CN115767024A - Bank branch monitoring video analysis system and method based on narrow-band network - Google Patents

Abstract

The invention discloses a bank outlet monitoring video analysis system based on a narrow-band network. The network monitoring cameras are arranged at all bank outlets and are communicated with the monitoring center through narrow-band network private lines to upload pictures. The first server is arranged in the monitoring center and used for instructing the network monitoring camera to carry out inspection and image capturing on the monitoring picture according to a preset image capturing strategy, and the second server is arranged in the monitoring center and used for analyzing the captured picture according to monitoring analysis requirements. The system adopts a video analysis server cluster based on the cloud, and sets a snapshot strategy according to monitoring analysis requirements, so that the existing monitoring equipment and narrow-band network of the bank outlets are not required to be modified, and further the intelligent modification of the video monitoring of the bank outlets can be completed with extremely low construction cost and equipment cost.

Description

Bank branch monitoring video analysis system and method based on narrow-band network

Technical Field

The invention relates to the technical field of communication, in particular to a system and a method for analyzing a monitoring video of a bank outlet based on a narrow-band network.

Background

The banking outlets are mainly used for handling various financial services for customers, and the requirements of the banking outlets on safety are difficult to meet only by means of manual inspection, so that the banking outlets usually deploy video monitoring systems and establish standard requirements of video monitoring of related financial industries.

Currently, a Network Camera (IP Camera) and a Network Video Recorder (NVR) are generally used for Video monitoring of business outlets of a bank. Based on the standard requirements of video monitoring in the financial industry, a single bank business office usually has tens of networked network cameras and a plurality of NVRs. In order to find abnormal behaviors in time and eliminate potential safety hazards, it is necessary to realize the intellectualization of the video monitoring system on the basis.

Currently, the intelligent modification of a video monitoring system is mainly realized by a method of replacing hardware equipment, specifically, an IP Camera with an AI function or an NVR with an AI function is adopted to replace a traditional IP Camera or NVR, and intelligent analysis is realized through a built-in AI algorithm. However, AI built in with cameras or video recorders is often computationally weak and the AI algorithms supported in one camera or video recorder are of a single kind, so they cannot support the more complex intelligent analysis needs of banks. In addition, the mode of changing hardware equipment is adopted to carry out intelligent transformation, and construction and equipment deployment need to be carried out to each business outlet. A bank in a city usually includes dozens or even hundreds of business sites, and its monitoring cameras often include thousands of routes, so that the site construction can introduce a great amount of engineering. Meanwhile, the network camera which is just transformed digitally is abandoned and replaced for a long time, and great waste is caused on material resources.

Disclosure of Invention

Aiming at the problems of part or all of the prior art, the invention firstly provides a bank branch monitoring video analysis system based on a narrow-band network, which comprises the following steps:

the network monitoring camera is arranged at a bank outlet and used for shooting a monitoring picture of the bank outlet and communicating with the monitoring center through a narrow-band network private line so as to upload one or more pictures in the monitoring picture;

the first server is arranged in the monitoring center and used for instructing the network monitoring camera to carry out routing inspection and picture capturing on a monitoring picture according to a preset picture capturing strategy so as to obtain the one or more pictures; and

and the second server is arranged in the monitoring center and used for analyzing the one or more pictures according to monitoring analysis requirements so as to identify the specified event.

Further, the bandwidth of the narrow-band network private line is less than 10Mps.

Further, the second server comprises a graphic image processing chip and a deep learning chip, and the second server comprises at least one algorithm.

Further, the second server includes a plurality of servers, each of which includes a different algorithm.

Further, the system further comprises a display module, which is arranged in the monitoring center and used for displaying the analysis result of the second server.

Further, the monitoring analysis requirements includes: equipment inspection and confirmation, linkage door opening detection, counter cash stop detection, workflow specification detection, cash transport and handover flow detection, equipment damage identification, uniform identification personnel on duty, image quality detection, article leaving identification, malicious advertisement identification, overtime retention identification and voice identification.

Based on the system, the invention further provides a method for analyzing the monitoring video of the bank outlets based on the narrow-band network, which comprises the following steps:

instructing a network monitoring camera to carry out inspection and picture capturing on a monitoring picture according to a preset picture capturing strategy through a first server so as to obtain one or more pictures;

according to monitoring analysis requirements of different areas, distributing the one or more pictures to different second servers for analysis; and

and the second server transmits the structural information and the picture obtained by analysis back to the first server.

Further, the grab strategy comprises:

the number of pictures uploaded to the first server per second is determined according to the bandwidth of the narrow-band network private line, the image transmission bearing coefficient and the picture resolution; and

and the image capturing frequency of different areas of the bank outlets is determined according to the security level and the monitoring analysis requirements of each area.

Further, the grab strategy further comprises:

the frequency of the grab pictures at different time periods is different for the same area.

Further, the grab strategy further comprises:

the frequency of the grab-chart is different for the same area in the working time period and the non-working time period.

Further, the grab strategy further comprises:

and increasing the image capturing frequency of the designated area in the designated time period.

Further, the grab strategy further comprises:

and when a preset event is detected, increasing the image capture frequency of the corresponding area.

Further, the grab strategy further comprises:

and adjusting the resolution of the captured picture according to the monitoring analysis requirement.

Further, the second server returns data in a Restful API format.

The bank outlet monitoring video analysis system and method based on the narrow-band network provided by the invention adopt the video analysis server cluster based on the cloud, further do not need to modify the existing bank outlet monitoring equipment, and simultaneously can support more complex algorithms, can support multiple algorithms and support the series analysis of different algorithms compared with a camera or a video recorder with an AI (advanced information acquisition) device. The system and the method adopt a static picture capturing mode to replace video streaming transmission, can effectively reduce the bandwidth required by data transmission, and further enable the original narrow-band network to be continuously used without equipment upgrading or replacing. In addition, the method sets a capture strategy according to monitoring analysis requirements of different areas of a bank outlet, can also quickly acquire multi-channel monitoring real-time images under low bandwidth, and improves the system response real-time property. In conclusion, the system and the method can complete the intelligent improvement of video monitoring of each business outlet with extremely low construction cost and equipment cost.

Drawings

To further clarify the above and other advantages and features of various embodiments of the present invention, a more particular description of various embodiments of the invention will be rendered by reference to the appended drawings. It is appreciated that these drawings depict only typical embodiments of the invention and are therefore not to be considered limiting of its scope. In the drawings, the same or corresponding parts will be denoted by the same or similar reference numerals for clarity.

Fig. 1 is a schematic structural diagram of a monitoring video analysis system of a banking outlet based on a narrowband network according to an embodiment of the present invention; and

fig. 2 is a schematic flow chart of a monitoring video analysis method of a banking outlet based on a narrowband network according to an embodiment of the present invention.

Detailed Description

In the following description, the present invention is described with reference to various embodiments. One skilled in the relevant art will recognize, however, that the embodiments can be practiced without one or more of the specific details, or with other alternative and/or additional methods or components. In other instances, well-known structures or operations are not shown, or are not described in detail, to avoid obscuring aspects of the invention. Similarly, for purposes of explanation, specific numbers and configurations are set forth in order to provide a thorough understanding of the embodiments of the invention. However, the invention is not limited to these specific details.

Reference in the specification to "one embodiment" or "the embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the invention. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment.

It should be noted that the method steps are described in a specific order according to the embodiments of the present invention, which is only for illustrating the specific embodiments and not for limiting the sequence of the steps. On the contrary, in different embodiments of the present invention, the sequence of the steps can be adjusted according to the adjustment of actual requirements.

If the mode of replacing hardware equipment is adopted, the intelligent reconstruction construction engineering quantity of the bank outlets is extremely large. Meanwhile, limited by the size of the device, only one AI algorithm can be usually built in the camera or NVR, and then only one monitoring and analyzing task, such as face recognition, people counting, license plate recognition, behavior analysis, thermal imaging recognition or intrusion detection, can be completed. For the bank outlets, the requirements for monitoring and analyzing are often complex, and the number of required devices is also large if the monitoring and analyzing requirements of the bank outlets are to be met. In addition, in the existing bank monitoring system, a narrow-band network dedicated line communication with a bandwidth less than 10Mbps is usually adopted between a bank outlet and a monitoring center, and the number of videos or pictures which can be uploaded at each moment is extremely limited, so that if the real-time performance of monitoring intelligent analysis is ensured, a communication network needs to be further modified when the intelligent modification of the bank outlet is realized by replacing hardware equipment, and the modification cost is further increased. Based on the above, the invention provides a monitoring video analysis system of a bank outlet based on a narrowband network, which does not need to replace the existing bank monitoring system, but deploys a second server in a monitoring center for video analysis, and simultaneously designs a snap-in strategy aiming at the problem of small bandwidth so as to fully utilize the picture which can be uploaded at each moment and cover the monitoring analysis requirements of each area in the bank outlet.

The solution of the invention is further described below with reference to the accompanying drawings of embodiments.

Fig. 1 shows a schematic structural diagram of a monitoring video analysis system of a banking outlet based on a narrowband network according to an embodiment of the present invention. As shown in fig. 1, a monitoring video analysis system of a banking outlet based on a narrowband network includes a network monitoring camera 101, a first server 102, and a second server 103. The network monitoring camera 101 is deployed at a bank outlet, and a plurality of network monitoring cameras and/or NVRs may be respectively disposed in different areas of each bank outlet to monitor each area in real time. The first server 102 and the second server 103 are deployed in a monitoring center, and the banking outlets and the monitoring center realize communication through a narrow-band network dedicated line.

In an embodiment of the present invention, for example, a branch switch may be set at each bank branch, and the branch switch, a dedicated router of a monitoring center and a central switch together form a network architecture between the bank branch and the monitoring center, so as to implement data communication between the bank branch and the monitoring center, and further, the first server 102 may instruct a network monitoring camera and/or NVR of each bank branch to perform inspection and snapshot on the shot monitoring video according to a preset snapshot policy, and send the inspection and snapshot to the second server 103 for analysis.

The first server 1021 is mainly used for instructing business outlets of all banks to perform inspection and grab pictures of monitoring pictures, and bearing a monitoring platform for presenting results. In an embodiment of the present invention, the first server 102 may adopt, for example, a hua cheng kunpeng server, which is used as a branch general server cluster and directly communicates with a central switch set in a monitoring center, on one hand, the network monitoring camera 101 performs a capture of a monitoring video according to a preset capture policy instruction, and on the other hand, the first server performs data interaction with other devices of the monitoring center, such as the second server 103, a display module, and an office computer, for example, the captured image is sent to a designated second server according to monitoring analysis requirements, an analysis result of the second server is received, and the captured image and/or the analysis result is sent to the display module for display, or is reported to an public security system through a dedicated router.

In an embodiment of the present invention, the snapshot policy mainly includes the number of pictures uploaded to the first server per second, and snapshot frequencies of different areas of the banking outlet.

The number of pictures which can be transmitted per second, namely the upper limit of the number of pictures uploaded to the first server, can be determined according to the bandwidth of the narrow-band network private line, the image transmission bearing coefficient and the picture resolution:

the number of pictures that can be transmitted per second = (bandwidth of narrowband network dedicated line =imagetransfer bearer coefficient)/size of specific resolution picture file,

wherein:

the image transmission bearing coefficient refers to the proportion of the bandwidth of the narrow-band network private line which can be used for image transmission; and

the size of the specific resolution picture file is calculated as follows: the file size of a single original uncompressed coded JPEG picture = (picture wide pixel value x picture high pixel value x pixel width)/8 bytes. In practical application, under the narrow-band condition of the bank business network, the following picture resolutions are usually adopted for capturing pictures, and by combining with the JPEG compression format, the loss compression is 90%, and the size of the corresponding single picture file is as follows:

1280x720 resolution (high definition): 90KB or 720Kbit;

720x576 resolution (standard definition): 40KB or 320Kbit; and

352x288 resolution (CIF format): 10KB or 80Kbit, or a combination thereof,

for example, if the bandwidth of a certain business outlet is 4Mbps, the picture transmission carrying factor is 0.5, and 720 × 576 resolution is selected for transmission, the number of pictures that can be transmitted per second is about 6.

The image capturing frequency of different areas of the bank outlets is determined according to the security level and the monitoring analysis requirement of each area of the bank outlets, and the image capturing frequency of different areas is determined according to the security level and the monitoring analysis requirement, so that the distribution of the captured images in unit time is reasonably distributed. In addition, according to the service characteristics of the banking outlets, different grapping frequencies can be further set for the same area in different time periods, for example, different grapping frequencies are set in working time periods and non-working time periods, or the grapping frequency of the corresponding area is adjusted in a specified time period or when a specified event occurs.

In one embodiment of the invention, the frequency of the snapshot of different areas of the banking outlet can be performed according to the following method:

the levels of the different regions are determined and the frequency of the grab images at the different levels is determined, e.g. 1 grab image every 30 seconds in a low frequency scan scenario, 1 grab image every 5 or 10 seconds in a medium frequency scan scenario, and 1 grab image 1 every 1 second in a high frequency scan scenario. Specifically, different areas of a bank outlet can be divided into a low-frequency scanning scene, a medium-frequency scanning scene and a high-frequency scanning scene, the capture frequency is different at different levels, and in order to better cover the monitoring and analysis requirements of the different areas, different levels can be further set for the areas at different time periods or states according to time intervals or specific events:

the strategy according to the time period is that different grades are set for corresponding areas according to whether the areas are in the working time period or the special event occurrence time period, for example, the grade of the areas such as a business hall and the like is higher than that of the areas in the non-working time period in the working time in the daytime, so that the scanning and image-capturing frequency of the cameras of a specific combination is increased in the working time in the daytime, and the grade of the non-working time period in the night in the areas such as an entrance, a special channel and a linkage door is higher than that in the working time; for example, in the arrival time period of the securicar, the machine room inspection time period and the like, the grades of corresponding areas such as the securicar parking area, the securicar adding room, the machine room and the like can be improved; and

the strategy according to the specific event refers to a picture-grabbing strategy triggered by a specified event, for example, after the arrival of the cash carrier is detected, the grade of a cash box along the area is improved, so that the picture-grabbing frequency of the camera in the corresponding area is improved, and for example, after the existence of people in an ATM self-service area is detected, the grade of the self-service equipment area is improved, so that the picture-grabbing frequency of the related camera is improved, and the like.

In addition, in order to further save bandwidth, in an embodiment of the present invention, the capture strategy further includes dynamically adjusting the resolution of captured pictures according to different service requirements. For example, for a pinhole camera of an ATM, the face pictures taken by the pinhole camera are taken in a short distance, and if face recognition is needed, 720x576 resolution is adopted, and if only the number of people is detected, a CIF resolution format is adopted.

Taking a certain network point of the postal savings bank as an example, 39 cameras are arranged in total, images are captured by adopting 720x576 resolution, about 2 times per second, 1Mbps bandwidth is occupied, and the capture frequency is shown in Table 1:

TABLE 1

The second server 103 is mainly used for video decoding and/or analyzing the captured pictures based on a computer vision algorithm and a deep learning algorithm so as to identify a specified event and meet the monitoring and analyzing requirements of different areas of a bank outlet. In an embodiment of the present invention, an algorithm may be deployed in the second server 103 to implement the following monitoring and analyzing requirements: equipment inspection and confirmation, linkage door opening detection, counter cash stop detection, workflow specification detection, cash transport and handover flow detection, equipment damage identification, uniform identification personnel on duty, image quality detection, article leaving identification, malicious advertisement identification, overtime retention identification, voice identification and the like.

In an embodiment of the present invention, the second server 103 implements multiple algorithms by internally setting a dedicated hardware, such as a chip or a board, for the graphic image processing and the deep learning acceleration. In particular, virtualization techniques, such as containers, can be employed to implement the deployment of the different algorithms.

In an embodiment of the present invention, the system includes a plurality of second servers, and each second server is deployed with a specific one or more algorithms, so that different algorithms can be selected according to actual requirements, and a series analysis between different algorithms is performed. Specifically, the platform software located in the first server 102 acquires a picture according to a service requirement, and then distributes the picture to a specific second server 103 according to the type of an algorithm, and if a single picture needs to be analyzed in multiple ways, distributes the picture to multiple second servers at the same time. After the second server completes the analysis, the analyzed picture and the structural information are transmitted back to the platform software on the first server 102 along with the picture. In one embodiment of the invention, the second server returns data in Restful API format for easy expansion.

Based on the foregoing system, fig. 2 shows a schematic flow chart of a monitoring video analysis method of a banking outlet based on a narrowband network according to an embodiment of the present invention, and as shown in fig. 2, a monitoring video analysis method of a banking outlet based on a narrowband network includes:

first, in step 201, a picture is captured. Instructing a network monitoring camera to carry out routing inspection and picture capture on a monitoring picture according to the picture capture strategy through a first server;

next, at step 202, the picture is distributed. And according to the monitoring analysis requirements of different areas, the captured pictures are distributed to different second servers for analysis. The first server distributes the captured pictures to a specific second server according to the type of the needed algorithm, and if a single picture needs to be analyzed in multiple ways, the single picture is distributed to a plurality of second servers at the same time; and

finally, in step 203, the picture is analyzed. And the second server analyzes the received picture, and transmits the structural information and the picture obtained by analysis back to the first server by adopting a Restful API format so as to perform subsequent display or upload a public security system.

The bank outlet monitoring video analysis system and method based on the narrow-band network provided by the invention adopt the video analysis server cluster based on the cloud, further do not need to modify the existing bank outlet monitoring equipment, and simultaneously can support more complex algorithms, can support multiple algorithms and support the series analysis of different algorithms compared with a camera or a video recorder with an AI (advanced information acquisition) device. The system and the method adopt a static picture capturing mode to replace video stream transmission, can effectively reduce the bandwidth required by data transmission, and further enable the original narrow-band network to be continuously used without equipment upgrading or replacing. In addition, the method sets a capture strategy according to monitoring analysis requirements of different areas of a bank outlet, can also quickly acquire multi-channel monitoring real-time images under low bandwidth, and improves the system response real-time property. In conclusion, the system and the method can complete the intelligent improvement of video monitoring of each business outlet with extremely low construction cost and equipment cost.

While various embodiments of the present invention have been described above, it should be understood that they have been presented by way of example only, and not limitation. It will be apparent to persons skilled in the relevant art that various combinations, modifications, and changes can be made thereto without departing from the spirit and scope of the invention. Thus, the breadth and scope of the present invention disclosed herein should not be limited by any of the above-described exemplary embodiments, but should be defined only in accordance with the following claims and their equivalents.

Claims (14)

1. A bank outlet monitoring video analysis system based on a narrow-band network is characterized by comprising:

the network monitoring camera is arranged at a bank outlet and is configured to shoot a monitoring picture of the bank outlet and communicate with the monitoring center through a narrow-band network private line so as to upload one or more pictures of the monitoring picture;

the system comprises a first server, a second server and a monitoring center, wherein the first server is arranged in the monitoring center and is configured to instruct a network monitoring camera to carry out inspection grab drawing on a monitoring picture according to a preset grab drawing strategy so as to obtain one or more pictures; and

and the second server is arranged in the monitoring center and is configured to analyze the one or more pictures according to monitoring analysis requirements and return an analysis result to the first server.

2. The bank outlet surveillance video analysis system of claim 1, wherein the bandwidth of the narrowband network private line is less than 10Mps.

3. The system of claim 1, wherein the second server comprises a graphic image processing chip and a deep learning chip, and the second server comprises at least one algorithm.

4. A banking site surveillance video analysis system according to claim 3, wherein the second server includes a plurality of servers, and each second server includes a different algorithm therein.

5. The system of claim 1, further comprising a display module disposed in a monitoring center, communicatively coupled to the first server, and configured to display the analysis result of the second server.

6. The bank outlet surveillance video analysis system of claim 1, wherein the surveillance analysis requirements include: equipment inspection and confirmation, linkage door opening detection, counter cash stop detection, workflow specification detection, cash transport and handover flow detection, equipment damage identification, uniform identification personnel on duty, image quality detection, article leaving identification, malicious advertisement identification, overtime retention identification and voice identification.

7. A monitoring video analysis method of a bank outlet based on a narrow-band network is characterized by comprising the following steps:

instructing a network monitoring camera to carry out routing inspection picture capture on a monitoring picture according to a preset picture capture strategy through a first server so as to obtain one or more pictures;

according to monitoring analysis requirements of different areas, the one or more pictures are distributed to different second servers for analysis; and

and the second server transmits the structural information and the pictures obtained by analysis back to the first server.

8. The method for analyzing surveillance video at a banking site as recited in claim 7, wherein the snap-down strategy comprises:

the number of pictures uploaded to the first server every second is determined according to the bandwidth of the narrow-band network private line, the image transmission bearing coefficient and the picture resolution; and

and the image capturing frequency of different areas of the bank outlets is determined according to the security level and the monitoring analysis requirements of each area.

9. The banking outlet surveillance video analytics method as claimed in claim 8, wherein said snap strategy further comprises:

the frequency of the grab pictures at different time periods is different for the same area.

10. The method for analyzing surveillance video at a banking site as recited in claim 8, wherein the snap strategy further comprises:

the frequency of the grab pictures is different in the working time period and the non-working time period for the same area.

11. The banking outlet surveillance video analytics method as claimed in claim 8, wherein said snap strategy further comprises:

and increasing the image capturing frequency of the designated area in the designated time period.

12. The method for analyzing surveillance video at a banking site as recited in claim 8, wherein the snap strategy further comprises:

and when the preset event is detected, increasing the image capture frequency of the corresponding area.

13. The banking outlet surveillance video analytics method as claimed in claim 8, wherein said snap strategy further comprises:

and adjusting the resolution of the captured picture according to the monitoring analysis requirement.

14. A banking outlet monitoring video analysis method according to claim 7, wherein said second server returns data in Restful API format.

Images