WO2023050603A1 - Accurate person recognition apparatus and method for using same - Google Patents

Abstract

An accurate person recognition apparatus and a method for using same. The apparatus comprises a facial recognition module (3), which is mounted on a side face of a coal conveyor belt (1); a limb feature extraction unit (2), which is mounted on the side face of the coal conveyor belt (1) and is located below the facial recognition module (3); a connection unit (4), which is fixedly mounted above the coal conveyor belt (1); a driving unit (5), which is fixedly mounted above the connection unit (4); a working unit (9), which is fixedly mounted below the connection unit (4); and a conveying amount distribution board (10), a conveying amount extension board (11) and a conveying amount equal-distribution board (12), which are connected to the bottom of the working unit (9). An intention of an inspector is determined by means of the combination of a facial expression and limb performance, such that the accuracy of person recognition is significantly improved, and a blank area for rapid conveying is formed in an inspection area of the inspector without affecting subsequent coal conveying, such that the inspection of a conveyor belt by the inspector is facilitated; moreover, the effect of a noise factor on the accurate recognition of a person is reduced, thereby further improving the accuracy of person state recognition.

Description

An accurate personnel identification device and its application method technical field

The invention relates to the field of personal identification devices, in particular to an accurate personal identification device and a use method.

Background technique

The general workflow of the personnel emotion recognition device is as follows: 1) input the picture into the convolutional neural network, and extract the feature map information of the picture through convolution operation and pooling operation; 2) the feature map information generates candidate areas through the area generation network, and recommend The candidate area of ∗ only means that there is a target, and the classification of the target is completed in the next step; 3) For the recommended candidate area, ROI pooling is performed, and the input features of different lengths are converted into output features of fixed length. For the neural network VGG, ROI pooling The output of the transformation layer is a dimensional feature vector, and the fixed-length feature vector is classified and regressed to obtain the final classification result.

During the process of transporting coal to the ground by the conveyor belt under the mine, due to the long-term operation of the conveyor belt during the transmission process, some supporting structures will be in a damaged state, and manual timing is required to inspect the conveyor belt. During the inspection and maintenance of the conveyor belt, the conveyor belt cannot be transported, which seriously affects the transmission efficiency of the conveyor belt, and the damaged part cannot be quickly found. In this way, we need to overhaul the conveyor belt in the working state underground, and there will be some problems. During the detection process, we cannot quickly form a blank area for transportation in the inspection area of the inspector. The way the inspector repeatedly presses the switch seriously affects the detection. Efficiency, for this reason we need to quickly identify the inspection intention of the inspectors; and due to the combined effect of shaking and noise in the dark environment of the mine and during the maintenance process, it will have a negative impact on the collection of pictures, sounds and other signals.

In order to solve the above problems, we combine the personnel identification device with the inspection of the conveyor belt under the mine to propose an accurate personnel identification device and its usage method.

Contents of the invention

The main purpose of the present invention is to provide an accurate personnel identification device and its usage method, which can effectively solve the problems in the background technology.

In order to achieve the above object, the technical scheme that the present invention takes is:

An accurate personnel identification device, comprising a facial recognition module installed on the side of a coal conveyor belt, a body feature extraction unit is installed on the side of the coal conveyor belt and below the facial recognition module, and fixedly installed above the coal conveyor belt There is a connection unit, a drive unit is fixedly installed above the connection unit, a working unit is fixedly installed below the connection unit, and the bottom of the working unit is connected with a transportation volume distribution board, a transportation volume extension board and a transportation volume equalizer The middle part of the working unit is slidably connected with a noise buffer unit, and the end of the noise buffer unit is fixedly installed with an end buffer sleeve.

The further improvement of the present invention lies in that the front end of the transportation volume distribution plate is fixedly equipped with a distribution tip, the upper surface of the transportation volume distribution plate is fixedly installed with connecting rods, and the two sides of the bottom end of the connecting rod are fixedly installed with upper stops The tail of the transportation volume distribution plate is fixedly equipped with a tail arc-shaped plate, and the sides of the transportation volume distribution board, transportation volume extension board and transportation volume equalizing board are all rotatably connected with side transmission wheels, and the connecting rods A driven gear is fixedly installed on the upper end, and a connecting flexible board is fixedly installed on the front end of the transportation volume extension board, and the end of the connection flexible board far away from the transportation volume extension board is fixedly connected with the tail arc-shaped board.

Through the above structure, it can be realized that the body feature extraction unit and the facial recognition module can accurately identify the intention of the inspector, and when it is recognized that the inspector has just moved to the designated position, the driving unit will drive the transportation volume equalizing board and the transportation volume distribution board Rotate to move the coal on the surface of the coal conveyor belt to the side of the capacity inspector, and form a blank area for transportation in the inspection area of the inspector, so as to facilitate the inspection of the conveyor belt by the inspector.

The further improvement of the present invention lies in that the structure of the transportation volume distribution plate is the same as that of the transportation volume distribution board, the rotation direction of the transportation volume distribution plate is opposite to that of the transportation volume distribution board, and the tail end of the transportation volume extension board is set as Arc-shaped, the tail end of the transportation volume equalizing plate is fixedly installed with a limited arc-shaped plate, and the upper end of the transportation volume equalizing plate is fixedly equipped with a No. 2 gear.

Through the above structure, it can be realized that the transportation volume equalizing plate and the transportation volume distribution plate rotate in opposite directions, and at the same time, under the action of the arc set at the tail end of the transportation volume extension plate, the increased transportation volume on the inspection side moves along the transportation volume equalizing plate, transportation volume The extension plate and the transportation volume distribution plate cross the inspection site and reach the rear side of the inspection site again, thereby preventing the inspection process from affecting the transportation of the subsequent transportation section.

A further improvement of the present invention is that the number of the working unit is the same as that of the face recognition module, the output end of the driving unit extends to the inside of the working unit, and its end is connected to a driving wheel, and the two sides of the driving wheel Mesh transmission with No. 2 gear and driven gear respectively.

Through the above structure, it can be realized that the driving wheel at the output end of the driving unit is set to drive the second gear and the driven gear to rotate, thereby driving the transportation volume equalizing plate and the transportation volume distribution plate to rotate in reverse.

The further improvement of the present invention lies in that the inside of the end buffer sleeve is rotatably connected with cooling fins, the end of the end buffer sleeve is fixedly installed with a protective frame, and the inner surface of the protective frame is fixedly installed with an expansion bag body.

Through the above structure, it can be realized that under the action of the heat dissipation fan, the heat of the gas is exported by the heat dissipation fan, and at the same time, the gas drives the heat dissipation fan to rotate to quickly cool down the gas, so that in the process of providing buffer for the maintenance personnel and the facial recognition module, Provide cooling to the facial recognition module and personnel during the maintenance process to reduce the risk of accidents during the maintenance process.

The further improvement of the present invention lies in that two airflow transmission ports are opened on the side of the upper surface of the noise buffer unit, and the airflow transmission ports are connected with the inside of the end buffer sleeve, and one side of the noise buffer unit is respectively opened. rack and the No. 2 rack, and the No. 2 rack is meshed with the driven gear for transmission.

A further improvement of the present invention is that a gas transmission module is fixedly installed in the middle of the working unit, and transmission units are provided on both sides of the gas transmission module. When the noise buffer unit moves to the end, the transmission unit The output end is in communication with the airflow transmission port, the bottom of the working unit is fixedly equipped with an airflow circulation pipe, and the output end of the end buffer sleeve is connected with the airflow circulation pipe.

Through the above structure, it can be realized that the drive unit drives the noise buffer unit and the end buffer sleeves to move When the inner buffer sleeve moves to the right below the facial recognition module, the transmission unit is connected to the airflow transmission port, and the gas expands the expansion capsule to provide buffer for the facial recognition module, and at the same time, the lower part of the expansion capsule provides protection for the maintenance personnel, so that the face The identification module clearly identifies the inspectors during the maintenance process, reducing the risk of accidents during the maintenance process.

The further improvement of the present invention is that a connecting cavity is provided inside the working unit, and a sliding unit is slidably connected inside the connecting cavity, and the sliding unit slides inside the connecting cavity, and the upper surface of the sliding unit is fixedly installed There is an upper slide plate, which is sealed to the transmission unit, and the bottom of the slide unit is meshed with a meshing gear, and the meshing gear is rotatably installed on the inner edge of the connecting cavity.

Through the above structure, it can be realized that the No. 1 rack is meshed with the meshing gear, and the meshing gear is meshed with the sliding unit on the upper side for transmission. In this way, when the inspector needs to repair the place, the noise buffer unit and the end buffer sleeve will go here. The moving pair of gas inflates the inflatable bladder to provide cushioning for the facial recognition module, while the underside of the inflated bladder provides protection to the operator.

A further improvement of the present invention is that the facial recognition module includes a noise extraction unit, an accurate facial expression recognition unit, an identity prior unit, and a control unit, and the accurate facial expression recognition unit is suitable for inputting a detected picture into a convolutional neural network, through Convolution operation and pooling operation to extract the feature map information of the picture, and compare the extracted feature map information with the expression image stored in the storage to accurately identify the type of employee expression information. The expression types are: relaxed, thinking or painful , the facial expression feature extraction unit shoots the same scene through two cameras with different focal lengths, so as to solve the over-fitting problem in scene reproduction.

The identity prior unit is suitable for checking the face and physical features of the inspector, and using the face check to extract the feature map as the face prior information, and adding the facial expression to the face check and extract feature map Distributed to the control unit.

The body feature extraction unit is adapted to extract the body features, and compare with the body features in the repository to identify the maintenance process of the device by the inspector.

The control unit is suitable for introducing the real face image and the face check extraction feature map extracted by the facial expression feature extraction unit in the dark alley into the discriminator, so as to judge the expression of the real face image at the feature level. Under the joint action of the feature extraction unit, the body features of the inspector are extracted, and the facial expression and body performance cooperate to judge the intention of the inspector, and at the same time control the work of the noise buffer unit and the working unit.

The noise extraction unit is suitable for extracting unclear noise information captured by the monitoring probe, and comparing it with the noise information in the noise library to find the cause of the noise.

Through the above structure, it can be realized that the recognition module is equipped with a facial expression feature extraction unit and an accurate facial expression recognition unit. With the extraction of audio information, the control unit can obtain the tiredness, relaxation or pain of the employees at the same station on both sides of the production line. The contrast coefficient, and control the work of the drive unit through the fatigue contrast value.

A further improvement of the present invention lies in a method for using an accurate personnel identification device, the steps of which are as follows:

A: Coal is transported on the coal conveyor belt. Before the inspectors inspect the coal conveyor belt, the identity prior unit checks the face and physical features of the inspector, and uses the face check to extract the feature map as the identity map. Face prior information, the feature map extracted by face nuclear inspection is added with expression, and then the face recognition model is constructed and distributed to the discriminator of the control unit;

B: After completing step A, each time the inspectors arrive at an inspection point, the real face image and the face verification extraction feature map extracted by the facial expression feature extraction unit in the dark alleyway are introduced into the discriminator to judge at the feature level The expression of the real face image, and at the same time, the body features of the inspector are extracted under the joint action of the body feature extraction unit. The facial expression and body performance cooperate to judge the intention of the inspector, and it is recognized that the inspector has just moved here When inspecting here, the driving unit drives the transport volume equalizing plate and the transport volume distribution plate to rotate, and moves the coal on the surface of the coal conveyor belt to the side of the capacity inspector, forming a blank for transportation in the inspection area of the inspector area, and does not affect the coal transmission behind, so as to facilitate the inspection of the conveyor belt by the inspector;

C: After completing step B, if noise is generated during the inspection process and the facial recognition module is not clearly recognized, extract the unclear noise information captured by the monitoring probe and compare it with the noise information in the noise library to find the source of the noise The reason is that at the same time, the body feature extraction unit and the facial recognition module recognize the frequent body movements of the inspector, which means that the inspector is here for maintenance, and the noise is caused by the maintenance process;

D: At the same time as step C is completed, the drive unit drives the noise buffer unit and the end buffer sleeve to move toward the inspector’s inspection position while driving the transportation volume equalizer plate and the transportation volume distribution plate to rotate. When the noise buffer unit When the end buffer sleeve and the end buffer cover are moved to directly below the facial recognition module, the transmission unit is connected to the airflow transmission port, and the gas expands the expansion capsule to provide buffer for the facial recognition module, and at the same time, the lower part of the expansion capsule provides protection for the maintenance personnel, so as to Make the facial recognition module clearly identify the inspectors during the maintenance process, reducing the risk of accidents during the maintenance process.

Compared with the prior art, the present invention has the following beneficial effects:

1. Through the combined use of the set face recognition module and the limb feature extraction unit, the identity prior unit first checks the face and physical features of the inspector, and uses the face check to extract the feature map as the face prior information , add the facial expression feature map and distribute it to the control unit, and introduce the real face image and face feature map extracted by the facial expression feature extraction unit in the dark alleyway into the discriminator to Judging the expression of the real face image at the feature level, and at the same time extracting the body features of the inspector under the joint action of the body feature extraction unit, the facial expression and body performance cooperate to judge the intention of the inspector, which significantly improves the accuracy of personnel recognition accuracy.

2. Through the combined use of the body feature extraction unit, facial recognition module, work unit, transportation volume distribution board, transportation volume extension board and transportation volume equalization board, the body feature extraction unit and facial recognition module will detect the intention of the inspector Accurate identification, when the inspector has just moved here to inspect this place, the drive unit drives the transportation volume equalizing plate and the transportation volume distribution plate to rotate, and moves the coal on the surface of the coal conveyor belt to the side of the capacity inspection personnel , forming a blank area for transportation in the inspection area of the inspector, and does not affect the subsequent coal transmission, so as to facilitate the inspection of the conveyor belt by the inspector.

The body feature extraction unit and the facial recognition module recognize that the inspector's body movements are frequent, indicating that the inspector is here for maintenance, and the maintenance causes shaking and noise in this section of the conveyor belt, resulting in unclear recognition by the facial recognition module, and the drive unit is driving the transport. During the rotation of the volume equalizing plate and the transportation volume distribution plate, the noise buffer unit and the end buffer sleeve are driven to move toward the inspection position of the inspector at the same time. When the noise buffer unit and the end buffer sleeve move to the right below the face recognition module When the transmission unit is in communication with the airflow transmission port, the gas inflates the expansion capsule to provide a buffer for the facial recognition module, and at the same time, the underside of the expansion capsule provides protection for the maintenance personnel, so that the facial recognition module can identify the inspector during the maintenance process. Clear, reducing the risk of accidents during maintenance.

Description of drawings

FIG. 1 is a schematic diagram of the overall structure of an accurate personal identification device according to the present invention.

Fig. 2 is a schematic diagram of the system connection of an accurate personal identification device according to the present invention.

Fig. 3 is a schematic diagram of the overall structure of a working unit of an accurate person identification device according to the present invention.

Fig. 4 is a schematic diagram of the connection between the transportation volume distribution plate and the connecting rod of an accurate personnel identification device according to the present invention.

Fig. 5 is a schematic diagram of the connection between the noise buffer unit and the end buffer sleeve of an accurate personal identification device according to the present invention.

Fig. 6 is a schematic diagram of connection of a noise buffer unit, an end buffer sleeve and a working unit of an accurate personal identification device according to the present invention.

Fig. 7 is an internal schematic diagram of a working unit of an accurate person identification device according to the present invention.

Fig. 8 is an exploded schematic view of an end buffer sleeve of an accurate personal identification device according to the present invention.

FIG. 9 is a schematic diagram of a method of using an accurate personal identification device according to the present invention.

Instructions for marks in the figure:

1. Coal conveyor belt; 2. Limb feature extraction unit; 3. Facial recognition module; 4. Connection unit; 5. Drive unit; 6. Noise buffer unit; 7. End buffer sleeve; 8. Gas transmission module; 9. Working unit; 10. Transport volume distribution plate; 11. Transport volume extension plate; 12. Transport volume equalizer plate; 13. Distribution tip; 14. Side transmission wheel; 15. Upper baffle plate; 16. Tail arc plate; 17 , driven gear; 18, connecting rod; 31, noise extraction unit; 32, accurate facial expression recognition unit; 33, identity prior unit; 34, control unit; 61, airflow transmission port; 62, No. 1 rack; 63 , No. 2 rack; 71. Protective frame; 72. Expansion capsule; 73. Cooling blades; 91. Air circulation pipe; 92. Transmission unit; 93. Connection chamber; 94. Slide unit; , meshing gear; 97, driving wheel; 98, No. 2 gear; 111, connecting flexible plate; 121, limit arc plate.

Detailed ways

In order to make the technical means realized by the present invention, creative features, goals and effects easy to understand, in the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", " The orientations or positional relationships indicated by "right", "vertical", "horizontal", "inner", "outer", etc. are based on the orientation or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, and It is not to indicate or imply that the device or element referred to must have a particular orientation, be constructed in a particular orientation, or operate in a particular orientation, and thus should not be construed as limiting the invention. In addition, the terms "first", "number one", "number two", "number three", and "number four" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. The present invention will be further described below in combination with specific embodiments.

Example 1

As shown in Figure 1-9, a precise personnel identification device includes a facial recognition module 3 installed on the side of the coal conveyor belt 1, and a limb feature extraction unit is installed on the side of the coal conveyor belt 1 and below the facial recognition module 3 2. A connecting unit 4 is fixedly installed above the coal conveyor belt 1, a drive unit 5 is fixedly installed above the connecting unit 4, a working unit 9 is fixedly installed below the connecting unit 4, and the bottom of the working unit 9 is connected with a transportation volume distribution plate 10. The transportation volume extension board 11 and the transportation volume equalization board 12, the middle part of the working unit 9 is slidingly connected with a noise buffer unit 6, and the end of the noise buffer unit 6 is fixedly installed with an end buffer sleeve 7; the facial recognition module 3 includes a noise Extraction unit 31, facial expression accurate identification unit 32, identity prior unit 33 and control unit 34, facial expression accurate identification unit 32 is suitable for detecting the picture input convolutional neural network, through convolution operation and pooling operation, extract the picture At the same time, compare the extracted feature map information with the expression images stored in the storage to accurately identify the type of employee expression information. The type of expression is: relaxed, thinking or painful; a relaxed expression means that there is no problem with the inspection or The problem is very small, and the expression of thinking represents that there is a problem here and needs to be repaired, while the painful expression represents that the problem here is large and needs to be repaired urgently; check, and use the face check to extract the feature map as the face prior information, add the facial expression to the face check and extract feature map and distribute it to the control unit 34; the limb feature extraction unit 2 is suitable for extracting the limb feature, and with The body features in the repository are compared to identify the maintenance process of the device by the inspector; the control unit 34 is suitable for introducing the real face image and the feature map extracted by the facial expression feature extraction unit in the dark alley into the discriminator In this method, the expression of the real face image is judged at the feature level, and at the same time, the body features of the inspector are extracted under the joint action of the body feature extraction unit 2, and the facial expression and body performance cooperate to judge the intention of the inspector. Control the work of the noise buffer unit 6 and the working unit 9; the noise extraction unit 31 is suitable for extracting the unclear noise information captured by the monitoring probe, and comparing it with the noise information in the noise library to find the cause of the noise.

Through this embodiment, it can be realized that: through the cooperative use of the set facial recognition module 3 and the limb feature extraction unit 2, the identity prior unit 33 first checks the face and physical features of the inspector, and uses the face check to extract The feature map is used as the prior information of the face, and the feature map extracted by the face verification is added to the expression and distributed to the control unit 34, and the real face image extracted by the facial expression feature extraction unit in the dark alleyway and the face verification The extracted feature map is introduced into the discriminator to judge the expression of the real face image at the feature level. At the same time, under the joint action of the body feature extraction unit 2, the body features of the inspector are extracted. The intention of the inspector can be significantly improved the accuracy of person identification.

Example 2

As shown in Figure 1-9, a precise personnel identification device includes a facial recognition module 3 installed on the side of the coal conveyor belt 1, and a limb feature extraction unit is installed on the side of the coal conveyor belt 1 and below the facial recognition module 3 2. A connecting unit 4 is fixedly installed above the coal conveyor belt 1, a drive unit 5 is fixedly installed above the connecting unit 4, a working unit 9 is fixedly installed below the connecting unit 4, and the bottom of the working unit 9 is connected with a transportation volume distribution plate 10. The transport capacity extension plate 11 and the transport capacity equalizing plate 12, the middle part of the working unit 9 is slidingly connected with the noise buffer unit 6, and the end of the noise buffer unit 6 is fixedly installed with the end buffer sleeve 7; the transport capacity distribution plate 10 A distribution tip 13 is fixedly installed at the front end, a connecting rod 18 is fixedly installed on the upper surface of the transportation volume distribution plate 10, upper baffles 15 are fixedly installed on both sides of the bottom end of the connecting rod 18, and a tail portion is fixedly installed at the tail of the transportation volume distribution plate 10. The sides of the arc-shaped plate 16, the traffic volume distribution plate 10, the traffic volume extension plate 11 and the traffic volume equalizing plate 12 are all rotatably connected with side transmission wheels 14, and the upper end of the connecting rod 18 is fixedly equipped with a driven gear 17, and the traffic volume is extended. The front end of the board 11 is fixedly installed with a connecting flexible board 111, and the end of the connecting flexible board 111 away from the transportation volume extension board 11 is fixedly connected with the tail arc-shaped board 16; The transportation volume distribution plate 12 is opposite to the transportation volume distribution plate 10 rotation direction, and the tail end of the transportation volume extension plate 11 is arranged as an arc, and the tail end of the transportation volume distribution plate 12 is fixedly installed with a limit arc plate 121, and the transportation volume is evenly distributed. The upper end of the sub-board 12 is fixedly equipped with a No. 2 gear 98; the number of the working unit 9 is the same as that of the facial recognition module 3, and the output end of the drive unit 5 extends to the inside of the working unit 9, and its end is connected with a drive wheel 97 for transmission. The two sides of driving wheel 97 are respectively meshed with No. 2 gear 98 and driven gear 17 for transmission; facial recognition module 3 includes noise extraction unit 31, facial expression accurate recognition unit 32, identity prior unit 33 and control unit 34, and facial expression is accurate The identification unit 32 is suitable for inputting the detected picture into the convolutional neural network, and extracts the feature map information of the picture through convolution operation and pooling operation, and compares the extracted feature map information with the expression image stored in the storage device to accurately Identify the types of employee facial expressions. The types of facial expressions are: relaxed, thinking, or painful. A relaxed expression means that there is no problem or a small problem in this maintenance point. A thinking expression means that there is a problem here and needs to be repaired, while a painful expression means that there is a problem here. The problem is large and urgently needs to be fixed;

The identity prior unit 33 is suitable for checking the face and physical features of the inspector, and using the face verification and extraction feature map as the face prior information, and adding the facial expression to the face verification and extraction feature map before distributing to the control unit 34;

The body feature extraction unit 2 is adapted to extract body features, and compare with the body features in the repository to identify the maintenance process of the device by the inspector;

The control unit 34 is suitable for introducing the real face image and the face check extraction feature map extracted by the facial expression feature extraction unit in the dark alley into the discriminator, so as to judge the expression of the real face image at the feature level. Under the combined action of the extraction unit 2, the physical characteristics of the inspector are extracted, and the facial expression and body performance cooperate to judge the intention of the inspector, and at the same time control the work of the noise buffer unit 6 and the working unit 9;

The noise extraction unit 31 is adapted to extract noise information that is not clear from the monitoring probe's shooting, and compare it with the noise information in the noise library to find the cause of the noise.

The above technical features make the working unit 9 and the drive unit 5 not work when judging that the employee is expressing a relaxed expression and movement, and when judging the employee's expression and movement of thinking or suffering, the working unit 9 and the driving unit 5 are made to work.

Through this embodiment, it can be realized that the body feature extraction unit can be used in conjunction with the set limb feature extraction unit 2, facial recognition module 3, work unit 9, transportation volume distribution board 10, transportation volume extension board 11 and transportation volume equalization board 12. 2 and the facial recognition module 3 accurately identify the intention of the inspector, and when it is recognized that the inspector has just moved to this place for inspection, the drive unit 5 drives the transportation volume equalizing plate 12 and the transportation volume distribution plate 10 to rotate, The coal on the surface of the coal conveyor belt 1 is moved to the side of the capacity inspection personnel, forming a blank area for transportation in the inspection area of the inspection personnel, and does not affect the subsequent coal transmission, so as to facilitate the inspection personnel to detect the conveyor belt.

Example 3

As shown in Figure 1-9, a precise personnel identification device includes a facial recognition module 3 installed on the side of the coal conveyor belt 1, and a limb feature extraction unit is installed on the side of the coal conveyor belt 1 and below the facial recognition module 3 2. A connecting unit 4 is fixedly installed above the coal conveyor belt 1, a drive unit 5 is fixedly installed above the connecting unit 4, a working unit 9 is fixedly installed below the connecting unit 4, and the bottom of the working unit 9 is connected with a transportation volume distribution plate 10. The transport capacity extension plate 11 and the transport capacity equalizing plate 12, the middle part of the working unit 9 is slidingly connected with the noise buffer unit 6, and the end of the noise buffer unit 6 is fixedly installed with the end buffer sleeve 7; the transport capacity distribution plate 10 A distribution tip 13 is fixedly installed at the front end, a connecting rod 18 is fixedly installed on the upper surface of the transportation volume distribution plate 10, upper baffles 15 are fixedly installed on both sides of the bottom end of the connecting rod 18, and a tail portion is fixedly installed at the tail of the transportation volume distribution plate 10. The sides of the arc-shaped plate 16, the traffic volume distribution plate 10, the traffic volume extension plate 11 and the traffic volume equalizing plate 12 are all rotatably connected with side transmission wheels 14, and the upper end of the connecting rod 18 is fixedly equipped with a driven gear 17, and the traffic volume is extended. The front end of the board 11 is fixedly installed with a connecting flexible board 111, and the end of the connecting flexible board 111 away from the transportation volume extension board 11 is fixedly connected with the tail arc-shaped board 16; The transportation volume distribution plate 12 is opposite to the transportation volume distribution plate 10 rotation direction, and the tail end of the transportation volume extension plate 11 is arranged as an arc, and the tail end of the transportation volume distribution plate 12 is fixedly installed with a limit arc plate 121, and the transportation volume is evenly distributed. The upper end of the sub-board 12 is fixedly equipped with a No. 2 gear 98; the number of the working unit 9 is the same as that of the facial recognition module 3, and the output end of the drive unit 5 extends to the inside of the working unit 9, and its end is connected with a drive wheel 97 for transmission. The two sides of the driving wheel 97 are meshed with the No. 2 gear 98 and the driven gear 17 for transmission; the inner part of the end buffer sleeve 7 is connected with a heat dissipation blade 73, and the end of the end buffer sleeve 7 is fixedly installed with a protective frame 71. The inner surface of the protective frame 71 is fixed with an expansion bladder 72; the upper surface side of the noise buffer unit 6 is provided with two airflow transmission ports 61, and the airflow transmission ports 61 are connected to the interior of the end buffer sleeve 7, and the noise buffer unit The sides of the 6 are provided with a No. 1 rack 62 and a No. 2 rack 63 respectively, and the No. 2 rack 63 is meshed with the driven gear 17 for transmission; the middle part of the working unit 9 is fixedly equipped with a gas transmission module 8, and the two sides of the gas transmission module 8 A transmission unit 92 is provided. When the noise buffer unit 6 moves to the end, the output end of the transmission unit 92 is connected to the airflow transmission port 61. The bottom of the working unit 9 is fixedly equipped with an airflow circulation pipe 91. The end buffer sleeve 7 The output end is connected to the air circulation pipe 91; the inside of the working unit 9 is provided with a connecting cavity 93, and the inside of the connecting cavity 93 is slidably connected with a sliding unit 94. The sliding unit 94 slides inside the connecting cavity 93, and the upper surface of the sliding unit 94 is fixed. An upper sliding plate 95 is installed, and the upper sliding plate 95 is sealed to the transmission unit 92. The bottom of the sliding unit 94 is meshed with an engaging gear 96, and the engaging gear 96 is rotated and installed on the inner edge of the connecting cavity 93; the facial recognition module 3 includes a noise extraction unit 31, The accurate facial expression recognition unit 32, the identity prior unit 33 and the control unit 34, the accurate facial expression recognition unit 32 is suitable for inputting the detected picture into the convolutional neural network, and after convolution operation and pooling operation, the feature map information of the picture is extracted, At the same time, the extracted feature map information is compared with the expression images stored in the storage to accurately identify the type of employee expression information. The type of expression is: relaxed, thinking or painful. A relaxed expression means that there is no problem or a small problem at this maintenance point. The expression of thinking represents that there is a problem here and needs to be repaired, while the expression of pain represents that the problem here is serious and needs to be repaired urgently; the identity prior unit 33 is suitable for checking the face and physical features of the inspector, and Use the face verification and extraction feature map as the prior information of the face, add the facial expression to the face verification and extraction feature map and distribute it to the control unit 34; the body feature extraction unit 2 is suitable for extracting body features, and with the library. Body feature comparison to identify the maintenance process of the inspector on the device;

The control unit 34 is suitable for introducing the real face image and the face check extraction feature map extracted by the facial expression feature extraction unit in the dark alley into the discriminator, so as to judge the expression of the real face image at the feature level. Under the combined action of the extraction unit 2, the physical characteristics of the inspector are extracted, and the facial expression and physical performance cooperate to judge the intention of the inspector, and at the same time control the work of the noise buffer unit 6 and the working unit 9; the noise extraction unit 31 is suitable for Extract the unclear noise information captured by the monitoring probe, and compare it with the noise information in the noise library to find the cause of the noise.

Through this embodiment, it can be realized that the body feature extraction unit 2 and the facial recognition module 3 identify frequent body movements of the inspector, indicating that the inspector is overhauling here, and the overhaul causes shaking noise in this section of the conveyor belt, causing the facial recognition module 3 The identification is not clear, the drive unit 5 drives the noise buffer unit 6 and the end buffer sleeve 7 to move toward the inspector’s inspection position while the drive unit 5 is driving the transport volume equalizing plate 12 and the transport volume distribution plate 10 to rotate. When the unit 6 and the end buffer sleeve 7 are moved to directly below the facial recognition module 3, the transmission unit 92 is connected to the airflow transmission port 61, and the gas expands the expansion bladder 72 to provide cushioning for the facial recognition module 3, and at the same time, the expansion bladder 72 The lower part of the face provides protection for the overhauler, so that the facial recognition module 3 can clearly identify the patroller during the overhaul process, reducing the risk of accidents during the overhaul process.

It should be noted that the present invention is an accurate personnel identification device and its use method. When in use, firstly, coal is transported on the coal conveyor belt 1, and inspection personnel need to check the identity of the coal conveyor belt 1 before inspection. Unit 33 checks the face and physical features of the inspector to distinguish whether the inspector is an ordinary employee, and uses the face check to extract the feature map as the face prior information, and adds the facial expression to the face check extract feature map Afterwards, the face recognition model is constructed and distributed to the discriminator of the control unit 34, and secondly, every time the inspectors arrive at a patrol point, the facial expression feature extraction unit extracts the real face image and the face check extraction feature in the dark alleyway. The figure is introduced into the discriminator to judge the expression of the real face image at the feature level. At the same time, under the joint action of the body feature extraction unit 2, the body features of the inspector are extracted, and the facial expression and body performance cooperate to judge the inspector. When it is recognized that the inspector has just moved here to inspect this place, the drive unit 5 drives the transportation volume equalizing plate 12 and the transportation volume distribution plate 10 to rotate, and moves the coal on the surface of the coal conveyor belt 1 to the capacity inspection. On one side of the personnel, a blank area for transportation is formed in the inspection area of the inspector, and it does not affect the coal transmission behind, so as to facilitate the inspection of the conveyor belt by the inspector. If noise occurs during the inspection process, the facial recognition module 3 Identify the unclear situation, extract the unclear noise information captured by the monitoring probe, compare it with the noise information in the noise library to find the cause of the noise, and at the same time, the body feature extraction unit 2 and the facial recognition module 3 identify the body movements of the inspector Frequently, it means that the inspectors are overhauling here, and the noise is caused by the overhaul process. The drive unit 5 drives the noise buffer unit 6 and the terminal while driving the transportation volume equalizing plate 12 and the transportation volume distribution plate 10 to rotate. The upper buffer sleeve 7 moves toward the inspection position of the inspector. When the noise buffer unit 6 and the end buffer sleeve 7 move to directly below the face recognition module 3, the transmission unit 92 is connected to the airflow transmission port 61, and the gas makes the inflatable bag The expansion of the body 72 provides a buffer for the face recognition module 3, and at the same time, the underside of the expansion bladder 72 provides protection for the overhauler, so that the face recognition module 3 can clearly identify the inspector during the overhaul process, reducing the risk of accidents during the overhaul process. risk.

The basic principles and main features of the present invention and the advantages of the present invention have been shown and described above. Those skilled in the industry should understand that the present invention is not limited by the above-mentioned embodiments. What are described in the above-mentioned embodiments and the description only illustrate the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will also have Variations and improvements are possible, which fall within the scope of the claimed invention. The protection scope of the present invention is defined by the appended claims and their equivalents.

Claims (7)

1. An accurate personnel identification device is characterized in that it includes a facial recognition module installed on the side of the coal conveyor belt, a body feature extraction unit is installed on the side of the coal conveyor belt and below the facial recognition module, and the coal conveyor belt A connecting unit is fixedly installed above the connecting unit, a drive unit is fixedly installed above the connecting unit, a working unit is fixedly installed below the connecting unit, and the bottom of the working unit is connected with a transportation volume distribution board, a transportation volume extension board and Transport volume distribution plate, the middle part of the working unit is slidably connected with a noise buffer unit, and the end of the noise buffer unit is fixedly installed with an end buffer sleeve; the front end of the transport volume distribution plate is fixedly equipped with a distribution tip, so The upper surface of the transportation volume distribution plate is fixedly installed with connecting rods, and the upper baffles are fixedly installed on both sides of the bottom end of the said transportation volume distribution plate. The distribution plate, the transport capacity extension plate and the transport capacity equalizing plate are all rotatably connected with side transmission wheels, the upper end of the connecting rod is fixedly equipped with a driven gear, and the front end of the transport capacity extension plate is fixedly installed with a connecting flexible plate , the end of the connecting flexible plate away from the transportation volume extension plate is fixedly connected with the tail arc plate; the structure of the transportation volume distribution plate is the same as that of the transportation volume distribution board, and the transportation volume distribution plate The direction of rotation is opposite, the tail end of the transportation volume extension plate is set in an arc shape, the tail end of the transportation volume equalizing plate is fixedly installed with a limited arc-shaped plate, and the upper end of the transportation volume equalizing plate is fixedly installed with a No. 2 Gear; the facial recognition module includes a noise extraction unit, an accurate facial expression recognition unit, an identity prior unit and a control unit, and the accurate facial expression recognition unit is suitable for detecting a picture input convolutional neural network, after convolution operation and pooling The feature map information of the picture is extracted, and the extracted feature map information is compared with the expression image stored in the storage device to accurately identify the type of employee expression information; the identity a priori unit is suitable for inspectors The face and body features are checked, and the feature map is used as the prior information of the face, and the feature map is added to the facial expression and distributed to the control unit; the limb feature extraction unit is suitable for extracting Body features, and compared with the body features in the repository to identify the maintenance process of the device by the inspector; the control unit is suitable for the real face image extracted by the facial expression feature extraction unit in the dark alleyway and the face verification The extracted feature map is introduced into the discriminator to judge the expression of the real face image at the feature level. At the same time, under the joint action of the body feature extraction unit, the body features of the inspector are extracted, and the facial expression and body performance cooperate to judge the inspection The intention of the operator, while controlling the work of the noise buffer unit and the working unit; the noise extraction unit is suitable for extracting the unclear noise information captured by the monitoring probe, and comparing it with the noise information in the noise library to find the cause of the noise.
2. A precise personal identification device according to claim 1, wherein the number of the working unit is the same as that of the facial recognition module, the output end of the driving unit extends to the inside of the working unit, and its end drives A driving wheel is connected, and the two sides of the driving wheel are respectively meshed with the second gear and the driven gear for transmission.
3. A precise personnel identification device according to claim 1, characterized in that, the inside of the end buffer sleeve is rotatably connected with cooling fins, and the end of the end buffer sleeve is fixedly installed with a protective frame, so An expansion bladder is fixedly installed on the inner surface of the protective frame.
4. A precise personnel identification device according to claim 3, characterized in that two airflow transmission ports are opened on the side of the upper surface of the noise buffer unit, and the airflow transmission ports communicate with the interior of the end buffer sleeve , The side surfaces of the noise buffering unit are respectively provided with a No. 1 rack and a No. 2 rack, and the No. 2 rack is meshed with the driven gear for transmission.
5. The precise personnel identification device according to claim 1, wherein a gas transmission module is fixedly installed in the middle of the working unit, and transmission units are arranged on both sides of the gas transmission module, when the noise buffer When the unit moves to the end, the output end of the transmission unit is connected to the airflow transmission port, the bottom of the working unit is fixedly equipped with an airflow circulation pipe, and the output end of the end buffer sleeve is connected to the airflow circulation pipe.
6. A precise person identification device according to claim 1, wherein a connection cavity is opened inside the working unit, and a sliding unit is slidably connected to the inside of the connection cavity, and the sliding unit is in the connection cavity. Sliding inside, the upper surface of the sliding unit is fixedly equipped with an upper sliding plate, the upper sliding plate is sealed to the transmission unit, the bottom of the sliding unit is meshed with a meshing gear, and the meshing gear is rotatably installed on the inner edge of the connecting cavity.
7. A method for using an accurate personal identification device according to any one of claims 1-6, characterized in that the steps of use are as follows:

   A: Coal is transported on the coal conveyor belt. Before the inspectors inspect the coal conveyor belt, the identity prior unit checks the face and physical features of the inspector, and uses the face check to extract the feature map as the identity map. Face prior information, the feature map extracted by face nuclear inspection is added with expression, and then the face recognition model is constructed and distributed to the discriminator of the control unit;

   B: After completing step A, each time the inspectors arrive at an inspection point, the real face image and the face verification extraction feature map extracted by the facial expression feature extraction unit in the dark alleyway are introduced into the discriminator to judge at the feature level The expression of the real face image, and at the same time, under the joint action of the body feature extraction unit, the body features of the inspector are extracted, and the facial expression and body performance cooperate to judge the intention of the inspector. When the position is specified, the driving unit drives the transportation volume equalizing plate and the transportation volume distribution plate to rotate, and moves the coal on the surface of the coal conveyor belt to the side of the capacity inspection personnel, forming a blank area for transportation in the inspection area of the inspection personnel;

   C: After completing step B, if noise is generated during the inspection process and the facial recognition module is not clearly recognized, extract the unclear noise information captured by the monitoring probe and compare it with the noise information in the noise library to find the source of the noise The reason is that at the same time, the body feature extraction unit and the facial recognition module recognize the frequent body movements of the inspector, which means that the inspector is here for maintenance, and the noise is caused by the maintenance process;

   D: At the same time as step C is completed, the drive unit drives the noise buffer unit and the end buffer sleeve to move toward the inspector’s inspection position while driving the transportation volume equalizer plate and the transportation volume distribution plate to rotate. When the noise buffer unit When the buffer sleeve at the end moves to directly below the facial recognition module, the transmission unit communicates with the airflow transmission port, and the gas expands the expansion capsule to provide buffer for the facial recognition module, while the underside of the expansion capsule provides protection for the maintenance personnel.

Images