Coal mine face recognition method and device based on deep learning

Abstract

The invention relates to a coal mine face recognition method based on deep learning, and belongs to the technical field of face recognition, and the method comprises the following steps: S1, obtaining a to-be-recognized image, the to-be-recognized image comprises a face part and a background part, and the face area of the face part at least accounts for more than one half of the to-be-recognized image; s2, converting the image to be recognized into a gray matrix; s3, inputting the gray matrix into a convolutional neural network for recognition, firstly recognizing a background part, recognizing a human face under the condition that the background part is successfully recognized, matching a human face recognition result with human face images in a database, and matching one human face recognition result with a plurality of similar human face images; s4, constructing a random forest model; and S5, inputting the face recognition result and the similar face images matched with the face recognition result into the random forest model, and predicting the most accurate result in the similar face images matched with the face recognition result.

Description

Technical Field

[0001] The present invention belongs to the technical field of face recognition and relates to a coal mine face recognition method and device based on deep learning. Background Art

[0002] Face recognition technology has been a hot research topic in the field of image processing and recognition in recent years. It uses facial features to identify people and is a type of human biometric recognition technology. In practice, face recognition has broad application prospects in the fields of information security, criminal investigation, access control, etc. In specific working environments such as underground tunnels in coal mines, the conditions are complex, and the video surveillance system on the ground is difficult to be directly applied to underground tunnels. Specific research needs to be done based on the actual environmental conditions.

[0003] The face recognition systems used in daily life have high recognition rates and high recognition speeds, but this is done under very good recognition environment conditions. The environment in a coal mine is relatively harsh, with many conditions such as occlusion, lighting, and coal ash, which greatly reduce the efficiency of face recognition. Therefore, there are still many problems with face recognition in a coal mine environment. The light in a coal mine is generally dim, and coal miners often have coal stains on their faces. These conditions will lead to a lower recognition rate. Coal mines are high-risk sites. Once the recognition rate is low, if a worker cannot be recognized for half a day, it will cause congestion for the workers behind, thereby increasing the risk. Summary of the invention

[0004] In view of this, the object of the present invention is to provide a coal mine face recognition method and device based on deep learning.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] On the one hand, the present invention provides a coal mine face recognition method based on deep learning, comprising the following steps:

[0007] S1: Acquire an image to be identified, wherein the image to be identified includes a face portion and a background portion, and the face area of ​​the face portion occupies at least one half of the image to be identified;

[0008] S2: converting the image to be recognized into a grayscale matrix;

[0009] S3: Input the grayscale matrix into a convolutional neural network for recognition, first recognize the background part, and if the background part is successfully recognized, recognize the face, and match the face recognition result with the face image in the database, so that one face recognition result matches multiple similar face images;

[0010] S4: Build a random forest model;

[0011] S5: Inputting the face recognition result and the similar face image matched by it into the random forest model to predict the most accurate result among the similar face images matched by the face recognition result.

[0012] Furthermore, the step S2 of converting the image to be identified into a grayscale matrix specifically includes the following steps:

[0013] S21: converting the image to be recognized into a grayscale image;

[0014] S22: Calculate a grayscale matrix corresponding to the grayscale image, where the grayscale matrix includes grayscale probability values ​​corresponding to each pixel in the grayscale image.

[0015] Furthermore, in step S3, the background part of the image to be identified is first identified, and the working state of the face recognition device is preliminarily determined. If the background part of the image to be identified is not identified successfully, the identification is performed again until the preset maximum number of times is reached. If the identification is still unsuccessful, a fault alarm is issued.

[0016] Furthermore, the background part is identified in step S3 by globally and locally comparing the image to be identified with the background of the image collected in the database. When the background part of the image to be identified is the same as the background of the image collected in the database, it is determined that the identification is successful, otherwise it is determined that the identification fails.

[0017] Further, the random forest model is constructed in step S4, specifically including the following steps:

[0018] S41: Taking the face recognition result as the root node sample set;

[0019] S42: Randomly select a preset number of samples from the root node sample set as samples of the next branch node;

[0020] S43: Calculate the entropy value of the root node according to the samples of each of the lower branch nodes;

[0021] S44: Calculate the entropy value of the lower layer node according to the predicted probability of the sample after the sample is segmented through the root node;

[0022] S45: Calculate information gain according to the entropy value of the root node and the entropy value of the lower node;

[0023] S46: Establishing a prediction model according to the information gain;

[0024] S47: Construct the random forest model according to the prediction model.

[0025] Furthermore, the step S41 is to use all the face recognition results as the root node sample set, or to perform random independent repeated sampling on the face recognition results to obtain the root node sample set.

[0026] Further, the step S5 specifically includes the following steps:

[0027] S51: using the face recognition result as the input of the random forest model, and obtaining the prediction probability of each prediction model through each prediction model;

[0028] S52: According to the prediction probabilities of the prediction models, a recognition result with the highest prediction probability is obtained, which is the prediction result.

[0029] On the other hand, the present invention provides a coal mine face recognition system based on deep learning, comprising:

[0030] An acquisition module, used for acquiring an image to be recognized, wherein the image to be recognized includes a face portion and a background portion, and the face area of ​​the face portion occupies at least one half of the image to be recognized;

[0031] A conversion module, used for converting the image to be recognized into a grayscale matrix;

[0032] A recognition module inputs the grayscale matrix into a convolutional neural network for recognition, firstly recognizes the background part, and when the background part is successfully recognized, recognizes the face, and matches the face recognition result with the face image in the database, so that one face recognition result matches multiple similar face images;

[0033] Building modules for building random forest models;

[0034] The prediction module is used to input the face recognition result and the similar face image matched by it into the random forest model to predict the most accurate result among the similar face images matched by the face recognition result.

[0035] Furthermore, the conversion module comprises:

[0036] The first submodule is used to convert the image to be recognized into a grayscale image;

[0037] The second submodule is used to calculate the grayscale matrix corresponding to the grayscale image, and the grayscale matrix includes the grayscale probability value corresponding to each pixel in the grayscale image.

[0038] Furthermore, the building blocks include:

[0039] A split submodule is used to use the face recognition result as a root node sample set; randomly select a preset number of samples from the root node sample set as samples of the next branch node;

[0040] The first calculation module is used to calculate the entropy value of the root node according to the samples of each of the lower branch nodes; and calculate the entropy value of the lower layer node according to the predicted probability of the sample after the sample is segmented through the root node;

[0041] A second calculation module, used to calculate information gain according to the entropy value of the root node and the entropy value of the lower node;

[0042] A module is established to establish a prediction model according to the information gain; and the random forest model is constructed according to the prediction model.

[0043] The beneficial effects of the present invention are as follows: the present invention converts the image to be identified into a grayscale matrix for algebraic identification, and by constructing a random forest model, adaptively calculates the most accurate identification result from multiple similar results identified, so as to improve the accuracy of face recognition in underground coal mine environments.

[0044] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0046] Figure 1 It is a schematic diagram of the process of the coal mine face recognition method based on deep learning described in an embodiment of the present invention;

[0047] Figure 2 It is a block diagram of a coal mine face recognition system based on deep learning described in an embodiment of the present invention;

[0048] Figure 3 It is a block diagram of a coal mine face recognition device based on deep learning described in an embodiment of the present invention;

[0049] Reference numerals: 710 - acquisition module; 720 - conversion module; 730 - recognition module; 740 - construction module; 750 - prediction module; 800 - coal mine face recognition device; 801 - processor; 802 - memory; 803 - multimedia component; 804 - I / O interface; 805 - communication component DETAILED DESCRIPTION

[0050] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0051] Among them, the drawings are only used for illustrative explanations, and they only represent schematic diagrams rather than actual pictures, and should not be understood as limitations on the present invention. In order to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0052] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "front", "rear", etc. indicate the orientation or position relationship, they are based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0053] In the first aspect, the present application provides a coal mine face recognition method based on deep learning, comprising:

[0054] Acquire an image to be identified, wherein the image to be identified includes a face portion and a background portion, and the face area of ​​the face portion is at least one-half;

[0055] Converting the image to be recognized into a grayscale matrix;

[0056] Recognize the grayscale matrix with a convolutional neural network to obtain a recognition result, wherein the recognition result includes a face recognition result and a background recognition result, and the face recognition result is a comparison result between the image to be recognized and the face image in the database;

[0057] Under the condition that the background recognition result is a successful recognition, constructing a random forest model according to the face recognition result;

[0058] The recognition result is used as the input of the random forest model to obtain a prediction result, where the prediction result is the recognition result with the highest face recognition accuracy among the recognition results.

[0059] In the second aspect, the present application also provides a coal mine face recognition system based on deep learning, including:

[0060] An acquisition module, used for acquiring an image to be recognized, wherein the image to be recognized includes a face portion and a background portion, and the face area of ​​the face portion is at least one-half;

[0061] A conversion module, used for converting the image to be recognized into a grayscale matrix;

[0062] A recognition module, used to recognize the grayscale matrix using a convolutional neural network to obtain a recognition result, wherein the recognition result includes a face recognition result and a background recognition result, and the face recognition result is a comparison result between the image to be recognized and the face image in the database;

[0063] A construction module, used for constructing a random forest model according to the face recognition result under the condition that the background recognition result is a successful recognition;

[0064] The prediction module is used to use the recognition result as the input of the random forest model to obtain a prediction result, wherein the prediction result is the recognition result with the highest face recognition accuracy among the recognition results.

[0065] Embodiment 1:

[0066] See also Figure 1 , Figure 1 The figure shows a flowchart of a coal mine face recognition method based on deep learning in this embodiment. This embodiment provides a coal mine face recognition method based on deep learning, and the method includes:

[0067] Step S1, obtaining an image to be identified, wherein the image to be identified includes a face portion and a background portion, and the face area of ​​the face portion is at least one-half;

[0068] Step S2, converting the image to be identified into a grayscale matrix;

[0069] Step S3, identifying the grayscale matrix with a convolutional neural network to obtain a recognition result, wherein the recognition result includes a face recognition result and a background recognition result, and the face recognition result is a comparison result between the image to be identified and a face image in a database;

[0070] Step S4: under the condition that the background recognition result is a successful recognition, construct a random forest model according to the face recognition result;

[0071] Step S5: using the recognition result as the input of the random forest model to obtain a prediction result, wherein the prediction result is the recognition result with the highest face recognition accuracy among the recognition results.

[0072] It can be seen that the above method converts the image to be identified into a grayscale matrix for algebraic recognition, and adaptively calculates the most accurate recognition result from multiple similar results identified by constructing a random forest model, so as to improve the accuracy of face recognition in underground coal mine environments.

[0073] In this embodiment, considering that the underground position is relatively fixed when face recognition is performed underground, the background part of the image to be recognized is recognized first, so that the working status of the face recognition device can be preliminarily determined. If the background part of the image to be recognized is not recognized successfully, it can be recognized again. After a preset number of recognitions, if it is still unsuccessful, it can be determined that the face recognition device has a fault.

[0074] Under the condition that the background part is successfully recognized, face recognition can further improve the recognition efficiency.

[0075] In this embodiment, the background part of the image to be identified can be identified by global and local comparison with the background of the image collected in the database to reduce the amount of calculation of the above method. When the background part of the image to be identified is the same as the background of the image collected in the database, it is determined that the recognition is successful, otherwise it is considered a recognition failure.

[0076] In this embodiment, an implementation method of step S2 specifically includes the following methods:

[0077] Step S21, converting the image to be recognized into a grayscale image;

[0078] Step S22: Calculate a grayscale matrix corresponding to the grayscale image, wherein the grayscale matrix includes grayscale probability values ​​corresponding to each pixel in the grayscale image.

[0079] Through the above method, the image to be recognized can be converted into a matrix for deep learning, which can further improve the recognition efficiency and accuracy.

[0080] In this embodiment, one implementation of step S4 is specifically as follows:

[0081] Step S41, taking the face recognition result as the root node sample set;

[0082] Step S42, randomly selecting a preset number of samples from the root node sample set as samples of the lower branch node;

[0083] Step S43, calculating the entropy value of the root node according to the samples of each of the lower branch nodes;

[0084] Step S44, calculating the entropy value of the lower layer nodes according to the predicted probability of the sample after the sample segmentation through the root node;

[0085] Step S45, calculating information gain according to the entropy value of the root node and the entropy value of the lower layer node;

[0086] Step S46: establishing a prediction model according to the information gain;

[0087] Step S47: construct the random forest model according to the prediction model.

[0088] By calculating information gain, multiple layers of nodes are established to construct a prediction model. The above information gain can directly determine the optimal sample of the split branch node, that is, the sample with the highest prediction probability.

[0089] In this embodiment, the implementation method of step S41 is as follows:

[0090] Step S411, taking all the face recognition results as a root node sample set;

[0091] Or step S412, performing random independent repeated sampling on the face recognition results to obtain the root node sample set.

[0092] When the sample number of face recognition results is small, all recognition results can be used as the root node sample set to avoid the small number of samples affecting the prediction results.

[0093] In this embodiment, the implementation of step S5 is specifically as follows:

[0094] Step S51, using the recognition result as the input of the random forest model, and obtaining the prediction probability of each prediction model through each prediction model;

[0095] Step S52: According to the prediction probabilities of the prediction models, the recognition result with the highest prediction probability is obtained, which is the prediction result.

[0096] It can be seen that the constructed random forest model can efficiently and adaptively calculate the recognition result with the highest prediction probability to avoid recognition errors caused by inaccurate recognition results.

[0097] In this embodiment, while obtaining the prediction result, the prediction probability corresponding to the current prediction result can also be obtained, and whether the recognition is successful is determined by setting a threshold, specifically:

[0098] The probability threshold can be set to 0.8 to 0.9. If the prediction probability corresponding to the current prediction result is less than the above probability threshold, it can be determined that the recognition has failed, that is, the image to be recognized does not belong to the face image in the database.

[0099] In this embodiment, the image to be recognized includes a face image collected in real time and a face image after rotation and mirror symmetry. The above method can effectively expand the number of recognition samples for face recognition by the convolutional neural network, thereby improving the recognition accuracy.

[0100] In this embodiment, the following preprocessing method is also included before step S2:

[0101] Step S6: De-noise the image to be identified by Gaussian filtering.

[0102] The above method can improve the recognition effect of the convolutional neural network without affecting the gray matrix data samples.

[0103] Embodiment 2:

[0104] See also Figure 2 , Figure 2 is a block diagram of a coal mine face recognition device based on deep learning according to an exemplary embodiment, including:

[0105] An acquisition module 710 is used to acquire an image to be recognized, wherein the image to be recognized includes a face portion and a background portion, and the face area of ​​the face portion is at least one-half;

[0106] A conversion module 720, used for converting the image to be recognized into a grayscale matrix;

[0107] The recognition module 730 is used to recognize the grayscale matrix using a convolutional neural network to obtain a recognition result, wherein the recognition result includes a face recognition result and a background recognition result, and the face recognition result is a comparison result between the image to be recognized and the face image in the database;

[0108] A construction module 740 is used to construct a random forest model according to the face recognition result under the condition that the background recognition result is a successful recognition;

[0109] The prediction module 750 is used to use the recognition result as the input of the random forest model to obtain a prediction result, where the prediction result is the recognition result with the highest face recognition accuracy among the recognition results.

[0110] It can be seen that the above system converts the image to be identified into a grayscale matrix for algebraic recognition, and adaptively calculates the most accurate recognition result from multiple similar results identified by constructing a random forest model, so as to improve the accuracy of face recognition in underground coal mine environments.

[0111] In this embodiment, the conversion module 720 includes:

[0112] The first submodule is used to convert the image to be recognized into a grayscale image;

[0113] The second submodule is used to calculate the grayscale matrix corresponding to the grayscale image, and the grayscale matrix includes the grayscale probability value corresponding to each pixel in the grayscale image.

[0114] Through the above method, the image to be recognized can be converted into a matrix for deep learning, which can further improve the recognition efficiency and accuracy.

[0115] In this embodiment, the construction module 740 includes:

[0116] A split submodule is used to use the face recognition result as a root node sample set; randomly select a preset number of samples from the root node sample set as samples of the next branch node;

[0117] The first calculation module is used to calculate the entropy value of the root node according to the samples of each of the lower branch nodes; and calculate the entropy value of the lower layer node according to the predicted probability of the sample after the sample is segmented through the root node;

[0118] A second calculation module, used to calculate information gain according to the entropy value of the root node and the entropy value of the lower node;

[0119] A module is established to establish a prediction model according to the information gain; and the random forest model is constructed according to the prediction model.

[0120] By calculating information gain, multiple layers of nodes are established to construct a prediction model. The above information gain can directly determine the optimal sample of the split branch node, that is, the sample with the highest prediction probability.

[0121] It should be noted that, regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment of the method, and will not be elaborated here.

[0122] Embodiment 3:

[0123] Corresponding to the above method embodiment, a coal mine face recognition device 800 based on deep learning is also provided in this embodiment. The coal mine face recognition device 800 based on deep learning described below and the coal mine face recognition method based on deep learning described above can be referenced to each other.

[0124] Figure 3 is a block diagram of a coal mine face recognition device 800 based on deep learning according to an exemplary embodiment. Figure 3As shown, the coal mine face recognition device 800 based on deep learning may include: a processor 801, a memory 802. The coal mine face recognition device 800 based on deep learning may also include one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0125] The processor 801 is used to control the overall operation of the deep learning-based coal mine face recognition device 800 to complete all or part of the steps in the deep learning-based coal mine face recognition method. The memory 802 is used to store various types of data to support the operation of the deep learning-based coal mine face recognition device 800, which may include instructions for any application or method operating on the deep learning-based coal mine face recognition device 800, as well as application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, disk or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 802 or sent via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 804 provides an interface between the processor 801 and other interface modules, and the above-mentioned other interface modules can be keyboards, mice, buttons, etc. These buttons can be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the deep learning-based coal mine face recognition device 800 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G or 4G, or a combination of one or more of them, so the corresponding communication component 805 can include: Wi-Fi module, Bluetooth module, NFC module.

[0126] In an exemplary embodiment, the deep learning-based coal mine face recognition device 800 can be implemented by one or more application-specific integrated circuits (Application Specific Integrated Circuit, referred to as ASIC), digital signal processors (Digital Signal Processor, referred to as DSP), digital signal processing devices (Digital Signal Processing Device, referred to as DSPD), programmable logic devices (Programmable Logic Device, referred to as PLD), field programmable gate arrays (Field Programmable Gate Array, referred to as FPGA), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned deep learning-based coal mine face recognition method.

[0127] In another exemplary embodiment, a computer storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned coal mine face recognition method based on deep learning are implemented. For example, the computer storage medium can be the above-mentioned memory 802 including program instructions, and the above-mentioned program instructions can be executed by the processor 801 of the coal mine face recognition device 800 based on deep learning to complete the above-mentioned coal mine face recognition method based on deep learning.

[0128] Embodiment 4:

[0129] Corresponding to the above method embodiment, a storage medium is also provided in this embodiment. The storage medium described below and the coal mine face recognition method based on deep learning described above can refer to each other.

[0130] A storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the coal mine face recognition method based on deep learning in the above-mentioned method embodiment.

[0131] The storage medium may specifically be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or other storage medium that can store program codes.

[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A coal mine face recognition method based on deep learning, characterized by: The following steps are involved: S1: Acquire an image to be identified, wherein the image to be identified includes a face portion and a background portion, and the face area of ​​the face portion occupies at least one half of the image to be identified; S2: converting the image to be recognized into a grayscale matrix; S3: Input the grayscale matrix into a convolutional neural network for recognition, first recognize the background part, and when the background part is successfully recognized, recognize the face, and match the face recognition result with the face image in the database, so that one face recognition result matches multiple similar face images; S4: Build a random forest model; S5: Inputting the face recognition result and the similar face image matched by it into the random forest model to predict the most accurate result among the similar face images matched by the face recognition result.

2. The coal mine face recognition method based on deep learning according to claim 1 is characterized in that: The step S2 of converting the image to be identified into a grayscale matrix specifically includes the following steps: S21: converting the image to be recognized into a grayscale image; S22: Calculate a grayscale matrix corresponding to the grayscale image, where the grayscale matrix includes grayscale probability values ​​corresponding to each pixel in the grayscale image.

3. The coal mine face recognition method based on deep learning according to claim 1 is characterized in that: In step S3, the background part of the image to be identified is first identified to preliminarily determine the working status of the face recognition device. If the background part of the image to be identified is not successfully identified, the identification is performed again until the preset maximum number of times is reached. If the identification is still unsuccessful, a fault alarm is issued.

4. The coal mine face recognition method based on deep learning according to claim 1 is characterized in that: The background part is identified in step S3 by globally and locally comparing the image to be identified with the background of the image collected in the database. When the background part of the image to be identified is the same as the background of the image collected in the database, it is determined that the identification is successful, otherwise it is determined that the identification fails.

5. The coal mine face recognition method based on deep learning according to claim 1 is characterized in that: The random forest model is constructed in step S4, which specifically includes the following steps: S41: Taking the face recognition result as the root node sample set; S42: Randomly select a preset number of samples from the root node sample set as samples of the next branch node; S43: Calculate the entropy value of the root node according to the samples of each of the lower branch nodes; S44: Calculate the entropy value of the lower layer node according to the predicted probability of the sample after the sample is segmented through the root node; S45: Calculate information gain according to the entropy value of the root node and the entropy value of the lower node; S46: Establishing a prediction model according to the information gain; S47: Construct the random forest model according to the prediction model.

6. The coal mine face recognition method based on deep learning according to claim 5 is characterized in that: The step S41 is to use all the face recognition results as the root node sample set, or to perform random independent repeated sampling on the face recognition results to obtain the root node sample set.

7. The coal mine face recognition method based on deep learning according to claim 1 is characterized in that: The step S5 specifically comprises the following steps: S51: using the face recognition result as the input of the random forest model, and obtaining the prediction probability of each prediction model through each prediction model; S52: According to the prediction probabilities of the prediction models, a recognition result with the highest prediction probability is obtained, which is the prediction result.

8. A coal mine face recognition system based on deep learning, characterized by: include: An acquisition module, used for acquiring an image to be recognized, wherein the image to be recognized includes a face portion and a background portion, and the face area of ​​the face portion occupies at least one half of the image to be recognized; A conversion module, used for converting the image to be recognized into a grayscale matrix; A recognition module inputs the grayscale matrix into a convolutional neural network for recognition, firstly recognizes the background part, and when the background part is successfully recognized, recognizes the face, and matches the face recognition result with the face image in the database, so that one face recognition result matches multiple similar face images; Building modules for building random forest models; The prediction module is used to input the face recognition result and the similar face image matched by it into the random forest model to predict the most accurate result among the similar face images matched by the face recognition result.

9. The coal mine face recognition system based on deep learning according to claim 8, characterized in that: The conversion module comprises: The first submodule is used to convert the image to be recognized into a grayscale image; The second submodule is used to calculate the grayscale matrix corresponding to the grayscale image, and the grayscale matrix includes the grayscale probability value corresponding to each pixel in the grayscale image.

10. The coal mine face recognition system based on deep learning according to claim 8, characterized in that: The building blocks include: A split submodule is used to use the face recognition result as a root node sample set; randomly select a preset number of samples from the root node sample set as samples of the next branch node; The first calculation module is used to calculate the entropy value of the root node according to the samples of each of the lower branch nodes; and calculate the entropy value of the lower layer node according to the predicted probability of the sample after the sample is segmented through the root node; A second calculation module, used to calculate information gain according to the entropy value of the root node and the entropy value of the lower node; A module is established to establish a prediction model according to the information gain; and the random forest model is constructed according to the prediction model.

Images