WO2020015087A1

WO2020015087A1 - Method and system for large-scale processing of images, computer device, and computer storage medium

Info

Publication number: WO2020015087A1
Application number: PCT/CN2018/104643
Authority: WO
Inventors: 魏颢; 廖学聪; 朱威; 刘敬一; 斯科特·马修·罗伯特
Original assignee: 深圳码隆科技有限公司
Priority date: 2018-07-20
Filing date: 2018-09-07
Publication date: 2020-01-23
Also published as: CN108920281A

Abstract

A method and system for large-scale processing of images, a computer device, and a computer storage medium. The method for large-scale processing of images comprises the following steps: receiving an image processing request from a user (S1); storing, according to a service logic of the image processing request, the image processing request in a service logic message queue (S2); extracting the image processing request from the service logic message queue, and calling, according to the service logic of the image processing request and from data processing modules of a serverless computing device of a public cloud platform, a data processing module suitable for processing the image processing request (S3); using the data processing module to complete a processing operation requested by the image processing request (S4); and storing a completion result of the processing operation requested by the image processing request (S5).

Description

Large-scale picture processing method, system, computer equipment and computer storage medium

Cross-reference to related applications

This application claims priority from a Chinese patent application filed with the Chinese Patent Office on July 20, 2018 with the application number 201810802561.1, entitled "Large-scale Picture Processing Method and System", the entire contents of which are incorporated herein by reference.

Technical field

The present application relates to the technical field of multimedia data processing, and in particular, to a large-scale picture processing method, system, computer equipment, and computer storage medium.

Background technique

In today's network environment, network servers often need to process huge multimedia data submitted by users. Considering that the amount of multimedia data submitted by users, the time of submission, and the frequency of submission are uncertain, in order to be able to cope with the most extreme situations such as multiple users submitting a large number of multimedia data requests at the same time in the same time period, the existing A batch processing framework for multimedia data usually requires a sufficient number of high-performance servers, and multiple high-performance servers are always on standby. However, this method will cause most or even all standby servers to be in a useless state during a time period when there is less data to be processed, which is likely to cause server waste. In addition, the startup, upgrade, shutdown, and reassignment of a large number of servers will bring great operation and maintenance costs.

Application content

The present application provides a large-scale picture processing method, system, computer equipment, and computer storage medium, which are used to solve at least one of the above problems in the existing multimedia data processing technology.

According to an embodiment of the present application, a large-scale picture processing method is provided. The method includes the following steps:

S1. Receive a picture processing request from a user.

S2. Store the picture processing request in a service logic message queue according to the service logic of the picture processing request.

S3. Take out the picture processing request from the business logic message queue, and call from the data processing module of the serverless computing device of the public cloud platform to process the picture processing according to the business logic of the picture processing request. Requested data processing module;

S4. Use the data processing module to complete the processing work requested by the picture processing request;

S5. Save the completion result of the processing work requested by the picture processing request.

Preferably, the large-scale picture processing method further includes the following steps:

S0. Write the data processing module to upload the data processing module to the serverless computing device for storage, so that the serverless computing device can call the data processing module when processing the picture request.

Preferably, the data processing module uses artificial intelligence to perform data processing.

Preferably, the data processing module includes a machine learning algorithm model, a deep learning model, and code that can use the machine learning algorithm model and the deep learning model.

Preferably, in the step S0, the method of uploading the data processing module to the serverless computing device for storage is: integrating the data processing module into the serverless computing device in the form of a plug-in.

Preferably, in the step S1, the number of picture processing requests is multiple; in the step S2, the number of business logic message queues is multiple; and the step S2 includes:

Determining a business logic message queue applicable to the picture processing request according to the business logic of each picture processing request;

Each picture processing request is stored separately in a business logic message queue applicable to the picture processing request.

Preferably, the step S3 includes:

Dividing the business logic of the picture processing request into a plurality of business logic parts;

According to the plurality of business logic parts, a plurality of data processing modules respectively adapted to process the plurality of business logic parts are called from the serverless computing device of the public cloud platform.

Preferably, the step S4 includes:

Use the multiple data processing modules to complete the processing work requested by the multiple business logic sections, respectively.

Preferably, the step S5 includes:

After using the multiple data processing modules to separately complete the processing work requested by the multiple business logic parts, summarize the completion results of the processing work requested by the multiple business logic parts into an aggregated result;

The aggregation result is saved.

An embodiment of the present application further provides a large-scale image processing system, which includes a receiving module, a temporary storage module, an execution module, a public cloud platform, and a storage module; the temporary storage module includes a business logic message queue The execution module includes a business logic executor; the public cloud platform includes a serverless computing device, and the serverless computing device includes a data processing module; wherein:

The receiving module is configured to: receive a picture processing request from a user, and store the picture processing request in the business logic message queue according to the business logic of the picture processing request;

The temporary storage module is configured to temporarily store the picture processing request through the business logic message queue;

The execution module is configured to: take the picture processing request from the business logic message queue through the business logic executor, and from the serverless computing device of the public cloud platform according to the business logic of the picture processing request The data processing module calls a data processing module adapted to process the picture processing request;

The data processing module is configured to complete the processing work requested by the picture processing request;

The storage module is configured to save a completion result of processing work requested by the picture processing request.

Preferably, the storage module includes a plurality of storage devices and / or storage areas in a plurality of databases.

Preferably, the number of the data processing modules is multiple; the number of the business logic executors is multiple.

Preferably, the receiving module is further configured to: receive multiple picture processing requests from a user, determine a business logic message queue applicable to the picture processing request according to the business logic of each picture processing request; and process each picture request separately Stored in a business logic message queue suitable for this picture processing request.

An embodiment of the present application further provides a computer device, where the computer device includes a memory and a processor, where the memory is configured to store a computer program, and the processor runs the computer program to cause the computer device to perform The large-scale picture processing method according to the embodiment.

An embodiment of the present application further provides a computer-readable storage medium storing the computer program stored and used in the computer device.

According to the above embodiments, the present application essentially implements a large-scale image processing method and a corresponding system, computer device, and computer-readable storage medium based on a public cloud platform serverless computing device. The characteristics of serverless computing devices on public cloud platforms are: only actual calculations are performed, and no costs will be incurred; no calculations will not generate costs; at the same time, the increase of computing resources can be completed by itself as the call volume continues to increase. ; With the reduction in the number of calls, the reduction of computing resources is completed on its own. Therefore, in the technical solution of the present application, when a user submits a large amount of image data, the serverless computing device will increase the computing resources on its own to support a sufficient amount of data processing; when the image data submitted by the user is processed, it will no longer be When there is a request sent to a serverless computing device, no computing fees will be incurred at this time. In this way, compared with the existing image data processing technology, this application can obtain many advantages: it does not need to set up a large number of servers to work on its own, which effectively saves framework costs and operation and maintenance costs; it can easily and flexibly expand application programs , To obtain better availability and fault tolerance; can easily implement framework customization, without being limited by platform type and storage location; at the same time, this application can still support users to call peaks on the premise of obtaining these advantages The amount of work required. Obviously, this application has made various improvements to the technology for processing large-scale picture data. Since the most important part of multimedia data is picture data, the large-scale picture processing method and system of the present application is a major improvement in the field of multimedia data processing technology, and can effectively solve the problems in the prior art.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a block diagram showing a structure of a large-scale picture processing system according to an embodiment of the present application.

FIG. 2 is a schematic flowchart of main steps of a large-scale picture processing method according to an embodiment of the present application; FIG.

FIG. 3 is a schematic flowchart of main steps of a large-scale picture processing method according to an embodiment of the present application; FIG.

FIG. 4 is a schematic structural diagram of a computer device according to an embodiment of the present application.

detailed description

In order to make the purpose, technical solution, and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments.

In order to solve the existing multimedia data processing technology that requires a large number of servers, causing problems such as high costs and waste of resources, the embodiments of the present application provide a large-scale picture processing method and a corresponding method for implementing the large-scale picture processing Method for large-scale picture processing system. Since the most important part of multimedia data is picture data, the large-scale picture processing method and system of the present application is a major improvement in the field of multimedia data processing technology, and can effectively solve the problems existing in the prior art.

Since the large-scale picture processing method will be introduced when specific components of the large-scale picture processing system are introduced, in order to make the order of expression clear, the following first describes the implementation of the large-scale picture processing system based on FIG. 1 in the accompanying drawings. An introduction is made, and then the implementation manner of the large-scale picture processing method is described in detail based on FIG. 2.

Please refer to FIG. 1, which is a block diagram illustrating a structure of a large-scale picture processing system 100 according to an embodiment of the present application. The large-scale image processing system 100 includes a receiving module 10, a temporary storage module 20, an execution module 30, a public cloud platform 40, and a storage module 50.

The receiving module 10 may include, for example, an existing API (Application Programming Interface) gateway. The API gateway has a network communication connection with a larger number of actual network users, and the communication connection mode may be wired. Communication may also be wireless communication. The API gateway may receive data from each network user, for example, a picture processing request from a network user; the picture processing request carries a picture that the network user wants to process. The method for the API gateway to establish a network communication connection with a network user and to receive a picture processing request from the network user can be implemented by using related technologies, and no further description is needed here. In specific implementation, the receiving module may be configured to: receive a picture processing request from a user, and store the picture processing request in the service logic message queue according to the business logic of the picture processing request. Specifically, it may be further configured to: receive multiple picture processing requests from the user, and determine a business logic message queue applicable to the picture processing request according to the business logic of each picture processing request; and store each picture processing request to the Picture processing request in the business logic message queue.

The temporary storage module 20 may be a local temporary storage module or a temporary storage module provided on the cloud (for example, on the public cloud platform 40) to obtain a larger storage capacity. The receiving module 10 may establish a network communication connection with the temporary storage module 20 through, for example, an existing wired or wireless connection method, and may temporarily store data such as the above-mentioned picture processing request received from a network user to the temporary storage module 20 In the temporary storage module 20. The temporary storage module 20 may include multiple business logic message queues 21, and these business logic message queues 21 may respectively correspond to different types of business logic. In general, business logic refers to the rules and processes that an entity unit should have in order to provide services to another entity unit. Business logic can include (1) domain entities. Domain entities are abstract representations of various objects in a certain domain and can be represented by nouns (such as concrete or abstract nouns or even gerunds, as long as they are nominal), and constitute the skeleton of the entire business logic and Static model. Generally, each domain entity has its own attributes and behaviors. (2) Business rules. Business rules are rules that operate in a certain field and constitute the key points and dynamic model of the entire business logic. Business rules act on domain entities, and domain entities operate in compliance with business rules. (3) Integrity constraints. Domain entities and business rules construct the body of business logic, but on top of this body, there are also restrictions on integrity constraints. (4) Business processes and workflows. A business process or workflow is the process of initiating and hosting coordination domain entities to complete established rules. Business logic can be stored in a temporary queue in the form of a message queue. Different types of business logic can be set up with different message queues. Each message queue can run asynchronously. The receiving module 10 may determine a business logic message queue 21 applicable to the picture processing request according to the specific business logic type of each received picture processing request, and then temporarily store each picture processing request to the The business logic message queue 21 of the picture processing request, that is, the business logic message queue 21 corresponding to the business logic type of the picture processing request. Each of the business logic message queues 21 can store multiple picture processing requests with corresponding business logic types.

The execution module 30 may establish a network communication connection with the temporary storage module 20 through, for example, an existing wired or wireless connection method, and retrieve data from the temporary storage module 20, for example, temporarily save the temporary storage. The picture processing request in the business logic message queue 21 of the storage module 20 is stored. Specifically, the execution module 30 may include a plurality of business logic executors 31. These business logic executors 31 may be set to correspond to different types of business logic, and therefore may be based on their corresponding business logic types. A correspondence relationship is further established with the above-mentioned business logic message queue 21. Each business logic executor 31 may establish communication with its corresponding business logic message queue 21 through, for example, an existing wired or wireless connection method, and take out and store the business logic message queue from its corresponding business logic message queue 21 Image processing request in 21. It can be understood that the execution module 30 can also be set on the cloud (for example, on the public cloud platform 40) to save local resources.

The public cloud platform 40 may be a cloud platform that can be used by users provided by third-party providers in the current market, and is responsible for organizing and coordinating computing resources and providing various computing services according to user needs. The public cloud platform 40 may include a serverless computing device 41. Among them, the public cloud platform is a data processing module. The public cloud is generally available through the Internet. It is usually free or low-cost. Its core attribute is shared resource services, which can be considered as the main form of cloud computing. . The basic architecture and basic working principle of the public cloud platform 40 and the serverless computing device 41 therein can be implemented with reference to related technologies in the art, and therefore will not be repeated here. The execution module 30 may establish a network communication connection with the public cloud platform 40 through, for example, an existing wired or wireless connection method. In this embodiment, the serverless computing device 41 may include multiple data processing modules 42. Each data processing module 42 may use artificial intelligence to perform data processing. Specifically, each data processing module 42 may include, for example, Machine learning algorithm models, deep learning models, and codes that can use the machine learning algorithm models and deep learning models, and so on. These data processing modules 42 may be business logic basic units integrated in the serverless computing device 41 in the form of plug-ins, and they may correspond to different types of business logic, so they may be further based on their corresponding business logic types. Corresponds to the business logic message queue 21 and business logic executor 31 described above. After each business logic executor 31 takes the picture processing request stored in the business logic message queue 21 from the corresponding business logic message queue 21, it can query the corresponding data processing module 42 according to the corresponding business logic type. Then, the data processing module 42 is called to complete the processing work requested by the picture processing request (that is, the calculation work required by the picture processing request). In this way, all the calculation work required for the image processing request can be transferred to the serverless computing device 41 on the public cloud platform 40 for execution, without the need to perform these calculation work on the local server, thereby saving better. Local resources.

The storage module 50 may include a variety of local data storage devices such as hard disks, mobile hard disks, etc., or may include one of the data storage devices (e.g., the public cloud platform 40 or other available clouds). Storage area in one or more databases (such as MangoDB database, MySQL database, HBase database, etc.).

The execution module 30 may establish a network communication connection with the storage module 50 through, for example, an existing wired or wireless connection manner. After each of the data processing modules 42 completes the processing work requested by the picture processing request, the corresponding business logic executor 31 in the execution module 30 can obtain the processing results (such as picture processing) made by the data processing module 42 Request the required calculation result), and store the processing result in the storage module 50 for query and use by the user. Of course, the corresponding business logic executor 31 in the execution module 30 may also actively send the obtained processing results made by the data processing module 42 to the user. The specific implementation manner can be flexibly set according to user needs.

Please refer to FIG. 2 together, which illustrates a schematic flowchart of main steps of a large-scale picture processing method according to an embodiment of the present application. The large-scale picture processing method may use the large-scale picture processing system 100 shown in FIG. 1. to realise. This method may be executed by a large-scale picture processing system. For an implementation manner of the large-scale picture processing system, refer to FIG. 1, and details are not described herein again. Specifically, a large-scale picture processing system can be implemented using, for example, computer equipment. The steps of the large-scale picture processing method are detailed below:

S1. Receive a picture processing request from a user.

In step S1, for example, the receiving module 10 in the large-scale picture processing system 100 can receive picture processing requests from users. The number of picture processing requests may be one or more. The picture processing requests include: Pictures to be processed by the user, and the picture processing request corresponds to a business logic type. Since the API module can be used in the receiving module 10, which can receive a large amount of image data at the same time, the embodiment of the present application allows users to upload a large number of image processing requests in batches at the same time. Requests for 10, or even 10 million images.

In step S2, after receiving the image processing request (usually a large number of image processing requests), the receiving module 10 does not immediately hand over these image processing requests to a data processing device (such as a local server or the cloud) for processing and return to processing. As a result, these picture processing requests are first stored in the business logic message queue 21 of the temporary storage module 20 for asynchronous processing. Specifically, the number of the business logic queues 21 is preferably multiple, and these business logic message queues 21 may be set to correspond to different types of business logic, respectively; the receiving module 10 may process according to each picture received The specific business logic type of the request determines the business logic message queue 21 that is applicable to the picture processing request, and then temporarily stores each picture processing request to the business logic message queue 21 that is suitable for the picture processing request, that is, with the picture The business logic message queue 21 corresponding to the type of business logic processing the request. Each of the business logic message queues 21 can store multiple picture processing requests with corresponding business logic types.

S3. Take out the picture processing request from the business logic message queue, and call from the data processing module of the serverless computing device of the public cloud platform to process the picture processing according to the business logic of the picture processing request. Requested data processing module.

In step S3, the business logic executor 31 sequentially retrieves and stores the business logic from the corresponding business logic message queue 21 in a predetermined order (for example, the storage order of picture processing requests in the business logic message queue 21). Each picture in the message queue 21 processes a request, and then queries the data processing module 42 corresponding to the type of the business logic from the serverless computing device 41 of the public cloud platform 40 according to the business logic of the retrieved picture processing request, and calls The data processing module 42 completes the processing work requested by the picture processing request (that is, the calculation work required by the picture processing request). In this way, all the calculation work required for the picture processing request can be transferred to the serverless computing device 41 on the public cloud platform 40 for execution, without performing these calculation work on the local server. In addition, since the serverless computing device 41 on the public cloud platform 40 can include a sufficient number of data processing modules 42, it can process a large number of image processing requests at the same time. The processing power is very powerful and will not be affected by the capacity of the local server. limits. In addition, the serverless computing device 41 on the public cloud platform 40 can automatically call an appropriate amount of data processing module 42 according to the number of image processing requests, that is, adjust the computing resources on its own, while satisfying the needs of users, it can also reasonably use the resources and avoid Waste computing resources and generate extra computing costs. S4. Use the data processing module to complete the processing work requested by the picture processing request.

In step S4, each data processing module 42 processes the processing work requested by the image processing request corresponding to its type based on the corresponding business logic and the work order arranged by the corresponding business logic executor 31.

In this step S5, after each of the data processing modules 42 completes the processing work requested by the picture processing request, the corresponding business logic executor 31 in the execution module 30 can obtain the data processing module 42 A processing result (such as a calculation result required for a picture processing request), and the processing result is stored in the storage module 50 for a user to query and use. In particular, if the storage module 50 includes a plurality of storage units, such as a plurality of storage devices provided locally and / or a storage area provided in a variety of databases on the cloud, the execution module 30 may correspond to The business logic executors 31 of different business logic types respectively store the processing results made by the data processing module 42 corresponding to the corresponding business logic type into different types of storage units. For example, the The business logic executor 31 stores the processing results made by the data processing module 42 corresponding to the first business logic type into the MangoDB database, and the business logic executor 31 corresponding to the second business logic type will correspond to the second The processing result made by the data processing module 42 of each business logic type is stored in the MySQL database, and the data processing module 42 corresponding to the third business logic type is performed by the business logic executor 31 corresponding to the third business logic type. The output processing results are stored in the HBase database. Logic 31 performs the processing corresponding to the result of the fourth type of business logic of the data processing module 42 is made to the local storage device is stored. The above is only an exemplary description, and should not be considered as a limitation. In practical applications, the type of processing results used by each storage unit can be flexibly set according to requirements. The storage unit can be implemented by using various storage device hardware and various forms of databases, and the storage methods are flexible and diverse. In this manner provided by this embodiment, the storage amount of a single storage unit can be saved, on the one hand, the security of storing the processing results is improved, and on the other hand, it is convenient for users to sort and arrange the processing results.

Preferably, in a further embodiment, the large-scale picture processing method further includes the following steps:

This step S0 is preferably performed before the above steps S1-S5, and at least before step S3. Users or other system developers can write a data processing module 42 with predefined business logic according to the possible requirements in actual work, and upload the written data processing module 42 to the serverless of the public cloud platform 40 It is stored in the computing device 41. When the steps S1 to S5 are performed, the user can query and call the data processing module 42 written or stored in advance by himself or others for processing the processing work requested by the image processing request with the corresponding business logic type.

Considering that the business logic of the image processing request may be more complicated, in order to be able to process the image processing request more reasonably and efficiently, in a further embodiment, the step S3 includes:

The business logic of the picture processing request is divided into a plurality of business logic parts; each business logic part may correspond to a type of business logic.

According to the plurality of business logic parts, a plurality of data processing modules respectively adapted to process the plurality of business logic parts are called from the serverless computing device of the public cloud platform. In this case, multiple data processing modules can be called to decompose the image processing request with complex business logic, and the processing efficiency can be effectively improved through the simultaneous processing of multiple threads.

Based on dividing the business logic of the picture processing request into a plurality of business logic parts, the step S4 includes:

Since the business logic of the picture processing request is decomposed and processed by multiple data processing modules, the step S5 includes:

After using the plurality of data processing modules to complete the processing work requested by the plurality of business logic parts, respectively, the completion results of the processing work requested by the plurality of business logic parts are summarized into an aggregation result; the aggregation result is also That is the final processing result of the picture processing request.

The aggregation result is saved.

Specifically, when the business logic of a picture processing request is more complicated, in order to improve the data processing efficiency, the execution module 30 may divide the business logic of the picture processing request into a plurality of business logic parts, and determine the multiple The respective business logic type of the business logic part. According to the business logic types of the multiple business logic parts, the execution module 30 may use a plurality of business logic executors 31 corresponding to the business logic types of the multiple business logic parts, respectively, from the public cloud platform 40 The serverless computing device 41 calls multiple data processing modules 42 that are respectively suitable for processing the multiple business logic parts (that is, corresponding to the business logic types of the multiple business logic parts), and then uses all The plurality of data processing modules 42 respectively complete the processing work requested by the plurality of business logic sections. Obviously, such a data processing method can split the business logic of a picture processing request into multiple parts, and then process the computing work of the multiple parts at the same time in multiple threads, which is beneficial to improving the data processing efficiency. In order to ensure that the final data processing result obtained by the above data processing method is still complete (that is, the complete result required by the business logic of the entire image processing request is obtained), in this embodiment, if the plurality of data processing modules 42 One or more of the data processing modules 42 completed the processing work requested by their corresponding business logic part earlier than this time, then the data processing results obtained by the data processing module 42 are not immediately output and saved at this time; Until the multiple data processing modules 42 have completed the processing work requested by their respective business logic sections, the completion results of the processing work requested by the multiple business logic sections are first aggregated into aggregated results, and then The aggregation result is output and stored in the storage module 50. This can ensure that the data processing results obtained by the above multi-threaded processing method are complete.

Based on the foregoing embodiment, reference is again made to a schematic flowchart of main steps of a large-scale picture processing method according to another embodiment of the present application shown in FIG. 3. The large-scale picture processing method may use the large-scale picture processing shown in FIG. The system 100 implements the specific steps as follows: S00, writing a data processing module. The number of data processing modules may be multiple, and different data processing modules may correspond to different types of business logic.

S01. Integrate the data processing module into a serverless computing device in the form of a plug-in. The data processing module may establish a correspondence relationship with a business logic message queue and a business logic executor based on its corresponding business logic type.

S10. Receive multiple picture processing requests from a user. Each picture processing request corresponds to business logic and contains pictures to be processed.

S20: Determine a business logic message queue applicable to the picture processing request according to the service logic of each picture processing request. It can be understood that different business logic message queues can correspond to different types of business logic, so the business logic message queue to be stored can be determined based on the business logic of each picture processing request.

S21. Each picture processing request is stored in a business logic message queue applicable to the picture processing request.

S30. Divide the business logic of the picture processing request into multiple business logic parts. For the image processing request with complex business logic, the business logic of the image processing request can be divided into multiple parts, so that the serverless computing device of the public cloud platform can process multiple threads simultaneously for different parts.

S31. According to a plurality of business logic parts, a plurality of data processing modules respectively adapted to process a plurality of business logic parts are called from a serverless computing device of a public cloud platform.

S40. Use the multiple data processing modules to complete the processing tasks requested by the multiple business logic sections, respectively. In this way, multiple threads of processing of multiple parts of the image processing request can be processed simultaneously, effectively improving data processing efficiency.

S50: Summarize the completion results of the processing jobs requested by the multiple business logic parts into an aggregation result.

S51. Save the polymerization result. This aggregation result is also the processing result of the image processing request.

In summary, the above-mentioned method provided by this embodiment does not need to set up a large number of servers to work locally by using a public cloud platform, which effectively saves framework costs and operation and maintenance costs. Moreover, this method can not only be easily and flexibly expanded Applications for better usability and fault tolerance; it can also easily customize the framework without being limited by platform type and storage location. At the same time, under the premise of obtaining these advantages, the present application can still stably support the workload required by the user to call the peak. In addition, multi-thread simultaneous processing of picture processing requests with complex business logic is also better, which effectively improves processing efficiency.

An embodiment of the present invention provides a computer device. Referring to a schematic structural diagram of a computer device shown in FIG. 4, the computer device includes a processor 1, a memory 2, a bus 3, and a communication interface 4. The processor 1, The communication interface 4 and the memory 2 are connected through a bus 3; the processor 1 is configured to execute an executable module, such as a computer program, stored in the memory 2.

The memory 2 may include a high-speed random access memory (RAM, Random Access Memory), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface 4 (which can be wired or wireless), and the Internet, a wide area network, a local network, a metropolitan area network, and the like can be used.

The bus 3 may be an ISA bus, a PCI bus, an EISA bus, or the like. The bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only a two-way arrow is used in FIG. 7, but it does not mean that there is only one bus or one type of bus.

The memory 2 is used to store a program, and the processor 1 executes the program after receiving an execution instruction. The method executed by the apparatus defined by the flow process disclosed in any one of the foregoing embodiments of the present invention may be applied to processing. Processor 1, or implemented by processor 1.

The processor 1 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method may be completed by an integrated logic circuit of hardware in the processor 1 or an instruction in the form of software. The above processor 1 may be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc .; it may also be a Digital Signal Processor (DSP) ), Application Specific Integrated Circuit (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. Various methods, steps, and logical block diagrams disclosed in the embodiments of the present invention may be implemented or executed. A general-purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in combination with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or may be performed by using a combination of hardware and software modules in the decoding processor. The software module may be located in a mature storage medium such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, and the like. The storage medium is located in the memory 2. The processor 1 reads the information in the memory 2 and completes the steps of the foregoing method in combination with its hardware.

According to the above embodiments, the present application essentially implements a large-scale image processing method and a corresponding system, computer device, and computer-readable storage medium based on a public cloud platform serverless computing device. The characteristics of serverless computing devices on public cloud platforms are: only actual calculations are performed, and no costs are incurred; no calculations will not generate costs; at the same time, as the call volume continues to increase, the computing resources can be increased by themselves. ; With the reduction in the number of calls, the reduction of computing resources is completed on its own. Therefore, in the technical solution of the present application, when a user submits a large amount of image data, the serverless computing device will increase the computing resources on its own to support a sufficient amount of data processing; when the image data submitted by the user is processed, it will no longer be When there is a request sent to a serverless computing device, no computing fees will be incurred at this time. In this way, compared with the existing image data processing technology, this application can obtain many advantages: it does not need to set up a large number of servers to work on its own, which effectively saves framework costs and operation and maintenance costs; it can easily and flexibly expand application programs , To obtain better availability and fault tolerance; can easily implement framework customization, without being limited by platform type and storage location; at the same time, this application can still support users to call peaks on the premise of obtaining these advantages The amount of work required. Obviously, this application has made various improvements to the technology for processing large-scale picture data. Since the most important part of multimedia data is picture data, the large-scale picture processing method and system of the present application is a major improvement in the field of multimedia data processing technology, and can effectively solve the problems in the prior art.

In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may also be implemented in other ways. The device embodiments described above are only schematic. For example, the flowcharts and structure diagrams in the drawings show the possible architectures and functions of the devices, methods, and computer program products according to various embodiments of the present application. And operation. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, which contains one or more components for implementing a specified logical function Executable instructions. It should also be noted that in alternative implementations, the functions labeled in the blocks may also occur in a different order than those labeled in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be a dedicated hardware-based system that performs a specified function or action. To achieve, or can be implemented by a combination of dedicated hardware and computer instructions.

In addition, the functional modules or units in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application is essentially a part that contributes to the existing technology or a part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a smart phone, a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in the embodiments of the present application. The foregoing storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disks or optical disks and other media that can store program codes .

The above are only preferred embodiments of this application, and are not intended to limit this application. Any modification, equivalent replacement, or improvement made within the spirit and principles of this application shall be included in this application Within the scope of protection.

Industrial applicability:

By applying the technical solution of the present application, there is no need to set up a large number of servers to work by itself, which effectively saves framework costs and operation and maintenance costs; can easily and flexibly expand applications, and obtain better availability and fault tolerance; The framework can be customized and is not limited by the type of platform and storage location. At the same time, this application can still support the workload required by users to call the peak on the premise of obtaining these advantages.

Claims

A large-scale picture processing method includes the following steps:

S1. Receive a picture processing request from a user.

S2. Store the picture processing request in a service logic message queue according to the service logic of the picture processing request.

S3. Take out the picture processing request from the business logic message queue, and call from the data processing module of the serverless computing device of the public cloud platform to process the picture processing according to the business logic of the picture processing request. Requested data processing module;

S4. Use the data processing module to complete the processing work requested by the picture processing request;

S5. Save the completion result of the processing work requested by the picture processing request.
The method for processing large-scale pictures according to claim 1, further comprising the following steps:

S0. Write the data processing module to upload the data processing module to the serverless computing device for storage, so that the serverless computing device can call the data processing module when processing the picture request.
The large-scale picture processing method according to claim 2, wherein the data processing module uses artificial intelligence to perform data processing.
The large-scale image processing method according to claim 3, wherein the data processing module comprises a machine learning algorithm model, a deep learning model, and code that can use the machine learning algorithm model and the deep learning model.
The method for large-scale picture processing according to claim 2, wherein in the step S0, a method for uploading the data processing module to the serverless computing device for storage is: uploading the data processing module Integrated into the serverless computing device in the form of a plug-in.
The large-scale picture processing method according to claim 1, wherein in the step S1, the number of the picture processing requests is plural; in the step S2, the number of the business logic message queues There are multiple; the step S2 includes:

Determining a business logic message queue applicable to the picture processing request according to the business logic of each picture processing request;

Each picture processing request is stored separately in a business logic message queue applicable to the picture processing request.
The method for processing a large-scale picture according to claim 6, wherein the step S3 comprises:

Dividing the business logic of the picture processing request into a plurality of business logic parts;

According to the plurality of business logic parts, a plurality of data processing modules respectively adapted to process the plurality of business logic parts are called from the serverless computing device of the public cloud platform.
The large-scale picture processing method according to claim 7, wherein the step S4 comprises:

Use the multiple data processing modules to complete the processing work requested by the multiple business logic sections, respectively.
The large-scale picture processing method according to claim 8, wherein the step S5 comprises:

After using the multiple data processing modules to separately complete the processing work requested by the multiple business logic parts, summarize the completion results of the processing work requested by the multiple business logic parts into an aggregated result;

The aggregation result is saved.
A large-scale picture processing system, the large-scale picture processing system includes a receiving module, a temporary storage module, an execution module, a public cloud platform, and a storage module; the temporary storage module includes a business logic message queue; and the execution module includes a service A logic executor; the public cloud platform includes a serverless computing device, and the serverless computing device includes a data processing module; wherein:

The receiving module is configured to: receive a picture processing request from a user, and store the picture processing request in the business logic message queue according to the business logic of the picture processing request;

The temporary storage module is configured to temporarily store the picture processing request through the business logic message queue;

The execution module is configured to: take the picture processing request from the business logic message queue through the business logic executor, and from the serverless computing device of the public cloud platform according to the business logic of the picture processing request The data processing module calls a data processing module adapted to process the picture processing request;

The data processing module is configured to complete the processing work requested by the picture processing request;

The storage module is configured to save a completion result of processing work requested by the picture processing request.
The large-scale picture processing system according to claim 10, wherein the storage module comprises a plurality of storage devices and / or storage areas in a plurality of databases.
The large-scale picture processing system according to claim 10, wherein the data processing module uses artificial intelligence to perform data processing.
The large-scale image processing system according to claim 12, wherein the data processing module comprises a machine learning algorithm model, a deep learning model, and code that can use the machine learning algorithm model and the deep learning model.
The large-scale picture processing system according to claim 10, wherein the number of the data processing modules is plural; and the number of the business logic executors is plural.
The large-scale picture processing system according to claim 10, wherein the receiving module is further configured to: receive multiple picture processing requests from a user, and determine that the picture processing is applicable to the picture processing according to the business logic of each picture processing request Requested business logic message queue; each picture processing request is stored separately in a business logic message queue that is applicable to the picture processing request.
A computer device, characterized in that the computer device includes a memory and a processor, the memory is configured to store a computer program, and the processor runs the computer program to cause the computer device to execute the claims 1-9 The large-scale picture processing method according to any one of the above.
A computer-readable storage medium, characterized in that the computer program stored and used in the computer device according to any one of claims 10 to 15 is stored.