Algorithm interaction method for multiple AI application scenarios in coal mines

Abstract

The present invention proposes a method for interacting multiple AI application scenario algorithms in coal mines. A plurality of selectable coal mine underground AI application scenario algorithms are built in the AI ​​application scenario algorithm mall corresponding to the coal mine, and various algorithm images are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall; based on sample data of any target coal mine, various algorithm images corresponding to the target coal mine are trained to obtain various AI application scenario algorithm images; multiple target AI application scenario algorithm images that are superior to various AI application scenario algorithms in various AI application scenario algorithm images are quality-checked to replace the corresponding algorithm images in the AI ​​application scenario algorithm mall. Thus, based on the coal mine sample data, accurate training and deployment of coal mine AI application scenario algorithms are achieved, and for the differences between different coal mines with the same AI application scenario algorithms, the threshold for rapid application of AI application scenario algorithms to coal mine sites is reduced.

Description

Technical Field

[0001] The present invention relates to the field of coal mine technology, and in particular to a method and device for algorithm interaction in multiple AI application scenarios in coal mines. Background Art

[0002] With the country's requirements for safe, intelligent and efficient coal mine production and the continuous development of technologies such as artificial intelligence and machine learning, coal mines' demand for intelligence will continue to increase. Among the related technologies, coal mine artificial intelligence (AI) application scenario algorithms are generally selected directly from the AI ​​application scenario algorithm mall. Since the coal mine multi-AI application scenario algorithms are fixed and different coal mines are different under the same AI application scenario algorithm, the traditional AI application scenario algorithm is not suitable for all coal mine application scenarios, resulting in the application effect of the AI ​​application scenario algorithm being quite different from the actual application scenario of the coal mine. Therefore, a more reliable coal mine multi-AI application scenario algorithm interaction method is urgently needed. Summary of the invention

[0003] The present invention aims to solve one of the technical problems in the related art at least to a certain extent.

[0004] To this end, the first purpose of the present invention is to propose a method for interacting multiple AI application scenario algorithms in coal mines. Based on coal mine sample data, the precise training and deployment of coal mine AI application scenario algorithms can be realized. Due to the differences between different coal mines with the same AI application scenario algorithms, the threshold for rapid application of AI application scenario algorithms to coal mine sites is lowered.

[0005] The second purpose of the present invention is to propose a coal mine multi-AI application scenario algorithm interaction device.

[0006] A third objective of the present invention is to provide an electronic device.

[0007] A fourth object of the present invention is to provide a non-transitory computer-readable storage medium storing computer instructions.

[0008] To achieve the above-mentioned purpose, the first embodiment of the present invention proposes a coal mine multi-AI application scenario algorithm interaction method, comprising:

[0009] Obtain coal mine sample data from multiple coal mines operating underground;

[0010] The AI ​​application scenario algorithm mall corresponding to coal mines has a variety of optional coal mine underground AI application scenario algorithms built in, and the algorithm images of various coal mine underground AI application scenario algorithms are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall;

[0011] Based on the sample data of any target coal mine, various algorithm images deployed in the target image warehouse corresponding to the target coal mine are trained to obtain algorithm images for various AI application scenarios;

[0012] Based on the harmonic mean, precision and recall values ​​of each of the AI ​​application scenario algorithm images, multiple target AI application scenario algorithm images that are superior to each of the AI ​​application scenario algorithms are quality-checked, and the multiple target AI application scenario algorithm images are used to replace the corresponding algorithm images in the AI ​​application scenario algorithm mall corresponding to the coal mine, otherwise they will not be replaced.

[0013] To achieve the above-mentioned purpose, the second embodiment of the present invention proposes a coal mine multi-AI application scenario algorithm interaction device, comprising:

[0014] An acquisition module, used to acquire coal mine sample data when multiple coal mines are operating underground;

[0015] A deployment module is used for a variety of optional coal mine underground AI application scenario algorithms built into the AI ​​application scenario algorithm mall corresponding to coal mines, and the algorithm images of various coal mine underground AI application scenario algorithms are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall;

[0016] The training module is used to train various algorithm images deployed in the target image warehouse corresponding to the target coal mine based on the sample data of any target coal mine, so as to obtain the algorithm images of various AI application scenarios;

[0017] A replacement module is used to quality-check multiple target AI application scenario algorithm images that are superior to each of the AI ​​application scenario algorithms based on the harmonic mean, precision and recall values ​​of each of the AI ​​application scenario algorithm images, and replace the corresponding algorithm images in the AI ​​application scenario algorithm mall corresponding to the coal mine with the multiple target AI application scenario algorithm images, otherwise they will not be replaced.

[0018] To achieve the above-mentioned purpose, the third aspect of the present invention proposes an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in the first aspect.

[0019] In order to achieve the above-mentioned purpose, the fourth aspect of the present invention proposes a non-transitory computer-readable storage medium storing computer instructions, where the computer instructions are used to enable the computer to execute the method described in the first aspect.

[0020] The coal mine multi-AI application scenario algorithm interaction method, device, electronic device and storage medium of the embodiment of the present invention, the coal mine corresponding AI application scenario algorithm mall has a variety of selectable coal mine underground AI application scenario algorithms built in, and various algorithm images are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall; based on the sample data of any target coal mine, the various algorithm images corresponding to the target coal mine are trained to obtain various AI application scenario algorithm images; multiple target AI application scenario algorithm images that are superior to various AI application scenario algorithms in each AI application scenario algorithm image are quality inspected to replace the corresponding algorithm images in the AI ​​application scenario algorithm mall, thereby, based on the coal mine sample data, the accurate training and deployment of the coal mine AI application scenario algorithm is realized, and for the differences between different coal mines with the same AI application scenario algorithm, the threshold for rapid application of the AI ​​application scenario algorithm to the coal mine site is lowered.

[0021] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0023] Figure 1 A schematic diagram of a flow chart of a coal mine multi-AI application scenario algorithm interaction method provided by an embodiment of the present invention;

[0024] Figure 2 A schematic flow chart of another method for interacting with multiple AI application scenarios in coal mines provided by an embodiment of the present invention;

[0025] Figure 3 An implementation example diagram of the interaction of multiple AI application scenarios in coal mines provided by an embodiment of the present invention;

[0026] Figure 4 An implementation example diagram of a coal mine multi-AI application scenario algorithm provided by an embodiment of the present invention;

[0027] Figure 5 A schematic diagram of the structure of a coal mine multi-AI application scenario algorithm interaction device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0028] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0029] It should be noted that the acquisition, storage, use, and processing of data in the technical solution of the present invention are in compliance with the relevant provisions of national laws and regulations.

[0030] The following describes the coal mine multi-AI application scenario algorithm interaction method and device according to an embodiment of the present invention with reference to the accompanying drawings.

[0031] Figure 1 A flowchart of a coal mine multi-AI application scenario algorithm interaction method provided by an embodiment of the present invention.

[0032] like Figure 1 As shown, the method comprises the following steps:

[0033] Step 101, obtaining coal mine sample data of multiple coal mines in operation underground.

[0034] In some embodiments, an implementation method for obtaining coal mine sample data of multiple coal mines when they are operating underground can be to collect videos, pictures, point clouds and audio data of each coal mine when it is operating underground, and classify and label them to obtain different types of initial coal mine operation data; and use the different types of initial coal mine operation data corresponding to each coal mine as the coal mine sample data of each coal mine when it is operating underground.

[0035] Furthermore, it is possible to use mining intrinsically safe cameras or visual acquisition equipment such as lidar to obtain visually perceived two-dimensional or three-dimensional visual image sample data from underground coal mines through network protocols (ONVIF, RTSP) and customized network application layer protocols, or to use microphones to collect audio sample data, which are then stored in the coal mine's platform database for data management.

[0036] Optionally, the visual image sample data and audio sample data can be subjected to relevant labeling methods such as video, picture, point cloud and audio classification, detection and segmentation, so as to obtain different types of initial coal mine operation data corresponding to each coal mine, which can be used as coal mine sample data when each coal mine is operating underground.

[0037] Step 102, the coal mine corresponding AI application scenario algorithm mall has a variety of optional coal mine underground AI application scenario algorithms built in, and the algorithm images of various coal mine underground AI application scenario algorithms are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall.

[0038] In some embodiments, the artificial intelligence (AI) application scenario algorithm includes an algorithm training image (basic application scenario algorithm A and training program), an algorithm quality inspection image, and an algorithm deployment image.

[0039] Optionally, the AI ​​application scenario algorithm management of this method implements standardized construction of AI application scenario algorithms based on Docker containerization packaging, which can be packaged once and run everywhere. Provide three types of algorithm packaging standard formats: training type (algorithm training image), quality inspection type (algorithm quality inspection image) and deployment type (algorithm deployment image), run a dedicated Docker image repository Harbor to uniformly manage various container images in the system. By providing a standard container packaging sample Dockerfile, AI application scenario algorithm developers perform Docker image packaging operations in an environment that matches the operating environment, and push it to the corresponding dedicated Docker image repository of the AI ​​application scenario algorithm mall.

[0040] Among them, by publishing the algorithm training image to the AI ​​scene mall, the artificial intelligence basic model A, training program, deep learning framework (pytorch, tensorflow, etc.) operating environment, descriptive files and other related files are packaged into a Docker image and pushed to the corresponding dedicated Docker image warehouse of the AI ​​application scenario algorithm mall. Training can be carried out without the need for additional configuration of the deep learning operating environment dependency, and support for modification of hyperparameters (batchsize, epoch, learning rate).

[0041] Step 103: Based on the sample data of any target coal mine, various algorithm images deployed in the target image warehouse corresponding to the target coal mine are trained to obtain algorithm images for various AI application scenarios.

[0042] Optionally, the target coal mine may be any coal mine that operates normally within a preset time period, but is not limited thereto.

[0043] Furthermore, in the case where the AI ​​application scenario algorithm is a belt deviation detection algorithm, the sample data may include but is not limited to unit time and belt deviation distance, and this embodiment does not make specific limitations on this.

[0044] In addition, when the AI ​​application scenario algorithm is a coal flow monitoring algorithm, the sample data may include but is not limited to coal flow transportation time and total coal flow, and this embodiment does not specifically limit this.

[0045] Step 104, based on the harmonic mean, precision and recall values ​​of each AI application scenario algorithm image, quality inspect each AI application scenario algorithm image to find out multiple target AI application scenario algorithm images that are better than each AI application scenario algorithm, and replace the corresponding algorithm images in the corresponding AI application scenario algorithm mall of the coal mine with the multiple target AI application scenario algorithm images, otherwise they will not be replaced.

[0046] In some embodiments, the target AI application scenario algorithm image that has been replaced in the corresponding AI application scenario algorithm mall for coal mines and the unreplaced AI application scenario algorithm should be deployed to the surface graphics processing unit (GPU) server or the underground mining edge computing power controller to facilitate use and improve recognition accuracy.

[0047] In the coal mine multi-AI application scenario algorithm interaction method of the embodiment of the present invention, a plurality of selectable coal mine underground AI application scenario algorithms are built in the AI ​​application scenario algorithm mall corresponding to the coal mine, and various algorithm images are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall; based on the sample data of any target coal mine, various algorithm images corresponding to the target coal mine are trained to obtain various AI application scenario algorithm images; multiple target AI application scenario algorithm images that are superior to various AI application scenario algorithms in each AI application scenario algorithm image are quality inspected to replace the corresponding algorithm images in the AI ​​application scenario algorithm mall, thereby, based on the coal mine sample data, accurate training and deployment of coal mine AI application scenario algorithms are realized, and for the differences between different coal mines with the same AI application scenario algorithm, the threshold for rapid application of AI application scenario algorithms to coal mine sites is lowered.

[0048] In order to clearly illustrate the previous embodiment, this embodiment also provides a coal mine multi-AI application scenario algorithm interaction method. Figure 2 A flowchart of another method for algorithm interaction in multiple AI application scenarios in coal mines provided in an embodiment of the present invention.

[0049] like Figure 2 As shown, the method may include the following steps:

[0050] Step 201, obtaining coal mine sample data of multiple coal mines in operation underground.

[0051] Step 202: The AI ​​application scenario algorithm mall corresponding to the coal mine has a variety of optional coal mine underground AI application scenario algorithms built in. The algorithm images of various coal mine underground AI application scenario algorithms are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall.

[0052] It should be noted that, for the specific implementation of step 201 to step 202, reference may be made to the relevant description in the above embodiment.

[0053] Step 203: Divide the sample data of any target coal mine into training sample data and test sample data according to a preset ratio.

[0054] In some embodiments, a ratio of 7:3 or 8:2 may be selected to divide the training sample data and the test sample data, but the ratio is not limited thereto.

[0055] In addition, if the sample data is insufficient, data enhancement methods such as rotation, scaling, translation, and Gaussian blur can be selected to increase the number of coal mine sample data and provide generalization capabilities.

[0056] Step 204: Based on the training sample data, various algorithm training images deployed in the target image warehouse corresponding to the target coal mine are trained to obtain algorithm images for various initial AI application scenarios.

[0057] In some embodiments, by uploading training sample data to an algorithm training image, each initial AI application scenario algorithm image is trained after a preset maximum training time or other hyperparameters (Epoch).

[0058] Step 205: Verify each initial AI application scenario algorithm image based on the test sample data to obtain each corresponding AI application scenario algorithm image when the verification result is less than a preset error threshold.

[0059] Optionally, the preset error threshold may be set to 1%, but is not limited thereto.

[0060] Step 206, based on the algorithm quality inspection image, calculate the harmonic mean, precision and recall value of each AI application scenario algorithm image, so as to quality inspect each AI application scenario algorithm image and find out multiple target AI application scenario algorithm images that are superior to each AI application scenario algorithm.

[0061] Optionally, based on the algorithm quality inspection image, the harmonic mean F1-score, precision and recall values ​​of each AI application scenario algorithm image (AI) are calculated, and compared with each AI application scenario algorithm (A) to determine the pros and cons of each AI application scenario algorithm image. If AI is better than A, AI is used and published to the AI ​​application scenario algorithm mall as the optimal algorithm B for the AI ​​application scenario algorithm. If new training sample data is input for training, the quality inspection comparison uses B instead of A. On the contrary, if AI is worse than A, AI can be selected for transfer learning to continue training, and ultimately the expected algorithm training effect is achieved. Since it involves confidentiality issues to the outside world, the algorithm can be encrypted to ensure its security.

[0062] Step 207, publish multiple target AI application scenario algorithm images to the algorithm deployment image to replace the corresponding algorithm images in the AI ​​application scenario algorithm mall corresponding to the coal mine, otherwise they will not be replaced.

[0063] Optionally, the deployment of the target AI application scenario algorithm image usually requires artificial intelligence algorithm conversion and adaptation to the corresponding AI computing power chip. The adaptation solution varies according to the AI ​​computing power chip manufacturer. The converted target AI application scenario algorithm image is then repackaged into the algorithm deployment image according to the rules and sent to the coal mine GPU server or the underground mining edge computing power controller.

[0064] Furthermore, the modular design concept is adopted, and the deployment method is a separate modular design, which can be independently applied to the application algorithm update of multiple edge mining computing controllers. There are two packaging methods for algorithm deployment images: one is the docker method; the other is to combine and package multiple files such as algorithm files, deployment algorithms or configurations in a custom way, and specify the storage address of the edge computing controller. The edge computing controller runs the parsing program, parses the packaged files, and runs the algorithm program to complete the deployment.

[0065] In the coal mine multi-AI application scenario algorithm interaction method of the embodiment of the present invention, a plurality of selectable coal mine underground AI application scenario algorithms are built in the AI ​​application scenario algorithm mall corresponding to the coal mine, and various algorithm images are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall; the sample data of any target coal mine is divided into training sample data and test sample data according to a preset ratio; based on the training sample data, various algorithm training images deployed in the target image warehouse corresponding to the target coal mine are trained to obtain each initial AI application scenario algorithm image; based on the test sample data, each initial AI application scenario algorithm image is verified to obtain each corresponding AI application scenario algorithm image when the verification result is less than the preset error threshold; based on the algorithm quality inspection image, the harmonic mean, precision and recall value of each AI application scenario algorithm image are calculated to quality inspect multiple target AI application scenario algorithm images in each AI application scenario algorithm image that are better than each AI application scenario algorithm; multiple target AI application scenario algorithm images are published to the algorithm deployment image to replace the corresponding algorithm image in the AI ​​application scenario algorithm mall corresponding to the coal mine, otherwise they are not replaced. Therefore, by selecting algorithms from the AI ​​application scenario algorithm mall and training AI application scenario algorithm images, the difficulty of iterative upgrades of coal mine AI application scenario algorithms can be reduced, meeting the coal mines' ever-changing new demands for intelligence, and at the same time better adapting to changes in the coal mine site environment and the development of coal mine intelligence needs.

[0066] In summary, in order to better understand the present invention, an embodiment of the present invention provides an implementation example diagram of the interaction of multiple AI application scenarios in a coal mine, such as Figure 3As shown, when there is an AI application scenario algorithm corresponding to any target coal mine in the AI ​​application scenario algorithm mall, the corresponding AI application scenario algorithm of the target coal mine is selected, and the algorithm image (algorithm training image, algorithm quality inspection image and algorithm deployment image) of the AI ​​application scenario algorithm uploaded in the AI ​​application scenario algorithm mall is used to determine the training sample data, training parameters and test sample data, quality inspection parameters, and initiate the training and quality inspection (verification) of the algorithm image, and then use the above as part of the production algorithm task management (training and quality inspection support re-editing and re-execution operations), and then check the production algorithm task details, execution progress, training log, quality inspection report, algorithm file, etc. to support download, and determine that there is a target AI application scenario algorithm image in the AI ​​application scenario algorithm image that is better than each AI application scenario algorithm, and upload the target AI application scenario algorithm image to the AI ​​application scenario algorithm mall, otherwise repeat the training and quality inspection of the algorithm image, where the training set and test set support automatic division in proportion.

[0067] like Figure 3 As shown, when there is no AI application scenario algorithm corresponding to any target coal mine in the AI ​​application scenario algorithm mall, the corresponding AI application scenario algorithm of the target coal mine is created, and the algorithm image (algorithm training image, algorithm quality inspection image and algorithm deployment image) of the AI ​​application scenario algorithm is uploaded to the image warehouse, and the training and quality inspection (verification) of the algorithm image are initiated for the determined training sample data, training parameters and test sample data, quality inspection parameters, and then the above is used as another part of the production algorithm task management (AI application scenario algorithm and algorithm image can be re-uploaded), and then the production algorithm task details, execution progress, training log, quality inspection report, algorithm file, etc. are checked for download support. When it is determined that there is a target AI application scenario algorithm image in the AI ​​application scenario algorithm image that is superior to each AI application scenario algorithm, the target AI application scenario algorithm image is uploaded to the AI ​​application scenario algorithm mall.

[0068] In addition, an embodiment of the present invention provides an implementation example diagram of a coal mine multi-AI application scenario algorithm, such as Figure 4As shown, select the AI ​​application scenario algorithm mall corresponding to any target coal mine. When the AI ​​application scenario algorithm corresponding to the target coal mine exists in the AI ​​application scenario algorithm mall, determine the edge computing controller of the target coal mine corresponding to the AI ​​application scenario algorithm, and then convert the AI ​​application scenario algorithm according to the corresponding AI computing chip. The converted AI application scenario algorithm is repackaged into the algorithm deployment image according to the rules (training, quality inspection), and the AI ​​application scenario algorithm is sent to the coal mine GPU server or the edge computing controller for underground mining through the algorithm deployment image, so as to perform AI application scenario algorithm image parsing to start the parsing. AI application scenario algorithm image, AI application scenario algorithm mall response; if the AI ​​application scenario algorithm corresponding to the target coal mine does not exist in the AI ​​application scenario algorithm mall, create the corresponding AI application scenario algorithm for the target coal mine, and train and quality inspect the algorithm image corresponding to the AI ​​application scenario algorithm to obtain the AI ​​application scenario algorithm after quality inspection, and then send the quality inspected AI application scenario algorithm to the coal mine's GPU server or underground mining edge computing power controller through the algorithm deployment image, and then parse the quality inspected AI application scenario algorithm image to start the parsed AI application scenario algorithm image, and the AI ​​application scenario algorithm mall responds.

[0069] In order to implement the above embodiments, the present invention also proposes a coal mine multi-AI application scenario algorithm interaction device.

[0070] Figure 5 A schematic diagram of the structure of a coal mine multi-AI application scenario algorithm interaction device provided in an embodiment of the present invention.

[0071] like Figure 5 As shown, the coal mine multi-AI application scenario algorithm interaction device 50 includes: an acquisition module 51, a deployment module 52, a training module 53, and a replacement module 54.

[0072] An acquisition module 51 is used to acquire coal mine sample data when multiple coal mines are operating underground;

[0073] A deployment module 52 is used for deploying the algorithm images of various coal mine underground AI application scenario algorithms in the image warehouse corresponding to the AI ​​application scenario algorithm mall, which has a variety of selectable coal mine underground AI application scenario algorithms built in the AI ​​application scenario algorithm mall;

[0074] The training module 53 is used to train various algorithm images deployed in the target image warehouse corresponding to the target coal mine based on the sample data of any target coal mine, so as to obtain algorithm images for various AI application scenarios;

[0075] The replacement module 54 is used to quality inspect multiple target AI application scenario algorithm images that are superior to each of the AI ​​application scenario algorithms based on the harmonic mean, precision and recall values ​​of each of the AI ​​application scenario algorithm images, and replace the corresponding algorithm images in the AI ​​application scenario algorithm mall corresponding to the coal mine with the multiple target AI application scenario algorithm images, otherwise they will not be replaced.

[0076] Furthermore, in a possible implementation of the embodiment of the present invention, the acquisition module 51 is specifically configured to:

[0077] The videos, pictures, point clouds and audio data of each coal mine during its underground operation are classified and labeled to obtain different types of initial coal mine operation data;

[0078] The different types of initial coal mine operation data corresponding to each coal mine are used as the coal mine sample data when each coal mine is operating underground.

[0079] Furthermore, in a possible implementation of an embodiment of the present invention, the algorithm image includes an algorithm training image, an algorithm quality inspection image, and an algorithm deployment image.

[0080] Furthermore, in a possible implementation of the embodiment of the present invention, the training module 53 is specifically used to:

[0081] Divide the sample data of any target coal mine into training sample data and test sample data according to a preset ratio;

[0082] Based on the training sample data, various algorithm training images deployed in the target image warehouse corresponding to the target coal mine are trained to obtain algorithm images for various initial AI application scenarios;

[0083] Based on the test sample data, each of the initial AI application scenario algorithm images is verified to obtain each corresponding AI application scenario algorithm image when the verification result is less than a preset error threshold.

[0084] Furthermore, in a possible implementation of the embodiment of the present invention, the replacement module 54 is specifically configured to:

[0085] Based on the algorithm quality inspection image, the harmonic mean, precision and recall values ​​of each of the AI ​​application scenario algorithm images are calculated to quality inspect multiple target AI application scenario algorithm images that are superior to each of the AI ​​application scenario algorithms in each of the AI ​​application scenario algorithm images;

[0086] Publish multiple target AI application scenario algorithm images to the algorithm deployment image to replace the corresponding algorithm images in the AI ​​application scenario algorithm mall corresponding to the coal mine, otherwise they will not be replaced.

[0087] It should be noted that the above explanation of the method embodiment is also applicable to the device of this embodiment, and will not be repeated here.

[0088] In the coal mine multi-AI application scenario algorithm interaction device of the embodiment of the present invention, a plurality of selectable coal mine underground AI application scenario algorithms are built in the AI ​​application scenario algorithm mall corresponding to the coal mine, and various algorithm images are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall; based on the sample data of any target coal mine, various algorithm images corresponding to the target coal mine are trained to obtain various AI application scenario algorithm images; multiple target AI application scenario algorithm images that are superior to various AI application scenario algorithms in each AI application scenario algorithm image are quality inspected to replace the corresponding algorithm images in the AI ​​application scenario algorithm mall, thereby, based on the coal mine sample data, accurate training and deployment of coal mine AI application scenario algorithms are realized, and for the differences between different coal mines with the same AI application scenario algorithm, the threshold for rapid application of AI application scenario algorithms to coal mine sites is lowered.

[0089] In order to implement the above embodiment, the present invention further provides an electronic device, including:

[0090] at least one processor; and

[0091] a memory communicatively connected to the at least one processor; wherein,

[0092] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the aforementioned method.

[0093] In order to implement the above embodiment, the present invention further proposes a non-transitory computer-readable storage medium storing computer instructions, where the computer instructions are used to enable the computer to execute the above method.

[0094] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0095] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0096] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0097] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute the instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0098] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0099] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0100] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0101] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A coal mine multi-AI application scenario algorithm interaction method, characterized in that: The method comprises: Obtain coal mine sample data from multiple coal mines operating underground; The AI ​​application scenario algorithm mall corresponding to coal mines has a variety of optional coal mine underground AI application scenario algorithms built in, and the algorithm images of various coal mine underground AI application scenario algorithms are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall, wherein the algorithm image includes an algorithm training image, an algorithm quality inspection image, and an algorithm deployment image, and also includes: by publishing the algorithm training image to the AI ​​application scenario algorithm mall, the basic model of artificial intelligence, the training program, the deep learning framework operating environment, and the descriptive files are packaged into a Docker image and pushed to the dedicated Docker image warehouse corresponding to the AI ​​application scenario algorithm mall; Based on the sample data of any target coal mine, various algorithm images deployed in the target image warehouse corresponding to the target coal mine are trained to obtain algorithm images for various AI application scenarios; Based on the harmonic mean, precision and recall values ​​of each of the AI ​​application scenario algorithm images, multiple target AI application scenario algorithm images that are superior to each of the AI ​​application scenario algorithms are quality-checked, and the multiple target AI application scenario algorithm images are used to replace the corresponding algorithm images in the AI ​​application scenario algorithm mall corresponding to the coal mine, otherwise they will not be replaced.

2. The method according to claim 1, characterized in that: The method of obtaining coal mine sample data of multiple coal mines in operation underground includes: The videos, pictures, point clouds and audio data of each coal mine during its underground operation are classified and labeled to obtain different types of initial coal mine operation data; The different types of initial coal mine operation data corresponding to each coal mine are used as the coal mine sample data when each coal mine is operating underground.

3. The method according to claim 1, characterized in that Based on the sample data of any target coal mine, various algorithm images deployed in the target image warehouse corresponding to the target coal mine are trained to obtain algorithm images for various AI application scenarios, including: Divide the sample data of any target coal mine into training sample data and test sample data according to a preset ratio; Based on the training sample data, various algorithm training images deployed in the target image warehouse corresponding to the target coal mine are trained to obtain algorithm images for various initial AI application scenarios; Based on the test sample data, each of the initial AI application scenario algorithm images is verified to obtain each corresponding AI application scenario algorithm image when the verification result is less than a preset error threshold.

4. The method according to claim 1, characterized in that Based on the harmonic mean, precision and recall values ​​of each of the AI ​​application scenario algorithm images, multiple target AI application scenario algorithm images that are superior to each of the AI ​​application scenario algorithms are quality-checked, and the multiple target AI application scenario algorithm images are replaced with the corresponding algorithm images in the AI ​​application scenario algorithm mall corresponding to the coal mine, otherwise, they are not replaced, including: Based on the algorithm quality inspection image, the harmonic mean, precision and recall values ​​of each of the AI ​​application scenario algorithm images are calculated to quality inspect multiple target AI application scenario algorithm images that are superior to each of the AI ​​application scenario algorithms in each of the AI ​​application scenario algorithm images; Publish multiple target AI application scenario algorithm images to the algorithm deployment image to replace the corresponding algorithm images in the AI ​​application scenario algorithm mall corresponding to the coal mine, otherwise they will not be replaced.

5. A coal mine multi-AI application scenario algorithm interaction device, characterized in that: The device comprises: An acquisition module, used to acquire coal mine sample data when multiple coal mines are operating underground; A deployment module is used for the AI ​​application scenario algorithm mall corresponding to coal mines, which has a variety of selectable coal mine underground AI application scenario algorithms built in. The algorithm images of various coal mine underground AI application scenario algorithms are deployed in the image warehouse corresponding to the AI ​​application scenario algorithm mall, wherein the algorithm images include algorithm training images, algorithm quality inspection images, and algorithm deployment images, and also includes: by publishing the algorithm training image to the AI ​​application scenario algorithm mall, the basic model of artificial intelligence, the training program, the deep learning framework operating environment, and the descriptive files are packaged into a Docker image and pushed to the dedicated Docker image warehouse corresponding to the AI ​​application scenario algorithm mall; The training module is used to train various algorithm images deployed in the target image warehouse corresponding to the target coal mine based on the sample data of any target coal mine, so as to obtain the algorithm images of various AI application scenarios; A replacement module is used to quality-check multiple target AI application scenario algorithm images that are superior to each of the AI ​​application scenario algorithms based on the harmonic mean, precision and recall values ​​of each of the AI ​​application scenario algorithm images, and replace the corresponding algorithm images in the AI ​​application scenario algorithm mall corresponding to the coal mine with the multiple target AI application scenario algorithm images, otherwise they will not be replaced.

6. The device according to claim 5, characterized in that The acquisition module is specifically used for: The videos, pictures, point clouds and audio data of each coal mine during its underground operation are classified and labeled to obtain different types of initial coal mine operation data; The different types of initial coal mine operation data corresponding to each coal mine are used as the coal mine sample data when each coal mine is operating underground.

7. An electronic device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 4.

8. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-4.

Images