Image Automatic Annotation Method Based on Joint Learning of Multi-Level Feature Spaces of Smart Lamp Posts

By installing observation cameras and adaptive cameras on smart street lights, combined with lift control components and sensing modules, real-time calculation and early warning of the height of truck cargo is achieved, safety hazards in the height-limited road sections are solved, and road safety is improved.

CN116631189BActive Publication Date: 2025-07-18NANJING ZHONGZHI TENGFEI AVIATION SCI & TECH RES INST CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202310624276.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-05-30
Publication Date
2025-07-18
Estimated Expiration
2043-05-30

AI Technical Summary

Technical Problem

In the prior art, vehicles in other places are prone to accidents caused by excessive cargo height due to unfamiliar routes and loading conditions on high-level sections, and existing smart lamp pole systems are difficult to effectively mark and early warning.

Method used

Install an observation camera and an adaptive camera on the smart street light. Through joint learning of multi-level feature spaces, the truck features are identified, combined with lift control components and sensing modules, the camera height is adjusted in real time, cargo height is calculated and marked, and the driver is prompted on the reminder screen.

Benefits of technology

It has improved the safety of smart cities and reduced traffic accidents on high-limited road sections through real-time early warning and response measures.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116631189B_ABST
    Figure CN116631189B_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of smart light poles. Specifically, an image automatic annotation method based on joint learning of multi-level feature spaces of smart light poles includes smart street lamps, reminder screens, two observation cameras, a lifting control component, an adaptive camera, and a sensing module. A number of smart street lamps are selectively installed on reasonable sections in the city, and the smart street lamps are vertically symmetrically arranged on both sides of the road. Two observation cameras are set on the smart street lamps to perform multi-level feature calculations on passing vehicles, conduct multi-space image learning on vehicles with target features, and determine the driving risks of the vehicles. When it is necessary to annotate the loading height of a truck, the use of the lifting control component is coordinated to adjust the height position of the adaptive camera, so that the height of the target vehicle can be calculated and annotated even in harsh environments, and then timely next-step response operations are carried out to improve the road safety and response efficiency of the smart city.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of smart lamp posts, and specifically to an automatic image annotation method based on joint learning of multi-level feature spaces of smart lamp posts. Background Technique

[0002] Smart street lamps are based on road lighting lamp posts and use networks to integrate carriers of various traffic assistance functions. They are an important part of the construction of smart cities. Joint learning of multi-level feature spaces is an algorithm for target multi-level feature tracking and recognition through complementary information and consistency information between different views at different spatial positions.

[0003] In the prior art, when automatically annotating target multi-level feature images, the target situation is photographed and recorded, and the system algorithm automatically obtains the feature situation of the target to be tracked. Among them, there are some sections of the urban road that are height-limited sections. Trucks need to ensure that the stacking height of the goods in their carriages is lower than this value when passing through. However, in the case of unfamiliar routes for out-of-town vehicles and non-self-loaded goods, accidents are still likely to occur. If a smart lamp post with joint learning of multi-level feature spaces is used to automatically annotate the image and give a timely reminder, this event may be avoided. Summary of the Invention

[0004] The purpose of the present invention is to provide an automatic image annotation method based on joint learning of multi-level feature spaces of smart lamp posts to solve the problems raised in the above background technique.

[0005] To achieve the above purpose, the present invention provides the following technical solution: An automatic image annotation method based on joint learning of multi-level feature spaces of smart lamp posts, the automatic image annotation method based on joint learning of multi-level feature spaces of smart lamp posts includes smart street lamps, reminder screens, two observation cameras, a lifting control component, an adaptive camera, and a sensing module, and has the following steps:

[0006] Step 1: Selectively install several smart street lamps on reasonable sections in the city, and vertically and symmetrically arrange the smart street lamps on both sides of the road;

[0007] Step 2: Make the reminder screen face the driver's direction, and the height of the smart street lamp should be higher than the height of the height-limiting pole on the rear road;

[0008] Step 3: When the vehicle is driving normally, automatically identify the characteristics of the vehicles passing by on the road through the observation camera and the adaptive camera. When it is recognized that the vehicle is a truck, automatically obtain vehicle and driver information through Internet information;

[0009] Step 4: Then, through smart street lamps at multiple positions on the road, automatically analyze multiple images of the vehicle, observe the characteristics of the stacking of the goods carried by the vehicle, and calculate the safety and stability of the goods in multiple spaces and with multiple features.

[0010] Step Five: When the vehicle passes by the intelligent street lamp, the height of the adaptive camera is controlled by the lifting control component so that the center of the captured image is flush with the highest point of the upper end of the vehicle cargo, and at the same time, the mental state characteristics of the driver are directly recorded and analyzed;

[0011] Step Six: When the adaptive camera is at the height of the vehicle cargo, the accurate measurement of the cargo height is carried out through the set sensing module to avoid image errors caused by weather such as rain, snow, etc., and the cargo height is box-selected and marked in the collected images;

[0012] Step Seven: When the cargo height is lower than the height limit, only the stability of the cargo and the mental state of the driver are analyzed at multiple levels. When the cargo height is higher than the height limit, the stability of the cargo is poor, and the mental state of the driver is poor, a text prompt is given on the reminder screen, and a text message notification is sent through the driver information obtained in advance.

[0013] Preferably, the reminder words on the reminder screen are in eye-catching red fonts, and the font size can be selected as required according to the size and set height of the reminder screen. The license plate and risk problems can be respectively displayed on the reminder screens of two adjacent intelligent street lamps.

[0014] Preferably, after the reminder screen and text message reminder, when the vehicle continues to move forward and successively passes by the intelligent street lamps near the height limit position, the driver is notified by phone ringing, and preparations for possible road traffic accidents are arranged in advance to provide fast handling for road traffic and personnel safety.

[0015] Preferably, the intelligent street lamp includes a street lamp column and a cantilever bracket. The cantilever bracket is horizontally arranged on one side of the upper end of the street lamp column. The cantilever bracket is arranged in a U-shaped structure. A protective side plate is vertically arranged on one side of the cantilever bracket. The reminder screen is vertically embedded on one side of the protective side plate. Two observation cameras are respectively arranged on the cantilever bracket near both sides of the protective side plate, and the thickness of the cantilever bracket is greater than the thickness of the protective side plate.

[0016] Preferably, a lighting lamp is horizontally arranged at the lower end of the cantilever bracket. The lifting control component includes a lifting frame. The lifting frame is vertically installed on one side of the lower end of the cantilever bracket near the street lamp column. A movable groove is vertically penetrated through the center of one side of the lifting frame. Sliding grooves are symmetrically arranged on both sides in the movable groove. A movable box is horizontally inserted in the movable groove. Two sliding strips are vertically symmetrically arranged on both sides of the movable box, and the two sliding strips are respectively inserted into the two sliding grooves in a movable manner.

[0017] Preferably, a rotating groove is horizontally opened on one side of the movable box. A bearing cylinder is horizontally and movably inserted into the rotating groove through a bearing. A transparent cover is arranged on one side of the bearing cylinder. The adaptive camera is horizontally arranged in the rotating groove, and one side of the adaptive camera penetrates through the bearing cylinder and is inserted into the transparent cover.

[0018] Preferably, the movable box is inserted into the movable groove and has planar grooves symmetrically on both sides, the two planar grooves are respectively connected to the two sides of the rotating groove, the bearing cylinder is placed in the rotating groove and has a rotating gear on one side, the rotating gear is respectively inserted into the two planar grooves on both sides, two give way grooves are symmetrically opened on one side of the movable groove close to the rotating gear, a number of toggle teeth are symmetrically arranged in one side of the give way groove, and the rotating gear passes through one side of the rotating groove and is meshed with the toggle teeth.

[0019] Preferably, a cleaning cover is horizontally provided at the upper end of one side of the movable box, the upper end of the adaptable camera is plugged into the lower end of the cleaning cover, and a cleaning cloth is provided on the side of the cleaning cover in contact with the adaptable camera.

[0020] Preferably, the sensor module is horizontally inserted on the side of the movable box away from the transparent cover, a groove is vertically opened on the side of the street lamp column close to the lifting frame, a vertical plate is vertically inserted in the groove, a plurality of sensing marks are symmetrically arranged on one side of the vertical plate, and one side of the sensor module points to the sensing mark arrangement on one side of the vertical plate.

[0021] Preferably, rope connecting seats are provided at both upper and lower ends of the movable box, and a lifting rope is commonly connected to the two rope connecting seats. Two shaft support seats are symmetrically provided on the upper and lower sides of the movable groove, and a control wheel is vertically provided between the two shaft support seats. A rotating shaft is horizontally provided at the center of the control wheel. Two shaft support seats are movably inserted on both sides of the rotating shaft. The rotating shaft close to the lighting lamp side passes through one side of the shaft support seat and a lifting motor is horizontally provided. The lifting rope is sleeved on the two control wheels. An extension groove is inserted and provided on one side of the control wheel in the movable groove, and a guide roller is horizontally provided on the lifting frame on one side of the extension groove. The lifting rope is sleeved on one side of the shaft support seat and passes through the extension groove and overlaps the guide roller.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] By setting up two observation cameras on the smart street lights to perform multi-level feature calculations on passing vehicles, multi-space image learning is performed on vehicles with target features to determine the vehicle's driving risk. When it is necessary to mark the height of the truck's cargo, the height position of the adaptive camera is adjusted in conjunction with the use of the lifting control component, so that it can calculate and mark the height of the target vehicle even in harsh environments, and then take timely next steps to improve the road safety and response efficiency of smart cities. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram of the structure of the present invention;

[0025] Figure 2 For the present invention Figure 1 Schematic diagram of the structure of part A;

[0026] Figure 3 This is a schematic diagram of the three-dimensional structure of the smart street lamp in the present invention from a first perspective;

[0027] Figure 4 A schematic diagram of the second viewing angle of the three-dimensional structure of the smart street lamp in the present invention;

[0028] Figure 5 It is a schematic diagram of the cantilever bracket structure of the present invention;

[0029] Figure 6 This is a schematic diagram of the vertical plate structure of the present invention;

[0030] Figure 7 This is a schematic diagram of the structure of the lifting control assembly of the present invention;

[0031] Figure 8 It is a schematic diagram of the connection structure of the activity box of the present invention.

[0032] In the figure: smart street light 1, street light column 101, cantilever bracket 2, protective side panel 3, reminder screen 4, observation camera 5, lighting lamp 6, lifting frame 7, movable slot 8, slide slot 9, movable box 10, transparent cover 11, bearing cylinder 12, transfer gear 13, plane slot 14, adaptive camera 15, cleaning cover 16, slide bar 17, sensor module 18, vertical plate 19, induction mark 20, lifting rope 21, shaft support seat 22, control wheel 23, lifting motor 24, guide roller 25, give way slot 26, toggle tooth 27. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the technical solutions in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] Please refer to the attached Figure 1-8 , this application provides the following five preferred embodiments.

[0035] Embodiment 1

[0036] The automatic image annotation method based on the joint learning of multi-level feature space of smart lamp poles includes a smart street lamp 1, a reminder screen 4, two observation cameras 5, a lifting control component, an adaptation camera 15 and a sensor module 18, and has the following steps:

[0037] Step 1: Selectively install a number of intelligent street lights 1 on reasonable sections in the city, and vertically and symmetrically arrange the intelligent street lights 1 on both sides of the road. The intelligent street light 1 includes a street light column 101 and a cantilever bracket 2. The cantilever bracket 2 is horizontally arranged on one side of the upper end of the street light column 101. The cantilever bracket 2 is arranged in a U-shaped structure. A protective side plate 3 is vertically arranged on one side of the cantilever bracket 2. A reminder screen 4 is vertically embedded on one side of the protective side plate 3. Two observation cameras 5 are respectively arranged on the cantilever bracket 2 near both sides of the protective side plate 3, and the thickness of the cantilever bracket 2 is greater than the thickness of the protective side plate 3 to shield the reminder screen 4 to a certain extent and improve the service life;

[0038] Step 2: Make the reminder screen 4 face the driver's direction, and the height of the intelligent street light 1 should be higher than the height of the height limit pole on the rear road;

[0039] Step 3: When the vehicle is driving normally, automatically identify the characteristics of the vehicles passing on the road through the observation cameras 5 and the adaptive cameras 15. When it is recognized that the vehicle is a truck, automatically obtain the vehicle and driver information through the Internet information;

[0040] Step 4: Then, through the intelligent street lights 1 at multiple positions on the road, automatically analyze the multi-pictures taken of the vehicle, observe the characteristics of the stacking of the goods carried by the vehicle, and calculate the safety and stability of the goods in multiple spaces and with multiple characteristics.

[0041] Step 5: When the vehicle passes by the smart street lamp 1, the height of the adaptive camera 15 is controlled by the lifting control component so that the center of the shooting picture is flush with the highest point of the upper end of the vehicle cargo, and the mental state characteristics of the driver are directly recorded and analyzed. The lower end of the cantilever bracket 2 is horizontally provided with a lighting lamp 6, and the lifting control component includes a lifting frame 7. The lifting frame 7 is vertically installed on the lower end of the cantilever bracket 2 near the street lamp column 101. A movable groove 8 is opened through the center of one side of the lifting frame 7, and sliding grooves 9 are symmetrically opened on both sides of the movable groove 8. An movable box 10 is horizontally inserted in the movable groove 8, and two slide bars 17 are vertically symmetrically arranged on both sides of the movable box 10. The two slide bars 17 are respectively movable Two slide grooves 9 are arranged for dynamic insertion, a rotating groove is horizontally opened on one side of the movable box 10, a bearing cylinder 12 is arranged in the rotating groove through the horizontal movable insertion of the bearing, a transparent cover 11 is arranged on one side of the bearing cylinder 12, an adaptable camera 15 is arranged horizontally in the rotating groove, and one side of the adaptable camera 15 penetrates the bearing cylinder 12 and is inserted in the transparent cover 11, and the movable box 10 is inserted into the movable groove 8 and plane grooves 14 are symmetrically opened on both sides, and the two plane grooves 14 are respectively connected with the two sides of the rotating groove, and the bearing cylinder 12 is placed in the rotating groove and is sleeved with a transfer gear 13 on one side, and the transfer gear 13 is respectively inserted and placed in the two plane grooves 14 on both sides, and the movable groove 8 One side is symmetrically opened near the transfer gear 13 One side A yield groove 26 is symmetrically provided with a plurality of toggle teeth 27 in one side of the yield groove 26, and the gear 13 passes through one side of the rotating groove and is meshed with the toggle teeth 27. Rope connectors are provided at both ends of the movable box 10, and a lifting rope 21 is connected to the two rope connectors. Two shaft support seats 22 are symmetrically provided on the upper and lower sides of the movable groove 8, and a control wheel 23 is vertically provided between the two shaft support seats 22. A rotating shaft is horizontally provided at the center of the control wheel 23, and two shaft support seats 22 are movably inserted on both sides of the rotating shaft. A lifting motor 24 is horizontally provided on one side of the rotating shaft passing through the shaft support seat 22 close to the lighting lamp 6, and the lifting rope 21 is sleeved on the two control wheels 23. The slot 8 is provided with a protruding slot inserted and inserted on one side of the control wheel 23, and the lifting frame 7 is provided with a guide roller 25 horizontally on one side of the protruding slot. The lifting rope 21 sleeves the shaft support seat 22 on one side and respectively penetrates the protruding slot and overlaps the guide roller 25, so that the height position of the adaptation camera 15 can be raised and lowered. Under the guidance of the slide bar 17, the activity is stable. When there is frost and rain and snow on the surface of the transparent cover 11, the control activity box 10 is reciprocated up and down several times, and the gear 13 is turned to reciprocate along the toggle tooth 27 to make the transparent cover 11 and the cleaning cover 16 frictionally clean, so as to improve the accuracy of the shooting picture of the adaptation camera 15 in rainy and snowy weather and avoid detection errors of the sensor module;

[0042] The lifting rope 21 and the shaft support seat 22 can be a chain and sprocket structure, and the cleaning cover 16 can also be equipped with electric heating to assist drying.

[0043] Step 6: When the adaptive camera 15 is at the height of the vehicle cargo, through the set sensing module 18, accurately measure the cargo height, avoid image errors caused by weather such as rain and snow, and perform frame selection and annotation of the cargo height in the collected images;

[0044] Step 7: When the cargo height is lower than the height limit, only perform multi-level feature analysis on the stability of the cargo and the mental state of the driver. When the cargo height is higher than the height limit, the stability of the cargo is poor, and the mental state of the driver is poor, a text prompt is given on the reminder screen 4, and a text message notification is sent through the pre-obtained driver information.

[0045] The image automatic annotation method based on multi-level feature space joint learning of a smart light pole disclosed in the second embodiment of the present invention has basically the same structure as that in the first embodiment. The reminder words on the reminder screen 4 are in eye-catching red fonts, and the font size can be selected as needed according to the size and set height of the reminder screen 4. License plates and risk issues can be respectively displayed on the reminder screens 4 of two adjacent smart street lights 1.

[0046] The image automatic annotation method based on multi-level feature space joint learning of a smart light pole disclosed in the third embodiment of the present invention has basically the same structure as that in the second embodiment. After the reminder screen 4 and the text message reminder, when the vehicle continues to move forward and successively passes by the smart street lights 1 near the height limit position, a ringing notification is given to the driver, and preparations for possible road traffic accidents are arranged in advance, providing a quick handling for road traffic and personnel safety.

[0047] The image automatic annotation method based on multi-level feature space joint learning of a smart light pole disclosed in the fourth embodiment of the present invention has basically the same structure as that in the third embodiment. The difference is that a cleaning cover 16 is horizontally provided at the upper end of one side of the movable box 10, the upper end of the adaptive camera 15 is inserted and arranged at the lower end of the cleaning cover 16, and a cleaning cloth is provided on the side of the cleaning cover 16 in contact with the adaptive camera 15, which is convenient for cleaning.

[0048] The image automatic annotation method based on multi-level feature space joint learning of a smart light pole disclosed in the fifth embodiment of the present invention has basically the same structure as that in the fourth embodiment. The difference is that the sensing module 18 is horizontally inserted and arranged on the side of the movable box 10 away from the transparent cover 11. A groove is vertically opened on the side of the street lamp column 101 close to the lifting frame 7, and a vertical plate 19 is vertically inserted in the groove. A plurality of induction points 20 are symmetrically arranged on one side of the vertical plate 19, and one side of the sensing module 18 points to the induction points 20 on one side of the vertical plate 19, which is used for auxiliary calculation of the height position sensing module of the adaptive camera 15 to ensure that the loading height of the truck can still be measured and annotated in harsh environments.

[0049] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An image automatic annotation method based on joint learning of multi-level feature spaces of intelligent light poles, characterized in that The automatic image annotation method based on the joint learning of multi-level feature spaces of smart lamp posts includes smart street lamps (1), reminder screens (4), two observation cameras (5), a lifting control component, an adaptive camera (15), and a sensing module (18), and has the following steps: Step 1: Selectively install several smart street lamps (1) on reasonable sections in the city, and vertically and symmetrically arrange the smart street lamps (1) on both sides of the road; Step 2: Make the reminder screen (4) face the driver's direction, and the height of the smart street lamp (1) should be higher than the height of the height limit pole on the rear road; Step 3: When the vehicle is driving normally, automatically identify the characteristics of the vehicles passing by on the road through the observation camera (5) and the adaptive camera (15). When a truck is identified, automatically obtain vehicle and driver information through Internet information; Step 4: Then, through the smart street lamps (1) at multiple positions on the road, automatically analyze the multi-pictures taken of the vehicle, observe the characteristics of the stacking of the goods carried by the vehicle, and calculate the safety and stability of the goods through multi-space and multi-features; Step 5: When the vehicle passes by the smart street lamp (1), control the height of the adaptive camera (15) through the lifting control component, so that the center of the captured picture is flush with the highest point of the upper end of the vehicle's goods, and at the same time directly record and analyze the mental state characteristics of the driver; Step 6: When the adaptive camera (15) is at the height of the vehicle's goods, accurately measure the height of the goods through the set sensing module (18), avoid image errors caused by rain and snow weather, and frame and mark the height of the goods in the collected images; Step 7: When the height of the goods is lower than the height limit, only perform multi-level feature analysis on the stability of the goods and the mental state of the driver. When the height of the goods is higher than the height limit, the stability of the goods is poor, and the mental state of the driver is poor, give a text prompt on the reminder screen (4), and send a text message notification through the previously obtained driver information.

2. The automatic image annotation method based on joint learning of multi-level feature spaces of smart light poles according to claim 1, wherein: The reminder words on the reminder screen (4) are in eye-catching red fonts, and the font size is selected as needed according to the size and set height of the reminder screen (4). The license plate and risk issues are respectively displayed on the reminder screens (4) of two adjacent smart street lamps (1).

3. The automatic image annotation method based on the joint learning of multi-level feature spaces of smart lamp posts according to claim 2, wherein: After the reminder screen (4) and the text message reminder, if the vehicle continues to move forward and successively passes by the smart street lamps (1) near the height limit position, ring the driver's phone, and make preparations in advance for possible road traffic accidents, providing quick handling for road traffic and personnel safety.

4. The automatic image annotation method based on joint learning of multi-level feature spaces of intelligent lamp posts according to claim 3, characterized in that: The smart street lamp (1) includes a lamp post (101) and a cantilever bracket (2). The cantilever bracket (2) is horizontally arranged on one side of the upper end of the lamp post (101). The cantilever bracket (2) is arranged in a U-shaped structure. A protective side plate (3) is vertically arranged on one side of the cantilever bracket (2). The reminder screen (4) is vertically embedded on one side of the protective side plate (3). The two observation cameras (5) are respectively arranged on both sides of the cantilever bracket (2) close to the protective side plate (3), and the thickness of the cantilever bracket (2) is greater than the thickness of the protective side plate (3).

5. The automatic image annotation method based on multi-level feature space joint learning of smart light poles according to claim 4, wherein: The lower end of the cantilever bracket (2) is horizontally provided with an illumination lamp (6), and the lifting control assembly comprises a lifting frame (7), the lifting frame (7) is vertically installed on the lower end of the cantilever bracket (2) near the side of the street lamp column (101), a movable groove (8) is provided through the center of one side of the lifting frame (7), and sliding grooves (9) are symmetrically provided on both sides of the movable groove (8), and a movable box (10) is horizontally inserted in the movable groove (8), and two sliding bars (17) are vertically symmetrically provided on both sides of the movable box (10), and the two sliding bars (17) are respectively movably inserted into the two sliding grooves (9).

6. The automatic image annotation method based on joint learning of multi-level feature spaces of smart light poles according to claim 5, characterized in that: A rotating groove is horizontally opened on one side of the movable box (10), a bearing cylinder (12) is horizontally inserted in the rotating groove through a bearing, a transparent cover (11) is provided on one side of the bearing cylinder (12), an adaptable camera (15) is horizontally arranged in the rotating groove, and one side of the adaptable camera (15) penetrates the bearing cylinder (12) and is inserted into the transparent cover (11) to be arranged.

7. The automatic image annotation method based on multi-level feature space joint learning of intelligent light poles according to claim 6, wherein: The movable box (10) is inserted into the movable groove (8) and has planar grooves (14) symmetrically provided on both sides thereof. The two planar grooves (14) are respectively connected to the two sides of the rotating groove. The bearing cylinder (12) is arranged in the rotating groove and is sleeved with a rotating gear (13) on one side thereof. The two sides of the rotating gear (13) are respectively inserted and placed in the two planar grooves (14). Two clearance grooves (26) are symmetrically provided on one side of the movable groove (8) close to the rotating gear (13). A plurality of shifting teeth (27) are symmetrically provided in one side of the clearance groove (26). The rotating gear (13) passes through one side of the rotating groove and is meshed and connected with the shifting teeth (27).

8. The automatic image annotation method based on joint learning of multi-level feature spaces of intelligent light poles according to claim 7, wherein: A cleaning cover (16) is horizontally arranged at the upper end of one side of the movable box (10), the upper end of the adaptable camera (15) is plugged into the lower end of the cleaning cover (16), and a cleaning cloth is arranged on the side of the cleaning cover (16) that contacts the adaptable camera (15).

9. The method for automatically annotating images based on joint learning of multi-level feature spaces of intelligent lamp poles according to claim 8, wherein: The sensor module (18) is horizontally plugged into a side of the movable box (10) away from the transparent cover (11), and a groove is vertically provided on a side of the street lamp column (101) close to the lifting frame (7), and a vertical plate (19) is vertically plugged into the groove. A plurality of sensing marks (20) are symmetrically provided on one side of the vertical plate (19), and one side of the sensor module (18) is arranged to point to the sensing marks (20) on one side of the vertical plate (19).

10. The automatic image annotation method based on multi-level feature space joint learning of smart light poles according to claim 9, wherein: The movable box (10) is provided with rope connection seats at both upper and lower ends, and the two rope connection seats are connected together with a lifting rope (21). Two shaft support seats (22) are symmetrically provided on both upper and lower sides of the movable groove (8), and a control wheel disc (23) is vertically provided between the two shaft support seats (22). A rotating shaft is horizontally provided at the center of the control wheel disc (23), and two shaft support seats (22) are movably inserted on both sides of the rotating shaft. The rotating shaft close to the lighting lamp (6) passes through the shaft support seat (22) and a lifting motor (24) is horizontally provided on one side. The lifting rope (21) is sleeved with the two control wheels (23). A protruding groove is inserted and provided on one side of the control wheel disc (23) in the movable groove (8), and a guide roller (25) is horizontally provided on the lifting frame (7) on the side of the protruding groove. The lifting rope (21) sleeves on one side of the shaft support seat (22) and passes through the protruding groove and overlaps the guide roller (25).

Citation Information

Patent Citations

  • Automatic entry and exit management system and method thereof

    CN115588249A

  • Intelligent street lamp capable of serving as 5G base station

    CN211822075U