Representing a time stamp.
3. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 1, wherein in the step S3, the dual-range HSV red detection algorithm defines two detection masks, the red detection mask is a logical or operation result of the two masks, wherein the first mask covers the areas of hues H e [0,10], saturation S e [70,255], brightness V e [50,255], and the second mask covers the areas of hues H e [170,180], saturation S e [70,255], brightness V e [50,255 ].
4. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 3, wherein in the step S3, the dual-range HSV red detection algorithm defines the following two detection masks: ; ; The red detection mask is: ; wherein H, S, V represent hue, saturation and brightness channels of the HSV format, respectively.
5. The intelligent monitoring method for color change of multi-video stream based on multi-scale time sequence analysis according to claim 1, wherein in the step S4, the red ratio of the current frame is obtained by counting the total number of red pixels and dividing the total number of pixels by the total number of pixels of the image.
6. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 1, wherein in the step S4, the calculation formula of the red duty ratio of the current frame is: ; Wherein, the Representing the total number of red pixels in the current frame, And The width and height of the image, respectively.
7. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 1, wherein in the step S5, the fused red duty ratio is obtained by performing arithmetic average calculation on the red duty ratio of the latest L frames.
8. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 7, wherein in the step S5, the fused red duty ratio is obtained through moving average filtering, and the calculation formula is as follows: ; Wherein, the For a moving average of the filtered red duty cycle, For the filter window length.
9. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 1, wherein in the step S6, the reference red duty ratio is obtained by acquiring the red duty ratio of the previous N frames and performing weighted average calculation according to exponential decay weights.
10. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 1, wherein in the step S6, the initial value of the reference red duty ratio is obtained by weighted average calculation, and the formula is as follows: ; Wherein, the Is the first The weight coefficient of the frame, and satisfies: ; Parameters (parameters) And updating the reference value under the condition of no abnormality at regular time or continuously, wherein the updating formula is as follows: ; Wherein, the In order to update the weights, the weights are updated, Is the average red duty cycle over the most recent time window.
11. The intelligent monitoring method for color change of a multi-video stream based on multi-scale time sequence analysis according to claim 1, wherein in the step S7, the relative change rate is calculated by adopting the relative change rate when the reference red duty ratio is larger than a preset minimum value, and the absolute change rate is calculated when the reference red duty ratio is smaller than or equal to the preset minimum value, and the multi-scale color change criterion is judged to be abnormal when the fused red duty ratio exceeds an absolute threshold or the relative change rate exceeds the relative change rate threshold.
12. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 11, wherein in step S7, the relative change rate is The calculation formula of (2) is as follows: ; Wherein, the Minimum positive number for preventing zero-removal error, multi-scale color change criterion The expression of (2) is: ; Wherein, the As an absolute threshold value, Is a relative rate of change threshold.
13. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 12, wherein the adaptive absolute threshold in step S7 And adaptive relative threshold Through the statistical dynamic adjustment of historical data, the calculation formulas are respectively as follows: ; ; Wherein, the As a historical red duty cycle average, Is the standard deviation of the two-dimensional image, And In order to adjust the coefficient of the coefficient, The median absolute deviation, defined as the ratio of change sequence: 。
14. The intelligent monitoring method for color change of multiple video streams based on multi-scale time sequence analysis according to claim 1, wherein in the step S8, the anomaly information includes a time stamp, a red duty ratio, a change rate and an alarm state.
15. A multi-video stream color change intelligent monitoring system based on multi-scale time sequence analysis, characterized in that the multi-video stream color change intelligent monitoring method based on multi-scale time sequence analysis according to any one of claims 1-14 is adopted, the system comprises: the multi-stream management module is used for concurrent access and connection management of a plurality of video streams; The color space conversion module is connected with the multi-stream management module and used for executing RGB to HSV color space conversion; The red area detection module is connected with the color space conversion module and is used for detecting a red area based on a double-range HSV red detection algorithm; the red duty ratio calculation module is connected with the red area detection module and is used for calculating the red duty ratio of each frame; The time sequence fusion processing module is connected with the red duty ratio calculation module and is used for carrying out moving average filtering on the red duty ratio sequence; The self-adaptive reference updating module is connected with the time sequence fusion processing module and is used for dynamically updating the reference red duty ratio based on the weighted sliding window; The intelligent threshold optimization module is used for dynamically adjusting an absolute threshold and a relative change rate threshold based on historical data statistics; The abnormality detection and alarm module is respectively connected with the time sequence fusion processing module, the self-adaptive reference updating module and the intelligent threshold optimization module and is used for executing multi-scale criteria and triggering alarm.
16. The intelligent multi-video stream color change monitoring system based on multi-scale time sequence analysis of claim 15, wherein the red region detection module adopts a dual-range HSV red detection algorithm to cover a complete red region by combining hue ranges [0,10] and [170,180 ].
17. The intelligent monitoring system for color change of multiple video streams based on multi-scale time sequence analysis according to claim 15, wherein the adaptive reference update module updates the reference value using a sliding window algorithm of exponentially decaying weights.
18. The intelligent monitoring system for color change of multiple video streams based on multi-scale time sequence analysis according to claim 15, wherein the anomaly detection and alarm module performs color change detection by combining an absolute threshold and a relative change rate double criterion.
19. The intelligent monitoring system for color change of multiple video streams based on multi-scale time sequence analysis according to claim 15, wherein the multi-stream management module adopts a multi-thread architecture, each video stream corresponds to an independent monitoring thread, and the detection results of the streams are uniformly stored through a shared memory or a message queue.
20. A computer device comprising one or more processors, wherein the one or more processors have one or more executable programs stored thereon, and wherein the one or more executable programs, when executed by the one or more processors, are configured to implement the multi-video stream color change intelligent monitoring method based on multi-scale timing analysis according to any one of claims 1-14.

Description

Multi-video stream color change intelligent monitoring method and system based on multi-scale time sequence analysis Technical Field The invention belongs to the technical field of computer vision and intelligent monitoring, and particularly relates to an intelligent monitoring method and system for color change of a multi-video stream based on multi-scale time sequence analysis. Background With the continuous improvement of the industrial automation and the intelligent degree, higher requirements are put on the intelligent analysis capability of the video monitoring system. The traditional video monitoring system mainly relies on manual observation to find abnormal conditions, and the mode has the problems of low efficiency, easiness in fatigue, high omission rate and the like. Particularly in application scenes needing to monitor specific color changes for a long time, such as liquid leakage detection of chemical plants, color change monitoring of laboratory test paper and the like, the traditional method is difficult to meet the requirements of instantaneity and accuracy. The existing color change detection technology mainly has the following problems: the environmental adaptability is poor, and environmental factors such as illumination change, camera position fine adjustment and the like easily cause false alarm or missing detection; The noise sensitivity is high, single frame detection is easily affected by factors such as video noise, network jitter and the like; Lack of timing analysis, namely, insufficient utilization of time sequence information for robust detection; The threshold value is difficult to set, namely the fixed threshold value is difficult to adapt to different scenes and environmental conditions; the multi-stream processing capability is limited and the concurrent detection of multiple video streams cannot be managed and coordinated efficiently. Therefore, there is a need to develop a color change monitoring system with strong environmental adaptability, high noise immunity, intelligent thresholding and multi-stream concurrent processing capability. Disclosure of Invention The invention aims to solve the technical problem of providing an intelligent monitoring method and system for color change of a multi-video stream based on multi-scale time sequence analysis, so as to overcome the defects of poor environmental adaptability, high noise sensitivity, lack of time sequence analysis capability, low multi-stream processing efficiency and the like in the prior art. In order to solve the technical problems, the invention adopts the following technical scheme: A multi-video stream color change intelligent monitoring method based on multi-scale time sequence analysis comprises the following steps: s1, initializing a plurality of monitoring threads, and establishing independent processing channels for each video stream; s2, performing RGB to HSV color space conversion on the current frame of each video stream; s3, detecting a red area based on a double-range HSV red detection algorithm, wherein the double-range covers a red area with low tone and high tone in a hue circle; S4, calculating the red duty ratio of each frame; s5, carrying out moving average filtering on the red duty ratio sequence to obtain a fused red duty ratio; S6, dynamically updating the reference red duty ratio based on the weighted sliding window; s7, combining the self-adaptive absolute threshold value and the self-adaptive relative change rate threshold value, and executing a multi-scale color change criterion to judge whether abnormality occurs; and S8, when the judgment is abnormal, triggering an alarm mechanism and outputting abnormal information. Preferably, in the step S2, the mathematical expression of RGB to HSV color space conversion is: ; Wherein, the Is an input RGB image which is displayed as a picture,Is an HSV image of the output,The coordinates of the pixels are represented and,Representing a time stamp. Preferably, in the step S3, the dual-range HSV red detection algorithm defines two detection masks, where the red detection mask is a logical or operation result of the two masks, and the first mask covers the areas of hues H e [0,10], saturation S e [70,255], brightness V e [50,255], and the second mask covers the areas of hues H e [170,180], saturation S e [70,255], brightness V e [50,255 ]. Preferably, in the step S3, the two-range HSV red detection algorithm defines two detection masks as follows: ; ; The red detection mask is: ; wherein H, S, V represent hue, saturation and brightness channels of the HSV format, respectively. Preferably, in the step S4, the red ratio of the current frame is obtained by counting the total number of red pixels and dividing the total number of pixels by the total number of pixels of the image. Preferably, in the step S4, the red duty ratio calculation formula of the current frame is: ; Wherein, the Representing the total number of red pixels in the current frame,AndThe