CN216430965U

CN216430965U - But remote control's convolution neural network model trainer

Info

Publication number: CN216430965U
Application number: CN202122672650.9U
Authority: CN
Inventors: 廖景行; 冯蕾; 赵冉冉
Original assignee: China National Institute of Standardization
Current assignee: China National Institute of Standardization
Priority date: 2021-11-03
Filing date: 2021-11-03
Publication date: 2022-05-03
Anticipated expiration: 2031-11-03

Abstract

The utility model discloses a but remote control's convolution neural network model trainer, including base and trainer body, the first motor of top fixedly connected with of base, the output fixedly connected with mounting panel of first motor, the one end fixedly connected with second motor of mounting panel, the output fixedly connected with screw rod of second motor, the top of mounting panel is equipped with the spout, the outer wall that the mounting panel was run through to the one end of screw rod extends in the spout, the slider has been cup jointed to the one end screw thread that the screw rod is located the spout, the slider is connected through the connecting rod with one side of trainer body, the both ends of connecting rod rotate with slider and trainer body respectively and are connected. The utility model discloses in through remote control signal reception and sending device's setting, can carry out long-range control to the device to training device's application scope has been promoted.

Description

But remote control's convolution neural network model trainer

Technical Field

The utility model relates to a convolutional neural network model training technical field especially relates to a but remote control's convolutional neural network model trainer.

Background

With the continuous updating and development of computer vision technology, the target detection technology has an important position in a plurality of fields of intelligent transportation, image retrieval and face recognition. In recent years, the development of deep learning of more and more fire as a more efficient tool helps us to research and discover in the field of target detection. At present, the deep learning greatly surpasses the traditional visual algorithm in the field of target detection, the deep learning can obtain effective characteristics under big data, and the learned characteristics far surpass the algorithm characteristics designed by hand in quantity and performance.

However, the external structure of the conventional convolutional neural network model training device is simple, and the training device cannot be remotely controlled and adjusted, so that the application range of the training device is influenced.

SUMMERY OF THE UTILITY MODEL

The utility model aims at solving the problem that the convolutional neural network model training device external structure is comparatively simple among the prior art, and the training device can't carry out the remote control and adjust, but the convolutional neural network model training device of a remote control that proposes.

In order to realize the purpose, the utility model adopts the following technical scheme:

a convolutional neural network model training device capable of being remotely controlled comprises a base and a training device body, wherein the top of the base is fixedly connected with a first motor, the output end of the first motor is fixedly connected with a mounting plate, one end of the mounting plate is fixedly connected with a second motor, the output end of the second motor is fixedly connected with a screw rod, the top of the mounting plate is provided with a chute, one end of the screw rod penetrates through the outer wall of the mounting plate and extends into the chute, one end of the screw rod, which is positioned in the chute, is in threaded sleeve connection with a sliding block, the sliding block is connected with one side of the training device body through a connecting rod, the two ends of the connecting rod are respectively in rotating connection with the sliding block and the training device body, the top of the mounting plate is fixedly connected with a supporting block, the top of the supporting block is in rotating connection with the bottom of the training device body through a rotating rod, the top of the training device body is provided with a cavity, the cavity is internally provided with a controller, a memory, a power supply, a signal receiver and a signal transmitter, the output ends of the signal receiver and the power supply are connected with the input end of the controller, and the output end of the controller is respectively connected with the input ends of the memory, the first motor, the second motor, the signal transmitter and the training device body.

Preferably, the bottom of the base is rotatably connected with a plurality of universal self-locking wheels.

Preferably, the bottom fixedly connected with of mounting panel a plurality of slide bars that the symmetry set up, the bottom fixedly connected with guide block of slide bar, the top of base is equipped with the annular slide rail that corresponds with the guide block.

Preferably, the rotating rod is sleeved with a torsion spring, and two ends of the torsion spring are respectively and fixedly connected with the supporting block and the training device body.

Preferably, a guide rod is fixedly connected in the sliding groove, and a through hole corresponding to the guide rod is formed in the sliding block.

Preferably, the one end of cavity is equipped with the installing port, rotate through the pivot in the installing port and be connected with the baffle, the one end that the pivot was kept away from to the baffle passes through locking screw fixed connection with the installing port inner wall, be equipped with the thermovent on the baffle, the thermovent internal fixation is inserted and is equipped with the dust screen.

Preferably, a power supply connector is fixedly inserted into the outer wall of the training device body, and the output end of the power supply connector is connected with the input end of a power supply.

Has the advantages that:

1. through the setting of the signal receiver, a remote control instruction can be received, meanwhile, the setting of the signal transmitter can transmit a remote signal, the controller can control the first motor and the second motor, the first motor drives the mounting plate to rotate, the direction of the training device body is adjusted, the second motor can drive the screw rod to rotate, the training device body is driven to rotate through the slider and the connecting rod, and the direction of the training device body is adjusted;

2. the setting of universal auto-lock wheel can conveniently remove the base, and the setting of slide bar and guide block can prevent rocking of mounting panel simultaneously, and the guide arm can prevent rocking of slider simultaneously, and the setting of baffle can protect control element simultaneously, and the setting of dust screen can play certain dustproof effect.

Drawings

Fig. 1 is a schematic structural diagram of a convolutional neural network model training device capable of being remotely controlled according to the present invention;

fig. 2 is a schematic structural view of a portion a of fig. 1.

In the figure: the training device comprises a base 1, a training device body 2, a first motor 3, a mounting plate 4, a second motor 5, a screw 6, a slider 7, a connecting rod 8, a supporting block 9, a rotating rod 10, a controller 11, a memory 12, a power supply 13, a signal receiver 14, a signal transmitter 15, a universal self-locking wheel 16, a sliding rod 17, a guide block 18, a guide rod 19, a rotating shaft 20, a baffle 21, a locking screw 22, a dust screen 23 and a power supply connector 24.

Detailed Description

The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only some embodiments of the present invention, not all embodiments.

Referring to fig. 1-2, a convolutional neural network model training device capable of being remotely controlled comprises a base 1 and a training device body 2, wherein the bottom of the base 1 is rotatably connected with a plurality of universal self-locking wheels 16, the base 1 is convenient to move, the top of the base 1 is fixedly connected with a first motor 3 for driving a first mounting plate 4 to rotate, the output end of the first motor 3 is fixedly connected with the mounting plate 4, the bottom of the mounting plate 4 is fixedly connected with a plurality of symmetrically arranged slide bars 17 for supporting the mounting plate 4, the bottoms of the slide bars 17 are fixedly connected with guide blocks 18, the top of the base 1 is provided with annular slide rails corresponding to the guide blocks 18, and the slide bars 17 are prevented from falling off from the annular slide rails;

in the embodiment, one end of the mounting plate 4 is fixedly connected with a second motor 5 for driving the screw 6 to rotate, the output end of the second motor 5 is fixedly connected with the screw 6 for driving the slider 7 to move left and right, the top of the mounting plate 4 is provided with a chute, a guide rod 19 is fixedly connected in the chute, the slider 7 is provided with a through hole corresponding to the guide rod 19, one end of the screw 6 penetrates through the outer wall of the mounting plate 4 and extends into the chute, one end of the screw 6 in the chute is in threaded sleeve connection with the slider 7, the slider 7 is connected with one side of the training device body 2 through a connecting rod 8 for driving the training device body 2 to rotate, and two ends of the connecting rod 8 are respectively in rotating connection with the slider 7 and the training device body 2;

in this embodiment, the top of the mounting plate 4 is fixedly connected with a supporting block 9 for supporting a rotating rod 10, the rotating rod 10 is sleeved with a torsion spring, two ends of the torsion spring are respectively and fixedly connected with the supporting block 9 and the training device body 2 to support the training device body 2 with a certain torsion, the top of the supporting block 9 is rotatably connected with the bottom of the training device body 2 through the rotating rod 10, the top of the training device body 2 is provided with a cavity, one end of the cavity is provided with a mounting opening, and a baffle 21 is rotatably connected in the mounting opening through a rotating shaft 20 to protect a control element;

in this embodiment, one end of the baffle 21, which is far away from the rotating shaft 20, is fixedly connected with the inner wall of the mounting hole through a locking screw 22, so that the baffle 21 is conveniently fixed, the baffle 21 is provided with a heat dissipation hole, a dust screen 23 is fixedly inserted in the heat dissipation hole, so that dust is prevented from entering a cavity, and the controller 11, the memory 12, the power supply 13, the signal receiver 14 and the signal transmitter 15 are arranged in the cavity;

in this embodiment, the output ends of the signal receiver 14 and the power supply 13 are connected to the input end of the controller 11, and are used for sending a control signal to the controller 11, the output end of the controller 11 is connected to the input ends of the memory 12, the first motor 3, the second motor 5, the signal transmitter 15 and the training device body 2, and is used for controlling the first motor 3 and the second motor 5, the outer wall of the training device body 2 is fixedly inserted with the power supply connector 24, and the output end of the power supply connector 24 is connected to the input end of the power supply 13.

In this embodiment, through the setting of signal receiver 14, can receive the remote control instruction, signal transmitter 15's setting simultaneously, remote signal's transmission can be carried out, controller 11 can control first motor 3 and second motor 5, first motor 3 drives mounting panel 4 and rotates, adjusts the direction of trainer body 2, second motor 5 can drive screw rod 6 and rotate, drive trainer body 2 through slider 7 and connecting rod and rotate, adjusts the direction of trainer body 2.

In this embodiment, base 1 can conveniently be removed in the setting of universal auto-lock wheel 16, and slide bar 17 and guide block 18's setting can prevent rocking of mounting panel 4 simultaneously, and guide arm 19 can prevent rocking of slider 7 simultaneously, and the setting of baffle 21 can protect control element simultaneously, and certain dustproof effect can be played in the setting of dust screen 23.

The above, only be the concrete implementation of the preferred embodiment of the present invention, but the protection scope of the present invention is not limited thereto, and any person skilled in the art is in the technical scope of the present invention, according to the technical solution of the present invention and the utility model, the concept of which is equivalent to replace or change, should be covered within the protection scope of the present invention.

Claims

1. The utility model provides a but remote control's convolutional neural network model training device, includes base (1) and training device body (2), its characterized in that: the top of the base (1) is fixedly connected with a first motor (3), the output end of the first motor (3) is fixedly connected with a mounting plate (4), one end of the mounting plate (4) is fixedly connected with a second motor (5), the output end of the second motor (5) is fixedly connected with a screw rod (6), the top of the mounting plate (4) is provided with a chute, one end of the screw rod (6) penetrates through the outer wall of the mounting plate (4) and extends inwards into the chute, one end of the screw rod (6) positioned in the chute is sleeved with a slider (7) in a threaded manner, the slider (7) is connected with one side of the training device body (2) through a connecting rod (8), two ends of the connecting rod (8) are respectively rotatably connected with the slider (7) and the training device body (2), the top of the mounting plate (4) is fixedly connected with a supporting block (9), the top of the supporting block (9) is rotatably connected with the bottom of the training device body (2) through a rotating rod (10), the top of trainer body (2) is equipped with the cavity, be equipped with controller (11), memory (12), power (13), signal receiver (14) and signal transmitter (15) in the cavity, the output of signal receiver (14) and power (13) all is connected with the input of controller (11), the output of controller (11) is connected with the input of memory (12), first motor (3), second motor (5), signal transmitter (15) and trainer body (2) respectively.

2. The convolutional neural network model training device capable of being remotely controlled according to claim 1, wherein: the bottom of the base (1) is rotatably connected with a plurality of universal self-locking wheels (16).

3. The convolutional neural network model training device capable of being remotely controlled according to claim 1, wherein: the bottom fixedly connected with of mounting panel (4) symmetry sets up a plurality of slide bars (17), the bottom fixedly connected with guide block (18) of slide bar (17), the top of base (1) is equipped with the annular slide rail that corresponds with guide block (18).

4. The convolutional neural network model training device capable of being remotely controlled according to claim 1, wherein: the rotating rod (10) is sleeved with a torsion spring, and two ends of the torsion spring are fixedly connected with the supporting block (9) and the training device body (2) respectively.

5. The convolutional neural network model training device capable of being remotely controlled according to claim 1, wherein: a guide rod (19) is fixedly connected in the sliding groove, and a through hole corresponding to the guide rod (19) is formed in the sliding block (7).

6. The convolutional neural network model training device capable of being remotely controlled according to claim 1, wherein: the one end of cavity is equipped with the installing port, rotate through pivot (20) in the installing port and be connected with baffle (21), the one end that pivot (20) were kept away from in baffle (21) passes through locking screw (22) fixed connection with the installing port inner wall, be equipped with the thermovent on baffle (21), the thermovent internal fixation is inserted and is equipped with dust screen (23).

7. The convolutional neural network model training device capable of being remotely controlled according to claim 1, wherein: a power supply connector (24) is fixedly inserted on the outer wall of the training device body (2), and the output end of the power supply connector (24) is connected with the input end of the power supply (13).