TWI448975B - Dispersing-type algorithm system applicable to image monitoring platform - Google Patents

Description

Decentralized computing system applied to image monitoring platform

The present invention relates to a distributed computing system, and more particularly to a distributed computing system applied to an image monitoring platform.

In busy urban life, in order to achieve the purpose of maintaining law and order, many specific locations will have one or more image monitoring devices for remote monitoring, such as intersection image surveillance systems or license plate recognition systems.

An operator who provides remote monitoring services for consumers can first build an image monitoring platform that can be connected to a plurality of image monitoring devices. When the image monitoring platform receives image data from a plurality of image monitoring devices, The image data is transmitted to the consumer's smart phone. Of course, consumers can also use their smart phones to dial a specific door number to enter the image monitoring platform, and then actively obtain images captured by the image monitoring device. Nowadays, various operators are also competing to set up another server in their built image monitoring platform to implement the increasingly mature intelligent image recognition and analysis technology, so as to use the intelligent image recognition and analysis technology in the received image data. Identify specific events and proactively inform consumers when they identify specific events.

However, intelligent image recognition and analysis technology requires a large amount of computing resources, and a single server is often unable to fully load. Moreover, due to cost considerations, most operators are unable to set up multiple servers simultaneously in the image monitoring platform for intelligent image recognition and analysis. On the other hand, the image monitoring platform is mostly a cloud platform. In order to provide stable performance, the operators will plan a hypervisor in the image monitoring platform. Once the server is used to implement intelligent image recognition and analysis. The original processing performance of the image monitoring platform was also seriously affected by the interference of the management program.

In view of the various deficiencies of the prior art, the object of the present invention is to provide a distributed computing system that enables an image monitoring platform to implement intelligent image recognition and analysis without affecting the original performance.

In order to achieve the above and other objects, the present invention provides a distributed computing system for an image monitoring platform, comprising: a command module, configured to cut image data received by the image monitoring platform into a plurality of image units; And a plurality of proxy modules connected to the command module and configured to receive a plurality of image units cut by the command module for intelligent image recognition and analysis, and the proxy modules are identified and analyzed in the intelligent image. Upon completion, the identified and analyzed results are passed back to the command module for integration by the command module.

In an embodiment of the present invention, the command module detects the processing performance of each proxy module, and determines the number of cuts of the image unit according to the detected processing performance of each proxy module. The command module sends a larger number of image units to the proxy module with higher processing efficiency.

Compared with the prior art, the present invention constructs a plurality of proxy modules for intelligent image recognition and analysis in the original image monitoring platform, and is connected with a plurality of proxy modules and automatically according to the proxy module. The idle performance of the command module for cutting, distributing, and integrating image data, so that it is not necessary to require additional hardware devices for the image monitoring platform, and it will not affect the processing performance originally provided by the image monitoring platform.

The other embodiments of the present invention will be readily understood by those skilled in the art from this disclosure. Of course, the invention may be embodied or applied by other different embodiments.

Please refer to FIG. 1 to understand the system architecture of the distributed computing system applied to the image monitoring platform of the present invention. As shown, the distributed computing system 1 includes a command module 10 and a plurality of agent modules 11, 12, 13, 14 coupled to the command module 10.

The proxy modules 11, 12, 13, and 14 are used for intelligent image recognition and analysis, and can be selectively connected to the command module 10 by transmitting a broadcast packet. For example, the proxy modules 11, 12, 13, 14 can be set to automatically send broadcast packets every 20 seconds, and the command module 10 can be set to automatically and the proxy modules 11, 12, 13 after receiving the broadcast packets. And 14 establish a connection.

The command module 10 can cut the image data received by the image monitoring platform into a plurality of image units, and send the cut plurality of image units to the proxy modules 11, 12, 13, and 14 to make the proxy module 11, 12, 13, and 14 perform intelligent image recognition and analysis on the received image unit according to a specific algorithm. After the intelligent image recognition and analysis is completed, the agent modules 11, 12, 13, and 14 will transmit the identified and analyzed results to the command module 10 for integration by the command module 10.

Specifically, the command module 10 can detect the processing performance of the proxy modules 11, 12, 13, and 14 when the connection is established, that is, the available proxy modules 11, 12, 13, and 14 can be utilized. Resources. Then, according to the processing performance of the detected proxy modules 11, 12, 13, and 14, the image data received by the image monitoring platform is cut into a plurality of image units. Thereafter, the command module 10遂 can transmit the processed plurality of image units to the proxy modules 11, 12, 13, and 14 respectively according to the processing performance of the detected proxy modules 11, 12, 13, and 14. After the proxy modules 11, 12, 13, and 14 receive the image unit and perform intelligent identification and analysis, the respective obtained results can be transmitted back to the command module 10 for integration. Preferably, the results obtained by the intelligent identification and analysis of the proxy modules 11, 12, 13, and 14 may include an event index that can be linked to the image.

For example, if the command module 10 detects that the processing performance of the proxy modules 11, 12, 13, and 14 is 4, 2, 1, and 2 units, respectively, the command module 10 can display the image. The image data received by the monitoring platform is correspondingly cut into 9 image units, and according to the processing performance of the proxy modules 11, 12, 13, and 14, respectively, 4, 2, 1, and 2 image units are sent to The proxy modules 11, 12, 13, and 14, in other words, determine the number of cuts of the image unit according to the detected processing performance of each proxy module, and determine the processing of the image unit according to the processing performance of each proxy module. The number, so as to avoid the overloaded proxy module is always busy or idle state of the proxy module is always idle, and then the proxy module 11, 12, 13, 14 according to a specific algorithm respectively to the received image unit For intelligent image recognition and analysis, the specific algorithm can be different depending on the monitored object or use, such as face recognition or license plate recognition.

After the agent module 11, 12, 13, 14 finds a specific event after intelligently identifying and analyzing the received image unit, the event can be indexed and the result containing the index can be returned to the event. The command module 10 is configured for the command module 10 to integrate a plurality of results. For example, the image monitoring platform is a license plate recognition system. When the agent module is found from the image unit according to the license plate number indicated by the command module When the image segment is matched, the image segment is marked with an event, for example, indexed, so that the identification result can be indexed, and the subsequent processing is facilitated, that is, it is easy to know when the occurrence occurs by marking. What happened or the object to be searched for at what time. Secondly, the algorithms for the intelligent identification and analysis of the proxy modules 11, 12, 13, and 14 can be specified by the command module 10, that is, the proxy modules 11, 12, 13, and 14 can be based on the command mode. The algorithm specified by group 10 performs intelligent image recognition and analysis on the received image unit. Of course, the command module 10 can also replace the agent modules 11, 12, 13, and 14 for intelligent identification and analysis.

Furthermore, the command module 10 can also dynamically detect the status of the agent modules 11, 12, 13, 14 to detect whether an abnormality has occurred in the agent modules 11, 12, 13, 14 and detect an agent. When the status of the module is abnormal, the command module 10 can resend the image unit originally sent to the proxy module whose status is abnormal to another agent module that does not have an abnormality. For example, if the command module 10 detects that the status of the agent module 11 is abnormal, the command module 10 can reassign the image unit originally sent to the agent module 11 to other agent modules to maintain the distributed operation. System 1 runs smoothly as a whole.

On the other hand, when more proxy modules are added, the new proxy module can also choose to easily find the command module 10 by sending a broadcast packet, thereby becoming a member of the distributed computing system 1.

It should be added that the image monitoring platform for the distributed computing system 1 to process the image data source can be an image monitoring platform built on the Internet Service Provider (ISP) side, and can include network collection ( NC) a hardware structure such as a server, an event server, a relay server, an alarm server, a newsletter or a mail server, and an event database, and the command module provided by the present invention 10 and the proxy modules 11, 12, 13, and 14 can also be constructed in a logical planning manner. In other words, the command module 10 and the proxy modules 11, 12, 13, and 14 can be located in the NC server and the Event server. , Relay server, Alarm server, SMS or mail server, and Event database. In actual operation, the distributed computing system 1 of the present invention can be set to operate in the background of the image monitoring platform.

In addition, when the command module 10 cuts the image data into the image unit, in order to ensure that the correctness of the recognition is not affected by the cutting of the image data, the plurality of image units may be formed into a pattern in which the head and the tail overlap, thereby allowing the first The end of one image unit and the beginning of the second image unit have the same content, and the end of the second image unit and the beginning of the third image unit are the same, and so on.

Compared with the prior art, the present invention can selectively construct a plurality of proxy modules for intelligent image recognition and analysis in a logical planning manner in the original image monitoring platform, and construct and output The plurality of proxy modules are connected and can automatically perform the command module for cutting, distributing, and integrating the image data according to the idle performance of the proxy module. Therefore, in actual application, the distributed computing system applied to the image monitoring platform of the present invention will not Increasing the implementation cost of the operator will not affect the processing performance originally possessed by the image monitoring platform, and thus fully solve the various shortcomings of the prior art.

However, the above-described embodiments are merely illustrative of the principles and effects of the invention and are not intended to limit the invention. Modifications and variations of the above-described embodiments can be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, the scope of protection of the present invention should be as set forth in the scope of the claims described below.

1. . . Decentralized computing system

10. . . Command module

11, 12, 13, 14. . . Agent module

FIG. 1 is a system architecture diagram of a distributed computing system applied to an image monitoring platform of the present invention.

1. . . Decentralized computing system

10. . . Command module

11, 12, 13, 14. . . Agent module

Claims (6)

A distributed computing system applied to an image monitoring platform includes: a command module for cutting image data received by the image monitoring platform into a plurality of image units; and a plurality of proxy modules, and the command module Linking and receiving a plurality of image units cut by the command module for intelligent image recognition and analysis, and each agent module will recognize and analyze the results after the intelligent image recognition and analysis is completed. Passing to the command module for integration of the command module; wherein the command module detects the processing performance of each agent module, and determines according to the detected processing performance of each agent module The number of cuts of the image unit, and wherein the intelligently identified and analyzed results include an index to indicate as an event. The distributed computing system applied to the image monitoring platform according to claim 1, wherein the command module sends a larger number of image units to the proxy module with higher processing performance. The distributed computing system applied to the image monitoring platform according to the first aspect of the patent application, wherein each of the proxy modules performs intelligent image recognition and analysis on the received image unit according to the algorithm specified by the command module. . The distributed computing system applied to the image monitoring platform according to claim 1, wherein the command module dynamically detects each agent module to detect the status of a proxy module. When it is abnormal, it will send the original image to the proxy module with abnormal status. Meta, resend to other proxy modules. The distributed computing system applied to the image monitoring platform according to claim 1, wherein the proxy module is connected to the command module by sending a broadcast packet. The distributed computing system applied to the image monitoring platform according to the first aspect of the patent application, wherein the image monitoring platform is built on a network service provider.