A traffic event secondary identification method based on video cloud platform pull stream detection

Abstract

The application relates to the field of intelligent traffic technology and discloses a traffic event secondary identification method based on video cloud platform stream pulling detection, which comprises the following steps: S1, pulling low-bit-rate video data in the cloud, performing low-bit-rate round-patrol event detection through a traffic event identification system, and outputting a preliminary detection result; S2, if the preliminary detection result is a traffic event, pulling high-bit-rate video data corresponding to the low-bit-rate video data, performing high-bit-rate event rechecking detection through the traffic event identification system, and outputting a final identification result; wherein the operation of the traffic event identification system is realized through dynamic scheduling of cloud computing power and sharing of cloud resources. Through low-bit-rate round-patrol event detection and high-bit-rate event rechecking detection of traffic events, the application reduces bandwidth resource consumption, effectively saves network resource and computing resource investment, and ensures identification accuracy under the support of an event detection algorithm based on a video analysis technology.

Description

Technical Field

[0001] The present application relates to the field of smart transportation technology, and specifically is a secondary identification method for traffic events based on video cloud platform streaming detection. Background Art

[0002] At present, most provinces use the "road section-provincial cloud platform-ministerial cloud platform" architecture to build video cloud platforms, that is, to build road section video surveillance systems through local deployment and to build provincial cloud platforms through cloud deployment.

[0003] In addition, to meet the requirements for traffic event recognition accuracy, the traffic event recognition system uses a single recognition mode for event detection, that is, it usually directly pulls videos with a bit rate of 1Mbps or above for detection, which requires a large amount of computing resources and network resources to support.

[0004] However, single-shot recognition mode pulls full, real-time video data with a bitrate of at least 1Mbps from the data source system or platform via protocols such as RTMP, HTTP-HLS, and FLV, and sends it to the traffic incident recognition system for detection. Often, surveillance video detection results indicate no traffic incidents have occurred, resulting in significant resource overhead, consuming essential but not entirely necessary resources and failing to efficiently utilize cloud and network resources. Summary of the Invention

[0005] The purpose of this application is to provide a secondary identification method for traffic events based on video cloud platform streaming detection, and a secondary identification technology based on low-bitrate patrol event detection and high-bitrate event review detection, which can improve the accuracy of event recognition while making full use of resources.

[0006] To achieve the above objectives, the present application discloses the following technical solution: a secondary traffic incident identification method based on video cloud platform streaming detection, the method comprising the following steps:

[0007] S1- Pull low-bitrate video data from the cloud, perform low-bitrate patrol event detection through the traffic event recognition system, and output preliminary detection results;

[0008] S2-If the preliminary detection result is a traffic incident, the high-bitrate video data corresponding to the low-bitrate video data is pulled, and the high-bitrate event is rechecked and detected by the traffic incident recognition system, and the final recognition result is output;

[0009] The operation of the traffic event recognition system is achieved by sharing cloud resources through dynamic scheduling of cloud computing power.

[0010] Preferably, the low bit rate is lower than 1 Mbps; and the high bit rate is higher than 1 Mbps.

[0011] Preferably, the network framework of the traffic event recognition system includes:

[0012] The provincial video cloud platform deployed in the cloud is configured to: communicate with local business systems via IPsec VPN or SSL VPN;

[0013] The network egress is configured as follows: using firewalls and routers as egress devices, with security policies and routing policies;

[0014] The core layer is configured to: achieve high-speed data transmission through core switches;

[0015] The access layer is configured to connect application servers, data servers, and distributed storage devices through access switches.

[0016] Preferably, the network framework of the traffic event recognition system further includes:

[0017] A traffic load balancing module is provided between the network egress and the core layer, and the traffic load balancing module is configured to dynamically adjust the data transmission path according to real-time traffic.

[0018] Preferably, the dynamically adjusting the data transmission path according to the real-time traffic specifically includes:

[0019] D1-Real-time monitoring of the traffic load rate of the network egress and the traffic load rate of the core layer;

[0020] D2-Calculate the difference between the traffic load rate at the network egress and the traffic load rate at the core layer;

[0021] D3-Dynamically allocate path weights of transmission paths according to the difference and the maximum theoretical load rate of the network.

[0022] Preferably, the maximum theoretical load rate of the network is calculated by the following formula:

[0023]

[0024] in, It is the total bandwidth capacity of the network egress and core layer; The peak traffic demand of the network; It is the network transmission delay time; is the task execution cycle;

[0025] The difference is specifically: ;

[0026] in, is the traffic load rate of the network egress, is the traffic load rate of the core layer;

[0027] The path weight is specifically: ;

[0028] in, is the sensitivity coefficient, is the threshold parameter.

[0029] Preferably, the dynamic scheduling of cloud computing power specifically includes:

[0030] Reduce network bandwidth consumption through cloud data connection, and dynamically allocate computing resources based on the computing power requirements of the low-bitrate patrol event detection and the high-bitrate event review detection;

[0031] Adjust the allocation ratio of computing resources based on the urgency of the traffic incident.

[0032] Preferably, the dynamic allocation of computing resources specifically includes:

[0033] The number of video channels detected according to the low-bitrate patrol event And the corresponding code stream , and the number of video channels for high-stream event review and detection And the corresponding code stream , dynamically calculate the total computing power requirements Total computing power requirements Calculated by the following formula:

[0034]

[0035] in, The unit bitrate computing power coefficient of low bitrate video; The unit bitrate computing power coefficient of high bitrate video; is the actual bitrate value of the i-th low-bitrate video; is the actual bitrate value of the j-th high-bitrate video; and is the nonlinear adjustment index; is the priority weighting factor of the j-th high-bitrate video.

[0036] As a preference, the unit code rate computing coefficient of the low code rate video and the unit code rate computing coefficient of the high code rate video are calculated based on the data of historical detection tasks, and the average detection time of the low code rate video is calculated. and the average detection time of high bitrate video Get, specifically:

[0037]

[0038]

[0039] in, is the average bitrate value of low bitrate video; is the average bitrate value of high bitrate video; and is the correction factor; is the dynamic compensation factor.

[0040] Preferably, the correction coefficient has a value range of: , , and in the dynamic allocation of computing resources, the value satisfies .

[0041] Beneficial effects: The secondary identification method of traffic events based on video cloud platform streaming detection of the present application pulls low-bitrate video data in the cloud for low-bitrate patrol event detection, and when the preliminary detection result is a traffic event, pulls the corresponding high-bitrate video data for high-bitrate event review detection, and outputs the final identification result, which reduces bandwidth resource consumption, can effectively save the investment of network resources and computing resources, and, with the support of event detection algorithms based on video analysis technology, can ensure the recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0043] Figure 1 This is a flowchart of a secondary traffic incident identification method based on video cloud platform streaming detection provided in an embodiment of the present application. DETAILED DESCRIPTION

[0044] The following is a clear and complete description of the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0045] In this document, the term "comprising" is intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0046] For example, a traffic incident recognition system launched on a provincial cloud platform in a certain province has access to 10,096 surveillance videos. Using existing single-shot recognition technology, if a 1Mbps stream is used for full, real-time video event recognition, the total bandwidth required for the traffic incident recognition system is approximately 9.86GB. Based on a 70% bandwidth utilization rate, 14.09GB of bandwidth is required. Using secondary recognition technology, if a single patrol of 1,200 128kbps streams is used, and a single review of 60 1Mbps streams is used, the total bandwidth required for the traffic incident recognition system is approximately 0.21GB. Based on a 70% bandwidth utilization rate, 0.30GB of bandwidth is required.

[0047] Secondly, the recognition effectiveness of event detection algorithms based on video analysis technology depends heavily on image clarity. Furthermore, video clarity is directly affected by factors such as video bitrate and resolution. Therefore, the recognition effectiveness of event detection algorithms is also affected by video bitrate and resolution.

[0048] Video bitrate refers to the data flow rate used by a video file per unit of time, typically measured in kbps or Mbps. The bitrate directly determines video clarity and file size. Video resolution, which represents the number of pixels in a video image, affects video clarity and detail. At the same video resolution, a larger bitrate results in a lower compression ratio, and the decoded video file is closer to the original file, resulting in better image quality and clearer images. Therefore, the recognition accuracy of event detection algorithms is positively correlated with bitrate size.

[0049] To obtain clear images and improve algorithm recognition accuracy, existing technologies typically use high-bitrate (1Mbps or above) videos for event detection. This means using full, real-time high-bitrate videos for event recognition, which results in a significant resource overhead. To maximize resource utilization, this paper studies event recognition accuracy under different video bitrate conditions to explore an event recognition model that balances event recognition accuracy with resource consumption. By performing traffic event recognition on 3,000 videos with different bitrates (including 13,000 event samples), it was found that the recognition accuracy for 32K bitrate videos was 60%, for 128K bitrate videos was 85%, and for 1M bitrate videos was 95%, indicating that event recognition accuracy increases with the increase in video bitrate.

[0050] In order to meet the requirement of traffic event recognition accuracy ≥ 85% and maximize resource utilization, this embodiment proposes a traffic event secondary recognition technology of "low bit rate patrol event detection + high bit rate event review detection". Figure 1 A secondary traffic incident identification method based on video cloud platform streaming detection is shown, and the method includes the following steps:

[0051] S1- Pull low-bitrate (less than 1Mbps) video data from the cloud, perform low-bitrate patrol event detection through the traffic event recognition system, and output preliminary detection results;

[0052] S2-If the preliminary detection result is a traffic incident, the high-bitrate (greater than or equal to 1Mbps) video data corresponding to the low-bitrate video data will be pulled, and the high-bitrate event will be reviewed and detected through the traffic event recognition system to output the final recognition result; the low-bitrate / high-bitrate video pulling can be implemented based on existing video streaming protocols (such as RTSP, RTMP), and the bandwidth occupancy can be controlled by adjusting the bitrate resolution (such as switching from 480p to 1080p).

[0053] The traffic event recognition system operates by sharing cloud resources through dynamic scheduling of cloud computing power. The traffic event recognition system is deployed in the cloud and on the same cloud as the cloud platform. To meet the event detection needs of video cloud platforms with cloud-based video sources, the traffic event recognition system, combined with secondary traffic event recognition technology, achieves efficient utilization of computing resources through dynamic scheduling of cloud computing power. Furthermore, through cloud data integration, network resource investment is reduced, saving construction costs.

[0054] Results from a specific application show that in terms of network resources, the bandwidth required by the secondary recognition technology is approximately 2.13% of that of the single-shot recognition mode. Regarding computing resources, the computing power was calculated based on a single-channel 128kbps bitstream video detection of 3 TFLOPS (FP16) and a single-channel 1Mbps bitstream video detection of 9 TFLOPS (FP16). The computing power required for a single round of detection in the single-shot recognition mode is 90,864 TFLOPS; the maximum computing power required for a single round of secondary recognition is 4,140 TFLOPS, resulting in approximately 4.56% of the computing resources required for the primary mode. Compared to the single-shot recognition mode, secondary recognition technology can effectively save network and computing resources.

[0055] Therefore, based on the above, the secondary identification method of traffic events based on video cloud platform streaming detection in this embodiment pulls low-bitrate video data in the cloud for low-bitrate patrol event detection, filters out videos without events, and when the preliminary detection result is a traffic event, pulls the corresponding high-bitrate video data for high-bitrate event review detection, and outputs the final identification result, so as to realize high-bitrate review only for suspected events, reduce bandwidth resource consumption, effectively save network resources and computing resource investment, and, with the support of event detection algorithms based on video analysis technology, ensure recognition accuracy.

[0056] In one embodiment, the network framework of the traffic event recognition system includes:

[0057] The provincial video cloud platform deployed in the cloud is configured to: communicate with local business systems via IPsec VPN or SSL VPN;

[0058] The network egress mainly meets the north-south interconnection between the LAN and external networks. It is configured to use firewalls and routers as egress devices, and configure security policies and routing policies as needed to ensure network security.

[0059] The core layer is responsible for high-speed and efficient transmission of data and traffic between various parts of the network. It is configured to achieve high-speed data transmission through core switches, that is, to provide high-speed forwarding for traffic packets entering and leaving the LAN;

[0060] The access layer is responsible for connecting application servers, data servers, distributed storage and other devices to the local area network. It is configured to connect application servers, data servers and distributed storage devices through access switches. That is, business systems such as operation monitoring and video surveillance are deployed on servers in the local computer room and connected to the local area network through access switches.

[0061] Furthermore, in this embodiment, the network framework of the traffic event recognition system also includes:

[0062] A traffic load balancing module is set between the network egress and the core layer, and the traffic load balancing module is configured to dynamically adjust the data transmission path according to real-time traffic. Specifically, the dynamic adjustment of the data transmission path according to real-time traffic includes:

[0063] D1-Real-time monitoring of the traffic load rate of the network egress and the traffic load rate of the core layer;

[0064] D2-Calculate the difference between the traffic load rate at the network egress and the traffic load rate at the core layer;

[0065] D3-Dynamically allocate path weights of transmission paths according to the difference and the maximum theoretical load rate of the network.

[0066] The difference is specifically: ; is the traffic load rate of the network egress, is the traffic load rate of the core layer.

[0067] The maximum theoretical network load rate is used to quantify the upper limit of the network load and is calculated using the following formula: , The total bandwidth capacity of the network egress and core layer, in Gbps, can be obtained from the network equipment specifications, such as 10 Gbps. is the network peak traffic demand, in Gbps; The network transmission delay time, in seconds, can be obtained through Ping test or network monitoring tools, such as 50ms; The task execution period is in seconds and is set according to the detection task frequency, such as 10 seconds. This formula quantifies the impact of delay on network load capacity by introducing network transmission delay time, providing a more accurate theoretical basis for load balancing.

[0068] The path weight is specifically: ; is the sensitivity coefficient, which is used to control the steepness of the weight adjustment. It can be adjusted experimentally to control the steepness of the weight adjustment, for example, ; is a threshold parameter used to set the baseline value of network load difference, such as 50% of Lmax, that is, ; The value range is 0~1, which is used to indicate the path priority.

[0069] This formula achieves a smooth transition in weight adjustment, avoiding the path switching jitter caused by traditional linear adjustment and improving network transmission stability. It also dynamically adjusts path weights based on the upper limit of the network's maximum theoretical load rate, reducing the risk of network congestion.

[0070] It is feasible that, in this embodiment, the network peak traffic demand is calculated by the following formula:

[0071]

[0072] in, is the total number of video streams; is the bitrate value of the k-th video, in Mbps; is the activity coefficient of the k-th video, ranging from 0 to 1, indicating the activity probability of the video stream per unit time; is the fluctuation amplitude coefficient, which is used to characterize the periodic fluctuation of network load and is fitted through historical traffic data analysis; is the fluctuation frequency in rad / s, which is fitted by analyzing historical flow data; is the attenuation coefficient, which is used to characterize the attenuation characteristics of fluctuations over time and is fitted through historical flow data analysis; is the current time in seconds. This formula dynamically predicts network peak traffic demand by introducing a periodic fluctuation function and an attenuation term, improving its adaptability to periodic changes in network load.

[0073] Furthermore, the activity coefficient of the k-th video is calculated using the following formula:

[0074]

[0075] in, is the active duration of the k-th video in unit time, in seconds; is the total duration of the unit time, in seconds; The activity fluctuation coefficient is used to characterize the periodic changes in the activity state of the video stream and is fitted by the historical activity data of the video stream; is the phase angle of the kth video, measured in rad, and is fitted using the historical activity data of the video stream. This formula, by introducing a periodic fluctuation function, further optimizes the dynamic scheduling strategy of network resources and improves its adaptability to periodic changes in the activity state of video streams.

[0076] In one embodiment, the dynamic scheduling of cloud computing power specifically includes:

[0077] Reduce network bandwidth consumption through cloud data connection, and dynamically allocate computing resources based on the computing power requirements of the low-bitrate patrol event detection and the high-bitrate event review detection;

[0078] Adjust the allocation ratio of computing resources based on the urgency of the traffic incident.

[0079] Specifically, the allocation ratio is adjusted according to the degree of urgency, including the following steps:

[0080] P1 - Priority Division: Low-bitrate patrol event detection is defined as a low-priority task, and high-bitrate event review detection is defined as a high-priority task. The goal of the low-priority task is to perform preliminary event detection using low-bitrate (<1Mbps) video, covering all video channels, but only marking suspected events. The computing power requirements are: low computing power requirements for single-channel low-bitrate video detection. Features include: a large workload (covering all video channels), but low resource consumption for single-channel tasks, and no need for high-precision recognition. The goal of the high-priority task is to pull high-bitrate (≥1Mbps) videos of suspected events in low-bitrate patrol event detection for review, and output the final recognition results. The computing power requirements are: high computing power requirements for single-channel high-bitrate video detection. Features include: a small workload (only for suspected events), but high resource consumption for single-channel tasks, and high-precision recognition.

[0081] P2-priority-driven allocation: high-priority tasks take up computing resources first to ensure that their computing power requirements are met; low-priority tasks are allocated on demand from the remaining computing power.

[0082] Furthermore, the dynamic allocation of computing resources specifically includes:

[0083] The number of video channels detected according to the low-bitrate patrol event And the corresponding code stream , and the number of video channels for high-stream event review and detection And the corresponding code stream , dynamically calculate the total computing power requirements Total computing power requirements Calculated by the following formula:

[0084]

[0085] in, The unit bitrate computing power coefficient of low bitrate video, in TFLOPS / (Mbps); The unit bitrate computing power coefficient of high bitrate video, in TFLOPS / (Mbps); is the actual bitrate of the i-th low-bitrate video, in Mbps; is the actual bitrate of the j-th high-bitrate video, in Mbps; and It is a nonlinear adjustment index used to modify the computing power requirements for low and high bitrates, respectively. The computing power demand curves under different bitrate scenarios are fitted through experiments. is the priority weighting factor of the j-th high-bitrate video, ranging from 0 to 1, and is dynamically adjusted according to the urgency of the task, such as 0.9 for accident detection and 0.5 for illegal parking detection. This formula introduces a nonlinear adjustment index and , realizing dynamic nonlinear modeling of computing power requirements, such as when the computing power requirements of low-bitrate tasks increase nonlinearly with the increase of bitrate, improving the adaptability to extreme bitrate scenarios, and the priority weighting factor , ensuring that high-urgency tasks get priority computing power and avoiding waste of resources.

[0086] When adjusting the allocation ratio of computing resources, the computing power allocation ratio is dynamically adjusted according to the real-time task queue (the number of low-bitrate patrol event detection and high-bitrate event review detection tasks). For example, when the number of high-bitrate event review detection tasks surges, the system automatically increases the computing power allocation ratio of high-bitrate tasks (specifically by increasing ), reduce the number of concurrent paths for low-bitrate patrol event detection.

[0087] Furthermore, the unit code rate computing coefficient of the low code rate video and the unit code rate computing coefficient of the high code rate video are calculated based on the data of historical detection tasks to calculate the average detection time of the low code rate video. and the average detection time of high bitrate video Get, specifically:

[0088]

[0089]

[0090] in, The average bitrate of low-bitrate videos, in Mbps. The average bitrate of high-bitrate videos, in Mbps. and is a correction coefficient used to compensate for the differences in algorithm complexity of different bitstream videos. Generally, through experimental verification, it is feasible that the value range of the correction coefficient is: , , and in the dynamic allocation of computing resources, the value satisfies ; It is a dynamic compensation factor used to balance the relationship between detection time consumption and computing power allocation, and is adjusted according to the changes in algorithm complexity, such as This formula, by introducing a dynamic compensation factor, achieves adaptive correction of computing power allocation, balancing detection time and computing power allocation, and improving adaptability to changes in algorithm complexity. Secondly, the correction coefficient is verified through experiments to ensure that the computing power allocation for low-bitrate videos is not excessively underestimated due to differences in algorithm complexity, while also preventing the computing power requirements for high-bitrate videos from being overestimated.

[0091] In summary, this embodiment of the traffic incident secondary identification method based on video cloud platform streaming detection only requires high-bitrate event verification for videos initially identified as traffic incidents, significantly reducing computing resource overhead. Furthermore, through a series of matching designs such as dynamic computing power allocation, traffic load balancing, and priority-driven scheduling, it ensures rational resource allocation, effective utilization, efficient response, and smooth network operation.

[0092] In the embodiments provided herein, it should be understood that the embodiments described herein can be implemented using hardware, software, firmware, middleware, code, or any appropriate combination thereof. For hardware implementation, the processor may be implemented in one or more of the following: an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field-programmable gate array (FPGA), a processor, a controller, a microcontroller, a microprocessor, or other electronic units designed to implement the functionality described herein, or any combination thereof. For software implementation, some or all of the processes of the embodiments may be performed by a computer program instructing the relevant hardware. During implementation, the program may be stored in a computer-readable storage medium or transmitted as one or more instructions or codes on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein communication media includes any medium that facilitates the transmission of a computer program from one location to another. The storage medium may be any available medium that can be accessed by a computer. Computer-readable storage media may include, but are not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing the desired program code in the form of instructions or data structures and accessible by a computer.

[0093] Finally, it should be noted that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent replacements for some of the technical features therein. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A secondary traffic incident identification method based on video cloud platform streaming detection, characterized in that: The method comprises the following steps: S1 - Pull low-bitrate video data from the cloud, perform low-bitrate patrol event detection through the traffic event recognition system, and output preliminary detection results; the network framework of the traffic event recognition system includes: The provincial video cloud platform deployed in the cloud is configured to: communicate with local business systems via IPsec VPN or SSL VPN; The network egress is configured as follows: using firewalls and routers as egress devices, with security policies and routing policies; The core layer is configured to: achieve high-speed data transmission through core switches; The access layer is configured to connect application servers, data servers, and distributed storage devices through access switches; and, providing a traffic load balancing module between the network egress and the core layer, wherein the traffic load balancing module is configured to: dynamically adjust the data transmission path according to real-time traffic; S2-If the preliminary detection result is a traffic incident, the high-bitrate video data corresponding to the low-bitrate video data is pulled, and the high-bitrate event is rechecked and detected by the traffic incident recognition system, and the final recognition result is output; The operation of the traffic event recognition system is achieved by sharing cloud resources through dynamic scheduling of cloud computing power. The dynamic scheduling of cloud computing power specifically includes: Reduce network bandwidth consumption through cloud data connection, and dynamically allocate computing resources based on the computing power requirements of the low-bitrate patrol event detection and the high-bitrate event review detection; Adjust the allocation ratio of computing resources based on the urgency of traffic events; The dynamic adjustment of the data transmission path according to the real-time traffic flow specifically includes: D1-Real-time monitoring of the traffic load rate of the network egress and the traffic load rate of the core layer; D2-Calculate the difference between the traffic load rate at the network egress and the traffic load rate at the core layer; D3-dynamically assigns the path weight of the transmission path according to the difference and the maximum theoretical load rate of the network; The maximum theoretical load rate of the network is calculated using the following formula: in, The total bandwidth capacity of the network egress and core layer, in Gbps; is the network peak traffic demand, in Gbps; It is the network transmission delay time; is the task execution cycle; The difference is specifically: ; in, is the traffic load rate of the network egress, is the traffic load rate of the core layer; The path weight is specifically: ; in, is the sensitivity coefficient; is the threshold parameter; The network peak traffic demand is calculated using the following formula: in, is the total number of video streams; is the bitrate value of the k-th video; is the activity coefficient of the k-th video, ranging from 0 to 1, indicating the activity probability of the video stream per unit time; is the fluctuation amplitude coefficient, which is used to characterize the periodic fluctuation of network load and is fitted through historical traffic data analysis; The fluctuation frequency is fitted through historical traffic data analysis; is the attenuation coefficient, which is used to characterize the attenuation characteristics of fluctuations over time and is fitted through historical flow data analysis; is the current time in seconds; The activity coefficient of the k-th video is calculated using the following formula: in, is the active duration of the kth video per unit time; is the total duration of the unit time; The activity fluctuation coefficient is used to characterize the periodic changes in the activity state of the video stream and is fitted by the historical activity data of the video stream; is the phase angle of the k-th video, which is fitted by the historical activity data of the video stream.

2. The secondary traffic incident identification method based on video cloud platform streaming detection according to claim 1 is characterized in that: The low bit rate is lower than 1 Mbps; the high bit rate is higher than 1 Mbps.

3. The secondary traffic incident identification method based on video cloud platform streaming detection according to claim 1 is characterized in that: The dynamic allocation of computing resources specifically includes: The number of video channels detected according to the low-bitrate patrol event And the corresponding code stream , and the number of video channels for high-stream event review and detection And the corresponding code stream , dynamically calculate the total computing power requirements Total computing power requirements Calculated by the following formula: in, The unit bitrate computing power coefficient of low bitrate video; The unit bitrate computing power coefficient of high bitrate video; is the actual bitrate value of the i-th low-bitrate video; is the actual bitrate value of the j-th high-bitrate video; and is the nonlinear adjustment index; is the priority weighting factor of the j-th high-bitrate video.

4. The secondary traffic incident identification method based on video cloud platform streaming detection according to claim 3 is characterized in that: The unit code rate computing coefficient of the low code rate video and the unit code rate computing coefficient of the high code rate video are calculated based on the data of historical detection tasks, and the average detection time of the low code rate video is calculated. and the average detection time of high bitrate video Get, specifically: in, is the average bitrate value of low bitrate video; is the average bitrate value of high bitrate video; and is the correction factor; is the dynamic compensation factor.

5. The secondary traffic incident identification method based on video cloud platform streaming detection according to claim 4 is characterized in that: The value range of the correction coefficient is: , , and in the dynamic allocation of computing resources, the value satisfies .

Images