Signboard classification device based on image recognition

Abstract

The utility model relates to the technical field of recovery and classification of electric power signboards, and provides a signboard classification device based on image recognition, which is characterized in that a wide-angle camera outputs collected signboard images to an image processing unit, and the image processing unit processes the signboard images and then outputs the signboard images to a classification and recognition unit; the classification and recognition unit recognizes the types of the signboards and outputs signboard type information to the mechanical arm control unit, and the mechanical arm control unit obtains signboard placement coordinates according to the signboard type information and plans a mechanical arm movement path according to the signboard placement coordinate information; the signboards are placed in the classification barrels corresponding to the signboards; in the signboard placing process, the mechanical arm control unit communicates with the air pump control unit, the air pump control unit controls the electric suction cup to suck and loosen the signboards and is matched with the classification mechanical arm to transfer the signboards from the initial positions into the corresponding classification barrels, manual work is replaced, signboard classification and counting work is completed, and loads are relieved.

Description

TECHNICAL FIELD

[0001] The utility model relates to power signboard recycling classification technical field especially, a signboard classification device based on image recognition. BACKGROUND

[0002] Power signboards are tools used for prompting and warning, and are widely used in power facilities and equipment to ensure safe operation. They are usually placed in conspicuous positions such as near power equipment, substations, distribution boxes, and power poles, to ensure that personnel can clearly see the signboard content. The design and use of power signboards must comply with relevant regulations and standards, such as ISO, NSI, or regulations of the national power department, to ensure their effectiveness and consistency. These signboards can effectively reduce accidents and improve operational safety.

[0003] In actual use of power signboards, the existing signboards in substations vary in size, listed text, and images. When the signboards are used and then recycled, manual identification and classification are required for proper placement. Monthly maintenance requires manual inventory and inspection of correct placement, which is low in work efficiency and time-consuming and labor-intensive.

[0004] Therefore, the utility model provides a signboard classification device based on image recognition to replace manual classification and inventory of signboards, avoid repetitive work by operating personnel, and reduce work load. UTILITY MODEL CONTENT

[0005] The utility model aims at solving the problems of the prior art, providing a signboard classification device based on image recognition to replace manual classification and inventory of signboards, avoid repetitive work by operating personnel, and reduce work load.

[0006] To achieve the above-mentioned purpose, the utility model adopts the following technical solutions:

[0007] A signboard classification device based on image recognition includes a conveying device, a classification mechanical arm on one side of the conveying device, a plurality of classification barrels within the working range of the classification mechanical arm, and a control device on the base of the classification mechanical arm. The end of the classification mechanical arm is provided with an electric suction cup and a wide-angle camera. An air pump is also provided on the base of the classification mechanical arm. The control device is electrically connected with the classification mechanical arm, the wide-angle camera, the air pump, and the conveying device. The air inlet of the electric suction cup is communicated with the air outlet of the air pump through an air pipe.

[0008] The control device comprises an image processing unit, a classification identification unit, a mechanical arm control unit and a gas pump control unit; the image processing unit is electrically connected with the wide-angle camera output, the image processing unit output is electrically connected with the classification identification unit input, the classification identification unit output is electrically connected with the mechanical arm control unit input, and the mechanical arm control unit output and input are electrically connected with the gas pump control unit input and output.

[0009] Preferably, the classification mechanical arm comprises a base, a rudder seat arranged on the top of the base, a rudder platform rotatably connected with the rudder seat, a large arm having one end rotatably connected with the rudder platform, a small arm having one end rotatably connected with the other end of the large arm, a horizontal seat rotatably connected with the other end of the small arm, a second motor and a third motor arranged on the two sides of the rudder platform, a first motor arranged on the bottom of the rudder platform, and a crank rotatably connected with the rudder platform; the output shaft of the second motor is connected with the large arm, the output shaft of the third motor is connected with the crank, and the output shaft of the first motor is connected with the bottom of the rudder platform; the small arm is provided with a rear fulcrum at the position where the small arm is connected with the large arm; a third connecting rod is rotatably connected with the rear fulcrum, and the idle end of the third connecting rod is rotatably connected with the crank; the top end of the large arm is rotatably connected with a tripod, the bottom of the tripod is rotatably connected with the outside of the large arm, a first connecting rod is rotatably connected with the right end of the tripod, a second connecting rod is rotatably connected with the left end of the tripod, the idle end of the first connecting rod is rotatably connected with the rudder seat, and the idle end of the second connecting rod is rotatably connected with the horizontal seat; the horizontal seat is connected with a suction cup and a wide-angle camera; and the first motor, the second motor and the third motor are electrically connected with the control device.

[0010] Preferably, the electric suction cup comprises a connecting seat connected with the horizontal seat, a profile arranged on the bottom of the connecting seat, three groups of air nozzle fixing seats sleeved with the profile, and a four-way valve arranged on one side of the profile; each group of the three groups of air nozzle fixing seats is provided with at least two air nozzles, and the two air nozzles are arranged on the two ends of the air nozzle fixing seat; the four-way valve is in communication with a main gas pipe and three first branch gas pipes, one first branch gas pipe is in communication with two second branch gas pipes through a three-way valve, and the idle end of the second branch gas pipe is in communication with the air inlet end of the air nozzle.

[0011] Preferably, the profile is provided with a sliding groove on the left side, the right side and the bottom, and a sliding block is slidably connected in the sliding groove; the top of the air nozzle fixing seat is provided with a U-shaped groove, the U-shaped groove is sleeved with the profile, and the air nozzle fixing seat and the sliding block are connected through bolts.

[0012] Preferably, a plurality of air nozzle mounting holes are arranged on the air nozzle fixing seat.

[0013] Preferably, the control device adopts Jetson Nno B01.

[0014] The signboard classification device based on image recognition has the following advantages.

[0015] The utility model discloses a wide-angle camera is with the signboard image that gathers output is given image processing unit, and image processing unit is given classification identification unit after signboard image processing output, and image processing unit extracts the initial coordinate information of the signboard that is photographed in addition and transfers to classification identification unit, and classification identification unit identifies the kind of signboard, and signboard kind information and the initial coordinate information of signboard are output to mechanical arm control unit, and mechanical arm control unit obtains signboard placement coordinates according to signboard kind information, and mechanical arm control unit is according to signboard's initial coordinate information and signboard placement coordinate information and plans mechanical arm movement path, and places signboard in the classification bucket corresponding thereto, in the process of placing signboard, and mechanical arm control unit and air pump control unit communicate, when classification mechanical arm end control electric sucking disc moves to the signboard operation position that needs to suck, and air pump control unit controls electric sucking disc and sucks signboard, when classification mechanical arm end control electric sucking disc moves to the signboard corresponding classification bucket operation position, and air pump control unit controls electric sucking disc and releases signboard, and places signboard in the corresponding classification bucket, and replaces artificial and completes the classification, and the counting work of signboard, avoids operating personnel and develops repetitive work, and reduces the work load, and greatly releases human resources. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 It is the overhead view of the signboard classification device based on image recognition of the utility model.

[0017] Figure 2 It is the structure schematic view of the embodiment 2 of the utility model.

[0018] Figure 3 It is still another structure schematic view of the embodiment 2 of the utility model.

[0019] Figure 4 It is the structure schematic view of the embodiment 3 of the utility model.

[0020] Figure 5 It is Figure 4 The partial enlarged schematic view of the middle part.

[0021] Figure 6 It is the connection schematic view of the section bar and air cock fixed seat of the utility model.

[0022] Figure 7 It is the principle block diagram of the control device of the utility model.

[0023] In the accompanying drawings: 1. Conveying device; 2. Sorting robot arm; 201. Base; 202. Steering seat; 203. Steering platform; 204. Upper arm; 205. Lower arm; 206. Horizontal seat; 207. Tripod; 208. First connecting rod; 209. Second connecting rod; 210. Crank; 211. Rear fulcrum; 212. Third connecting rod; 213. First motor; 214. Second motor; 215. Third motor; 3. Sorting bucket; 4. Wide-angle camera; 5. Control device; 51. Image processing unit; 52. Classification and recognition unit; 53. Robotic arm control unit; 54. Air pump control unit; 6. Electric suction cup; 601. Connecting seat; 602. Profile; 603. Air nozzle fixing seat; 604. Air nozzle; 605. Main air pipe; 606. Four-way valve; 607. First air branch pipe; 608. Three-way valve; 609. Second air branch pipe; 610. Slider; 611. Bolt; 7. Air pump; 8. Air pipe. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0025] In the description of the present invention, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside", "outside", etc., indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they cannot be understood as limitations on the present invention. Example 1

[0026] Please refer to Figure 1 A signboard sorting device based on image recognition comprises: a conveying device 1, a sorting robot 2 mounted on one side of the conveying device 1, a plurality of sorting buckets 3 within the working range of the sorting robot 2, and a control device 5 mounted on the base 201 of the sorting robot 2. The sorting robot 2 is provided with an electric suction cup 6 and a wide-angle camera 4 at its end. An air pump 7 is also mounted on the base 201 of the sorting robot 2. The control device 5 is electrically connected to the sorting robot 2, the wide-angle camera 4, the air pump 7, and the conveying device 1. The air inlet of the electric suction cup 6 is connected to the air outlet of the air pump 7 via an air pipe 8. The sorting buckets 3 remain in place after placement. In this embodiment, the wide-angle camera 4 uses a JD-202 USB module.

[0027] like Figure 7As shown, the control device 5 comprises: an image processing unit 51, a classification recognition unit 52, a mechanical arm control unit 53, and a gas pump control unit 54; the image processing unit 51 is electrically connected with the output of the wide-angle camera 4, the output of the image processing unit 51 is electrically connected with the input of the classification recognition unit 52, the output of the classification recognition unit 52 is electrically connected with the input of the mechanical arm control unit 53, and the output and input of the mechanical arm control unit 53 are electrically connected with the input and output of the gas pump control unit 54; the wide-angle camera 4 outputs the collected signboard image to the image processing unit 51, the image processing unit 51 outputs the normalized, size-adjusted, and grayscale-processed signboard image to the classification recognition unit 52, the image processing unit 51 additionally extracts the initial coordinate information of the photographed signboard and transmits it to the classification recognition unit 52, the classification recognition unit 52 identifies the type of the signboard and outputs the signboard type information and the initial coordinate information of the signboard to the mechanical arm control unit 53, the mechanical arm control unit 53 obtains the signboard placement coordinates according to the signboard type information, and the mechanical arm control unit 53 plans the motion path of the classification mechanical arm 2 according to the initial coordinate information of the signboard and the signboard placement coordinate information to place the signboard into the corresponding classification barrel 3; during the placement of the signboard, the mechanical arm control unit 53 communicates with the gas pump control unit 54, when the end control electric suction cup 6 of the classification mechanical arm 2 moves to the signboard operation position that needs to be sucked, the gas pump control unit 54 controls the electric suction cup 6 to suck the signboard, when the end control electric suction cup 6 of the classification mechanical arm 2 moves to the operation position of the signboard corresponding classification barrel 3, the gas pump control unit 54 controls the electric suction cup 6 to release the signboard, and places the signboard into the corresponding classification barrel 3. It should be understood that the conveying device 1 remains stationary during the classification of the signboard. As an example, the classification recognition unit 52 classifies the signboard according to a convolutional neural network model, and the convolutional neural network model is trained using common signboards. The path planning algorithm of the classification mechanical arm 2 uses the RRT algorithm. In this embodiment, there are seven common signboards, which are “work here”, “enter and exit from here”, “go up and down from here”, “no closing line with people working”, “no closing with people working”, “cross operation with isolation point for primary equipment”, and “cross operation with isolation point for secondary equipment”.

[0028] As a preferred control device 5, Jetson Nno B01 is adopted to reduce the cost while ensuring the recognition accuracy. Embodiment 2

[0029] Based on Embodiment 1, please refer to Figure 2 and Figure 3The embodiment provides a preferred classification mechanical arm 2 comprising a base 201, a rudder seat 202 arranged on the top of the base 201, a rudder platform 203 rotationally connected with the rudder seat 202, a large arm 204 rotationally connected with one end of the rudder platform 203, a small arm 205 rotationally connected with the other end of the large arm 204, a horizontal seat 206 rotationally connected with the other end of the small arm 205, a second motor 214 and a third motor 215 arranged on the two sides of the rudder platform 203, a first motor 213 arranged on the bottom of the rudder platform 203, a crank 210 rotationally connected with the rudder platform 203; the output shaft of the second motor 214 is connected with the large arm 204, the output shaft of the third motor 215 is connected with the crank 210, and the output shaft of the first motor 213 is connected with the bottom of the rudder platform 203; the small arm 205 is provided with a rear supporting point 211 at the position where the small arm 205 is connected with the large arm 204; the rear supporting point 211 is rotationally connected with a third connecting rod 212, and the idle end of the third connecting rod 212 is rotationally connected with the crank 210; the top end of the large arm 204 is rotationally connected with a tripod 207, the bottom of the tripod 207 is rotationally connected with the outside of the large arm 204, the right end of the tripod 207 is rotationally connected with a first connecting rod 208, the left end of the tripod 207 is rotationally connected with a second connecting rod 209, the idle end of the first connecting rod 208 is rotationally connected with the rudder seat 202, and the idle end of the second connecting rod 209 is rotationally connected with the horizontal seat 206; the horizontal seat 206 is connected with a suction cup and a wide-angle camera 4; the first motor 213, the second motor 214 and the third motor 215 are electrically connected with a control device 5. The first motor 213 controls the rotation of the rudder platform 203 on the horizontal plane, the second motor 214 controls the rotation angle between the large arm 204 and the rudder platform 203, the third motor 215 controls the included angle between the small arm 205 and the large arm 204, the motion path of the electric suction cup 6 can be controlled by controlling the rotation angles of the first motor 213, the second motor 214 and the third motor 215, the second motor 214 and the third motor 215 are arranged on the rudder platform 203, the load of the classification mechanical arm 2 itself can be reduced, the motion precision of the end of the classification mechanical arm 2 is improved, and as an option, the first motor 213, the second motor 214 and the third motor 215 use HTS-35H rudders in the embodiment. Embodiment 3

[0030] Based on embodiment 1, please refer to Figures 4 to 6In the embodiment, in order to improve the suction effect of the electric suction cup 6, the number of suction nozzles is increased, and the electric suction cup 6 comprises a connecting seat 601 connected with the horizontal seat 206, a profile 602 arranged at the bottom of the connecting seat 601, three groups of suction nozzle fixing seats 603 sleeved with the profile 602, and a four-way pipe 606 arranged at one side of the profile 602. Each group of the three groups of suction nozzle fixing seats 603 is provided with at least two suction nozzles, and the two suction nozzles are arranged at two ends of the suction nozzle fixing seat 603. The four-way pipe 606 is respectively communicated with a main air pipe 605 and three first branch air pipes 607, one first branch air pipe 607 is respectively communicated with two second branch air pipes 609 through a three-way pipe 608, and the second branch air pipe 609 is communicated with the air inlet end of the suction nozzle 604. In the embodiment, the number of the suction nozzles 604 is six, and the suction nozzles 604 are uniformly distributed, so that the effect of sucking the signboard can be improved.

[0031] As preferred, the left side, the right side and the bottom of the profile 602 are provided with sliding grooves, and sliding blocks 610 are slidingly connected in the sliding grooves. The sliding grooves and the sliding blocks 610 are trapezoidal, the top of the suction nozzle fixing seat 603 is provided with a U-shaped groove, the U-shaped groove is sleeved with the profile 602, the suction nozzle fixing seat 603 is connected with the sliding block 610 through a bolt 611, and the sliding and locking of the suction nozzle fixing seat 603 and the profile 602 can be controlled by controlling the tightness of the bolt 611, so that the transverse distance between the suction nozzles 604 can be quickly adjusted to adapt to signboards of different sizes. As preferred, a plurality of suction nozzle 604 mounting holes are arranged on the suction nozzle fixing seat 603, the number of the suction nozzles 604 can be increased or the longitudinal distance between the suction nozzles 604 can be adjusted, and the air path needs to be adjusted when the number of the suction nozzles 604 is adjusted.

[0032] The above is only a preferred specific embodiment of the utility model, but the protection scope of the utility model is not limited to this. The replacement can be the replacement of part of the structure, device, method step, or the complete technical scheme. According to the technical scheme and the utility model concept of the utility model, equivalent replacement or change should be covered in the protection scope of the utility model.

Claims

1. A sign classification device based on image recognition, characterized in that: include: A conveying device (1), a classification robot arm (2) provided on one side of the conveying device (1), a plurality of classification barrels (3) provided within the working range of the classification robot arm (2), and a control device (5) provided on the base (201) of the classification robot arm (2); an electric suction cup (6) and a wide-angle camera (4) are provided at the end of the classification robot arm (2); an air pump (7) is also provided on the base (201) of the classification robot arm (2); the control device (5) is electrically connected to the classification robot arm (2), the wide-angle camera (4), the air pump (7), and the conveying device (1) respectively; the air inlet of the electric suction cup (6) and the air outlet of the air pump (7) are connected through an air pipe (8); The control device (5) includes: an image processing unit (51), a classification and recognition unit (52), a robotic arm control unit (53), and an air pump control unit (54); the input of the image processing unit (51) is electrically connected to the output of the wide-angle camera (4), the output of the image processing unit (51) is electrically connected to the input of the classification and recognition unit (52), the output of the classification and recognition unit (52) is electrically connected to the input of the robotic arm control unit (53), and the output and input of the robotic arm control unit (53) are electrically connected to the input and output of the air pump control unit (54), respectively.

2. The image recognition-based sign classification device according to claim 1, characterized in that: The classification robot arm (2) comprises: a base (201), a steering seat (202) arranged on the top of the base (201), a steering platform (203) rotatably connected to the steering seat (202), a large arm (204) rotatably connected to the steering platform (203), a small arm (205) rotatably connected to the large arm (204) at one end, a horizontal seat (206) rotatably connected to the other end of the small arm (205), a second motor (214) and a third motor (215) respectively arranged on both sides of the steering platform (203), a first motor (213) arranged at the bottom of the steering platform (203), and a crank (210) rotatably connected to the steering platform (203); an output shaft of the second motor (214) is connected to the large arm (204), an output shaft of the third motor (215) is connected to the crank (210), and an output shaft of the first motor (213) is connected to the bottom of the steering platform (203); the small arm (205) is connected to the small arm (214) and the crank (210) at the bottom of the small arm (215). 05) is provided with a rear fulcrum (211) at the connection with the big arm (204); a third connecting rod (212) is rotatably connected to the rear fulcrum (211), and the idle end of the third connecting rod (212) is rotatably connected to the crank (210); the top of the big arm (204) is rotatably connected to the tripod (207), the bottom of the tripod (207) is rotatably connected to the outside of the big arm (204), the right end of the tripod (207) is rotatably connected to the first connecting rod (208), and the left end of the tripod (207) is rotatably connected to the second connecting rod (209), the idle end of the first connecting rod (208) is rotatably connected to the steering seat (202), and the idle end of the second connecting rod (209) is rotatably connected to the horizontal seat (206); the horizontal seat (206) is connected to the suction cup and the wide-angle camera (4); the first motor (213), the second motor (214) and the third motor (215) are electrically connected to the control device (5) respectively.

3. The sign classification device based on image recognition according to claim 1, characterized in that: The electric suction cup (6) comprises: a connecting seat (601) connected to the horizontal seat (206), a profile (602) provided at the bottom of the connecting seat (601), three groups of air nozzle fixing seats (603) sleeved with the profile (602), and a four-way (606) provided on one side of the profile (602); each of the three groups of air nozzle fixing seats (603) is provided with at least two air nozzles, and the two air nozzles are respectively provided at both ends of the air nozzle fixing seat (603); the four-way (606) is respectively connected to the main air pipe (605) and three first air branch pipes (607); one first air branch pipe (607) is respectively connected to two second air branch pipes (609) through a three-way (608); and the idle end of the second air branch pipe (609) is connected to the air inlet end of the air nozzle (604).

4. The image recognition-based sign classification device according to claim 3, characterized in that: The left side, right side and bottom of the profile (602) are provided with sliding grooves, and sliders (610) are slidably connected in the sliding grooves. The top of the air nozzle fixing seat (603) is provided with a U-shaped groove, and the U-shaped groove is sleeved with the profile (602). The air nozzle fixing seat (603) and the slider (610) are connected by bolts (611).

5. The image recognition-based sign classification device according to claim 4, characterized in that: The gas nozzle fixing seat (603) is provided with a plurality of gas nozzle mounting holes.

6. The sign classification device based on image recognition according to claim 1, characterized in that: The control device (5) adopts Jetson Nno B01.

Images