CN219550088U

CN219550088U - Data annotation model training system

Info

Publication number: CN219550088U
Application number: CN202320365214.3U
Authority: CN
Inventors: 王华威; 杨早; 韩春; 贺信信
Original assignee: Qingdao Zhonglan Xinhe Information Technology Co ltd
Current assignee: Qingdao Zhonglan Xinhe Information Technology Co ltd
Priority date: 2023-03-02
Filing date: 2023-03-02
Publication date: 2023-08-18
Anticipated expiration: 2033-03-02

Abstract

The utility model discloses a data annotation model training system, which relates to the technical field of data annotation model training and aims at the problems that an existing data annotation model training system is inconvenient to automatically turn pages of pictures to be resolved, a camera cannot be rotated during recognition, more recognition angles cannot be provided, and the use effect is not ideal. The utility model has reasonable design, is convenient for automatically turning pages of pictures to be resolved, is convenient for rotating the camera during recognition, can provide more recognition angles, has good use effect and high working efficiency.

Description

Data annotation model training system

Technical Field

The utility model relates to the technical field of data annotation model training, in particular to a data annotation model training system.

Background

The data marking is to label the pictures to be identified and resolved by the computer in advance, so that the computer can continuously identify the characteristics of the pictures, and finally, the computer can automatically identify the pictures. The data label provides a large amount of labeled data for the artificial intelligence enterprises, so that the machine training and learning can be realized, and the effectiveness of the algorithm model can be ensured. The meaning of the data standard model training is that the accuracy of computer identification is improved, and the data standard model training efficiency can be improved, and the workload of staff is reduced, so that a data annotation model training system is needed.

However, the existing data annotation model training system is inconvenient to automatically turn pages of pictures to be resolved, is inconvenient to rotate a camera during recognition, cannot provide more recognition angles, and is not ideal in use effect.

Disclosure of Invention

The utility model aims to solve the defects that the existing data annotation model training system is inconvenient to automatically turn pages of pictures to be resolved, the camera is inconvenient to rotate during recognition, more recognition angles cannot be provided, and the using effect is not ideal.

In order to achieve the above purpose, the present utility model adopts the following technical scheme:

the utility model provides a data annotation model training system, includes the box, place the touch-control display screen on the bottom inner wall of box, fixedly mounted has servo motor on the bottom inner wall of box, fixedly welded with worm on servo motor's the output shaft, fixedly mounted has the cam on the worm, rotate on the one side inner wall of box and be connected with the cross axle that the level set up, fixedly mounted has worm wheel and initiative bevel gear on the cross axle, the worm with the worm wheel meshes, rotate on the rear side inner wall of box and be connected with the transmission shaft, fixedly mounted has driven bevel gear and disc on the transmission shaft, initiative bevel gear with driven bevel gear meshes;

two fixing bases are fixedly installed on the disc, two sliding connection is arranged on one sides, away from each other, of the fixing bases, a sliding rod is arranged on one side, away from each other, of the sliding rod, a touch control head is fixedly installed on one end, away from each other, of the sliding rod, the touch control head is matched with a touch control display screen, an installation seat is fixedly installed on the inner wall of a box body, an identification camera is rotatably arranged on the installation seat, a gear is fixedly sleeved on the identification camera, a rack is slidably arranged on the inner wall of the rear side of the box body, the rack is meshed with the gear, a cam is in rolling connection with the rack, a first spring is fixedly arranged on one end, away from the rack, of the cam, and the inner wall of the box body is fixedly arranged on one end, away from the first spring.

In a preferred embodiment, two stabilizing seats are fixedly welded on the inner wall of the rear side of the box body, and the rack and the stabilizing seats are arranged in a sliding manner.

In a preferred embodiment, the fixing base is provided with a sliding groove, the inner wall of the sliding groove is provided with two limiting grooves, the sliding rod is fixedly provided with a connecting seat, the connecting seat is arranged in the corresponding limiting groove in a sliding mode, the top of the connecting seat is fixedly provided with a second spring, and the top end of the second spring is fixedly arranged with the inner wall of the top of the limiting groove.

In a preferred embodiment, a placement groove is formed in the inner wall of the bottom of the box body, and the touch display screen is located in the placement groove.

In a preferred embodiment, four supporting legs are fixedly welded at the bottom of the box body, and the four supporting legs are symmetrically arranged in pairs.

In a preferred embodiment, an inner ring of a bearing is fixedly sleeved on the transverse shaft, a supporting rod is fixedly welded on an outer ring of the bearing, and the bottom end of the supporting rod is fixedly arranged with the inner wall of the bottom of the box body.

According to the data annotation model training system, a servo motor, a touch display screen and an identification camera are started, the rotation of an output shaft of the servo motor drives a worm to rotate, a cam arranged on the outer side of the worm is driven to rotate by the rotation of the worm, the cam is in rolling connection with a rack, the rack is driven to do left and right reciprocating motion under the action of elastic force of a first spring, and the rack and a gear are meshed to drive the identification camera to do reciprocating swing, so that the effect of model training on pictures displayed on the touch display screen at different angles is achieved;

according to the data annotation model training system, the worm is meshed with the worm wheel to drive the transverse shaft to rotate, the transverse shaft rotates to drive the driving bevel gear to rotate, the driving bevel gear is meshed with the driven bevel gear to drive the transmission shaft to rotate, the rotation of the transmission shaft drives the fixing seat to rotate, the rotation of the fixing seat drives the sliding rod to rotate, the effect of automatic image changing is achieved when the touch control head is contacted with the touch control display screen, and the training efficiency of the data annotation model is improved;

the utility model has reasonable design, is convenient for automatically turning pages of pictures to be resolved, is convenient for rotating the camera during recognition, can provide more recognition angles, has good use effect and high working efficiency.

Drawings

FIG. 1 is a schematic diagram of a three-dimensional structure of a training system for a data annotation model according to the present utility model;

FIG. 2 is a schematic diagram of a portion A of a training system for a data annotation model according to the present utility model;

FIG. 3 is a front view of a training system for a data annotation model according to the present utility model;

FIG. 4 is a schematic diagram of a cross-sectional structure of a portion B of a data annotation model training system according to the present utility model.

In the figure: 1. a case; 2. a touch display screen; 3. a servo motor; 4. a worm; 5. a cam; 6. a horizontal axis; 7. a worm wheel; 8. a drive bevel gear; 9. a transmission shaft; 10. a driven bevel gear; 11. a disc; 12. a fixing seat; 13. a slide bar; 14. a touch control head; 15. a mounting base; 16. identifying a camera; 17. a gear; 18. a rack; 19. a first spring; 20. a stabilizing seat; 21. a chute; 22. a limit groove; 23. a connecting seat; 24. and a second spring.

Detailed Description

The following description of the embodiments of the present utility model will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present utility model, but not all embodiments.

Referring to fig. 1-4, one embodiment of the present solution provides: the utility model provides a data annotation model training system, including box 1, touch-control display screen 2 has been placed on the bottom inner wall of box 1, servo motor 3 is fixed mounting on the bottom inner wall of box 1, fixedly welded with worm 4 on servo motor 3's the output shaft, fixedly mounted has cam 5 on worm 4, rotationally be connected with horizontal cross axle 6 that sets up on the inner wall of one side of box 1, fixedly mounted has worm wheel 7 and initiative bevel gear 8 on cross axle 6, worm 4 meshes with worm wheel 7, rotationally be connected with transmission shaft 9 on the rear side inner wall of box 1, fixedly mounted has driven bevel gear 10 and disc 11 on transmission shaft 9, initiative bevel gear 8 meshes with driven bevel gear 10;

two fixing bases 12 are fixedly mounted on the disc 11, one sides, away from each other, of the two fixing bases 12 are all connected with sliding rods 13 in a sliding mode, one ends, away from each other, of the two sliding rods 13 are all fixedly mounted with a touch control head 14, the touch control head 14 is matched with a touch control display screen 2, a mounting seat 15 is fixedly mounted on the inner wall of the box body 1, an identification camera 16 is rotatably arranged on the mounting seat 15, a gear 17 is fixedly sleeved on the identification camera 16, a rack 18 is slidably arranged on the inner wall of the rear side of the box body 1, the rack 18 is meshed with the gear 17, a cam 5 is in rolling connection with the rack 18, a first spring 19 is fixedly arranged at one end, away from the cam 5, of the rack 18, and one end, away from the rack 18, of the first spring 19 is fixedly arranged on the inner wall of the box body 1.

Referring to fig. 1, in this embodiment, a placement groove is formed in an inner wall of a bottom of a case 1, and a touch display screen 2 is located in the placement groove, so as to facilitate positioning of the touch display screen 2.

Referring to fig. 1, in this embodiment, four legs are fixedly welded at the bottom of the case 1, and the four legs are symmetrically arranged in pairs, so as to play a role in stabilizing and supporting the case 1.

Referring to fig. 3, in this embodiment, an inner ring of a bearing is fixedly sleeved on a transverse shaft 6, an outer ring of the bearing is fixedly welded with a supporting rod, and the bottom end of the supporting rod is fixedly arranged on the inner wall of the bottom of the box body 1, so that the rotation of the transverse shaft 6 is stably supported.

Referring to fig. 3, in the present embodiment, two stabilizing seats 20 are fixedly welded on the inner wall of the rear side of the case 1, and the rack 18 and the stabilizing seats 20 are slidably disposed to guide the movement of the rack 18

Referring to fig. 4, in the present embodiment, a sliding groove 21 is formed on the fixing base 12, two limiting grooves 22 are formed on the inner wall of the sliding groove 21, a connecting seat 23 is fixedly arranged on the sliding rod 13, the connecting seat 23 is slidably arranged in the corresponding limiting groove 22, a second spring 24 is fixedly arranged on the top of the connecting seat 23, and the top end of the second spring 24 is fixedly arranged on the inner wall of the top of the limiting groove 22, so that the touch control head 14 can slide on the touch control display screen 2 conveniently.

The working principle is that firstly, the servo motor 3, the touch display screen 2 and the identification camera 16 are started, the rotation of an output shaft of the servo motor 3 drives the worm 4 to rotate, the rotation of the worm 4 drives the cam 5 arranged on the outer side of the worm to rotate, the cam 5 is in rolling connection with the rack 18, the rack 18 is driven to do left-right reciprocating motion under the action of the elastic force of the first spring 19, the rack 18 is meshed with the gear 17 to drive the identification camera 16 to do reciprocating swing, the effect of model training of pictures displayed on the touch display screen 2 at different angles is achieved, the worm 4 is meshed with the worm gear 7 to drive the transverse shaft 6 to rotate, the rotation of the transverse shaft 6 drives the drive bevel gear 8 to rotate, the drive bevel gear 8 is meshed with the driven bevel gear 10 to drive the transmission shaft 9 to rotate, the rotation of the transmission shaft 9 drives the fixed seat 12 to rotate, the rotation of the fixed seat 12 drives the slide bar 13 to rotate, the automatic picture changing effect is achieved when the touch display screen 14 is contacted with the touch display screen 2, and the efficiency of data labeling model training is improved.

In the description of the present utility model, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings are merely for convenience in describing the present utility model and simplifying the description, and do not indicate or imply that the apparatus or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the present utility model.

Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present utility model, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.

The preferred embodiments of the utility model disclosed above are intended only to assist in the explanation of the utility model. The preferred embodiments are not exhaustive or to limit the utility model to the precise form disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the utility model and the practical application, to thereby enable others skilled in the art to best understand and utilize the utility model. The utility model is limited only by the claims and the full scope and equivalents thereof.

Claims

1. A data annotation model training system, comprising:

the automatic control device comprises a box body (1), wherein a touch display screen (2) is placed on the inner wall of the bottom of the box body (1), a servo motor (3) is fixedly installed on the inner wall of the bottom of the box body (1), a worm (4) is fixedly welded on an output shaft of the servo motor (3), a cam (5) is fixedly installed on the worm (4), a horizontal shaft (6) which is horizontally arranged is rotationally connected on the inner wall of one side of the box body (1), a worm wheel (7) and a driving bevel gear (8) are fixedly installed on the horizontal shaft (6), the worm (4) is meshed with the worm wheel (7), a transmission shaft (9) is rotationally connected on the inner wall of the rear side of the box body (1), a driven bevel gear (10) and a disc (11) are fixedly installed on the transmission shaft (9), and the driving bevel gear (8) is meshed with the driven bevel gear (10).

Two fixing bases (12) are fixedly installed on the disc (11), two the equal sliding connection in one side that fixing bases (12) kept away from each other has slide bar (13), two the equal fixed mounting of one end that slide bar (13) kept away from each other has touch-control head (14), touch-control head (14) with touch-control display screen (2) looks adaptation, fixed mounting has mount pad (15) on the inner wall of box (1), it is provided with discernment camera (16) to rotate on mount pad (15), fixed cover is equipped with gear (17) on discernment camera (16), sliding is provided with rack (18) on the rear side inner wall of box (1), rack (18) with gear (17) are meshed, cam (5) with rack (18) roll connection, rack (18) keep away from the one end of cam (5) is fixed to be provided with first spring (19), first spring (19) keep away from one end of rack (18) with the inner wall of box (1) is fixed.

2. A data annotation model training system as claimed in claim 1, wherein: two stabilizing seats (20) are fixedly welded on the inner wall of the rear side of the box body (1), and the racks (18) and the stabilizing seats (20) are arranged in a sliding mode.

3. A data annotation model training system as claimed in claim 1, wherein: the fixing seat (12) is provided with a sliding groove (21), the inner wall of the sliding groove (21) is provided with two limiting grooves (22), the sliding rod (13) is fixedly provided with a connecting seat (23), the connecting seat (23) is slidably arranged in the corresponding limiting groove (22), the top of the connecting seat (23) is fixedly provided with a second spring (24), and the top end of the second spring (24) is fixedly arranged with the inner wall of the top of the limiting groove (22).

4. A data annotation model training system as claimed in claim 1, wherein: a placing groove is formed in the inner wall of the bottom of the box body (1), and the touch display screen (2) is located in the placing groove.

5. A data annotation model training system as claimed in claim 1, wherein: four supporting legs are fixedly welded at the bottom of the box body (1), and the four supporting legs are symmetrically arranged in pairs.

6. A data annotation model training system as claimed in claim 1, wherein: an inner ring of a bearing is fixedly sleeved on the transverse shaft (6), a supporting rod is fixedly welded on the outer ring of the bearing, and the bottom end of the supporting rod is fixedly arranged on the inner wall of the bottom of the box body (1).