Intelligent traffic flow monitoring machine based on deep learning

Abstract

The invention provides an intelligent traffic flow monitoring machine based on deep learning, which comprises an unmanned aerial vehicle, an intelligent traffic information acquisition system and a supporting mechanism, a gear ring is fixed on a lower carrier plate at the bottom of the unmanned aerial vehicle, a driving gear and a track wheel are mounted on the upper side of each motor seat plate, and the driving gears are meshed with the outer edge of the fixed gear ring; a driving motor is fixed to the bottom of each motor base plate, a rotating shaft of each driving motor and a driving gear are installed together, a radial rod is fixed to the inner end of each motor base plate, a positioning sleeve is fixed to the inner end of each radial rod, a center column is fixed to the center of the bottom of the lower carrier plate, and the center column is sleeved with the positioning sleeves respectively. And supporting mechanisms are respectively fixed at the outer ends of the motor base plates of the rotary adjusting mechanisms. According to the invention, flexible rotation and adjustment of monitoring equipment are realized, requirements of different monitoring positions and angles are met, road traffic flow information is monitored in real time, traffic management is optimized, and road safety and traffic efficiency are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent traffic detection equipment, and in particular relates to an intelligent traffic flow monitoring machine based on deep learning. Background Art

[0002] The intelligent traffic flow monitoring system is a comprehensive system based on advanced information technology, sensor technology and communication technology, which is used to monitor and analyze road traffic conditions in real time. The system can collect and process a large amount of traffic data, including traffic volume, speed, vehicle classification, traffic density, vehicle spacing, lane occupancy, traffic events, traffic flow time distribution, road environment information and signal light status. The real-time monitoring and analysis of these data are of great significance to the traffic management department to optimize traffic signal control, improve road design, and improve traffic safety and efficiency, which helps to reduce traffic congestion, reduce traffic accident rates and improve travel efficiency.

[0003] Existing intelligent traffic flow monitoring systems mainly rely on fixed monitoring equipment, such as traffic cameras, geomagnetic coils, radars, and lidars. These devices are usually installed in fixed locations, such as roadsides, traffic signal poles, and under viaducts. These fixed devices can provide stable data collection, but their coverage is limited, and the monitoring location cannot be flexibly adjusted, making it difficult to meet the needs of dynamically changing traffic monitoring. In addition, some mobile monitoring equipment such as vehicle-mounted cameras and drones are also gradually being used, but their monitoring capabilities and endurance are subject to certain limitations, making it difficult to achieve long-term, high-precision monitoring.

[0004] Common problems and shortcomings of existing technologies. Fixed monitoring equipment can only cover specific areas, cannot flexibly adjust the monitoring position, and is difficult to cope with dynamically changing traffic conditions. Although some mobile monitoring equipment has been available, they are usually unable to maintain a stable horizontal state during the monitoring process, resulting in reduced data accuracy. Existing mobile devices usually lack precise leveling and position adjustment mechanisms. Existing mobile monitoring equipment is usually unable to stay on the road guardrail at any time and lacks the ability to perform stable monitoring at different locations. In addition, existing equipment generally has the problem of insufficient battery life, making it difficult to conduct long-term continuous monitoring, and lacks a convenient way to charge midway.

[0005] In order to achieve more comprehensive and flexible traffic monitoring, it is necessary to develop an intelligent traffic flow monitoring system that can move flexibly, stay stably, have a long endurance and can be charged midway. Summary of the invention

[0006] In response to the common defects and problems of existing similar products, the present invention provides an intelligent traffic flow monitoring machine based on deep learning, which aims to monitor road traffic information in real time and accurately, and provide stable data support to optimize traffic management and improve road safety and traffic efficiency.

[0007] The solution of the present invention to solve its technical problem is: adopt an intelligent traffic flow monitoring machine based on deep learning, including a drone, an intelligent traffic information collection system and a supporting mechanism, a rotation adjustment mechanism is installed at the bottom of the drone, the rotation adjustment mechanism includes a fixed gear ring, a motor seat plate and a driving motor, a gear ring is fixed on the download plate at the bottom of the drone, a driving gear and a track wheel are installed on the upper side of each motor seat plate, and a blocking cap is fixed on the top of the rotating shaft of each driving gear and track wheel; when the motor seat plate is assembled with the fixed gear ring, the driving gear is meshed with the outer edge teeth of the fixed gear ring The track wheel fits with the inner edge of the fixed gear ring, and each stop cap is located on the upper side of the inner and outer edges of the fixed gear ring; a driving motor is fixed to the bottom of each motor base plate, and the rotating shaft of each driving motor is installed together with the corresponding driving gear, wherein the driving motor has a self-locking function; a radial rod is fixed to the inner end of each motor base plate, and a positioning sleeve is fixed to the inner end of each radial rod, and a central column is fixed to the bottom center of the download plate at the same time, and each positioning sleeve is respectively mounted on the outer side of the central column and can rotate around the central column; a supporting mechanism is fixed to the outer end of the motor base plate of the rotation adjustment mechanism.

[0008] Preferably, the supporting mechanism includes a cross brace, a rotating sleeve and a supporting foot. The rotating sleeve is located at the end of the cross brace, and the supporting foot is located at the lower end of the rotating sleeve. When the controller controls each driving motor to rotate, each motor seat plate can be moved to a relatively concentrated position, so that each supporting mechanism can be brought closer to each other. The rotating sleeve in the middle is mounted on the inner side of the guardrail, and the rotating sleeves on both sides are supported on the outer side of the guardrail. The forces on both sides are balanced, and the rotational pressure of each driving gear causes the inner and outer rotating sleeves to squeeze and support the inner and outer sides of the guardrail respectively.

[0009] Preferably, an elastic pressing mechanism is added in each cross brace rod, wherein the elastic pressing mechanism is an inner sleeve rod sleeved on the inner side of the cross brace rod, a spring seat is fixed to the inner end of the inner sleeve rod, a fixed stop is provided in the inner cavity of the cross brace rod, a thrust spring is sleeved between the spring seat and the fixed stop, and an inner stop is fixed on the cross brace rod for limiting the spring seat and preventing it from moving outward.

[0010] Preferably, a central axis is vertically fixed to the outer end of the inner sleeve rod, outer baffles are respectively fixed to the upper and lower ends of the central axis, the rotating sleeve is mounted on the central axis between the upper and lower outer baffles; and the supporting foot is vertically fixed to the bottom of the central axis.

[0011] Preferably, the intelligent traffic information collection system is fixed to the top of the upper loading plate on the upper part of the drone, or is movably assembled on the top of the upper loading plate on the upper part of the drone.

[0012] Preferably, a plurality of cantilever fixing seats are evenly distributed and fixed around the periphery between the upper loading plate and the downloading plate, a cantilever is fixed to the outer end of each cantilever fixing seat, a rotor motor is fixed to the outer end of each cantilever, a rotor blade is installed on the upper side of the rotating shaft of each rotor motor, and an independent battery is fixed at the root of each cantilever, each independent battery supplies power to the corresponding rotor motor, and the area between the upper loading plate and the downloading plate is used for installing the flight controller, the power control system and the communication system.

[0013] Preferably, it also includes an auxiliary charging mechanism, a fixing seat is fixed on the top of the charging column, a hanging plate is fixed to the fixing seat through a connecting rod, a perforation is provided on the rear side of the hanging plate and a guide column is installed matchingly, a movable pressure plate is fixed at the rear end of each guide column, an inner plate is fixed on the front side of each guide column, and a return spring is mounted on the rear side of each guide column; elastic conductors are respectively fixed on the front sides of the left and right inner plates, each elastic conductor is respectively connected to the power cord, and a cylindrical conductor is fixed in the middle of the central axis, and the two cylindrical conductors are respectively connected to the input of the charging controller through the power cord.

[0014] Beneficial effects of the present invention: The intelligent traffic flow monitoring machine based on deep learning has significant advantages in real-time monitoring of road traffic information, optimizing traffic management, and improving road safety and traffic efficiency.

[0015] 1. Flexibility and adaptability: By fixing the gear ring, driving gear and track wheel, the monitoring equipment can be flexibly rotated and adjusted to meet the needs of different monitoring positions and angles. The support mechanism design allows the equipment to stay firmly on the road guardrail or the ground to adapt to a variety of road conditions and environments.

[0016] 2. Battery life and charging convenience: Charging poles are set up on highways, so that the monitoring machine can be charged conveniently, improving battery life and ensuring the continuity of long-term monitoring. The contact design of elastic conductor and columnar conductor ensures good electrical contact during charging, improving charging efficiency and reliability.

[0017] 3. Stability and accuracy: The flight power supply and the detection instrument power supply are separated to avoid the interference of the endurance current and ensure the accuracy of data collection. When the drone moves, the intelligent traffic information collection system is automatically leveled to ensure that the equipment always remains horizontal and improve the monitoring accuracy.

[0018] 4. Versatility and comprehensiveness: It can collect traffic volume, speed, vehicle classification, traffic density, vehicle spacing, lane occupancy, traffic events, traffic flow time distribution, road environment information and signal light status, etc., and provide comprehensive traffic information. It combines video surveillance, sensors, GPS data and mobile device data to ensure the comprehensiveness and reliability of information collection.

[0019] 5. Safety and stability: The slider linkage design prevents the impact of bad weather (such as strong wind) on the charging process, improving the safety and stability of the charging process. The self-locking function of the drive motor ensures the stability of the device in a fixed position to prevent accidental movement or vibration. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the monitoring machine of the present invention in the unfolded state; Figure 2 It is a schematic diagram of a retention state of the monitoring machine of the present invention; Figure 3 yes Figure 1 Schematic diagram of the lower structure; Figure 4 yes Figure 1 Schematic diagram comparing the state of looking down and looking up; Figure 5 It is a schematic diagram of the coordination relationship between the rotary adjustment mechanism and the supporting mechanism; Figure 6 It is a schematic diagram of the internal structure of the support mechanism; Figure 7 It is a structural schematic diagram of an auxiliary charging mechanism; Figure 8 is a schematic diagram of the structure of another auxiliary charging mechanism; Fig. 9 It is a partial enlarged structural schematic diagram of the rotary adjustment mechanism.

[0021] Reference numerals in the figure: UAV 1; intelligent traffic information collection system 2; rotation adjustment mechanism 3; support mechanism 4; auxiliary charging mechanism 5; automatic locking mechanism 6; auxiliary leveling mechanism 7; loading board 11; unloading board 12; cantilever fixing seat 13; cantilever 14; rotor motor 15; rotor blade 16; independent battery 17; fixed gear ring 31; fixing hole 32; motor seat plate 33; driving gear 34; driving motor 35; track wheel 36; radial rod 37 ; Positioning sleeve 38; cross brace rod 41; inner sleeve rod 42; spring seat 43; thrust spring 44; central axis 45; rotating sleeve 46; outer baffle 47; supporting foot 48; inner baffle 49; fixing seat 51; hanging plate 52; movable pressure plate 53; inner plate 54; guide column 55; reset spring 56; elastic conductor 57; cylindrical conductor 58; annular groove 61; pull rope 62; upper left slider 63; upper right slider 64; lower left slider 65; lower right slider 66. DETAILED DESCRIPTION

[0022] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0023] Example 1: An intelligent traffic flow monitoring machine based on deep learning is used to monitor road traffic conditions in real time. It can be automatically or remotely controlled to move with the traffic flow for monitoring; or to stay steadily at the road guardrail for monitoring; or to stay on the ground for monitoring. The improvement aims to solve the problems of limited coverage, insufficient mobile monitoring function, and insufficient staying and endurance capabilities in the prior art, provide a more comprehensive and efficient traffic monitoring solution, and improve the accuracy and effectiveness of traffic management. Figure 1 Figure 5 As shown, the monitoring machine mainly includes a drone 1, an intelligent traffic information collection system 2, a rotation adjustment mechanism 3 and a support mechanism 4. Figure 2 The drone 1 shown in the figure includes an upload board 11, a download board 12, a cantilever fixing seat 13, a cantilever 14, a rotor motor 15, a rotor blade 16 and an independent battery 17. A plurality of cantilever fixing seats 13 are evenly fixed around the upload board 11 and the download board 12. A cantilever 14 is fixed to the outer end of each cantilever fixing seat 13. A rotor motor 15 is fixed to the outer end of each cantilever 14. A rotor blade 16 is installed on the upper side of the rotating shaft of each rotor motor 15. In addition, an independent battery 17 is fixed at the root of each cantilever 14. Each independent battery 17 supplies power to the corresponding rotor motor 15, so that the flight power supply is separated from the power supply of the detection instrument to avoid interference with the accuracy of the monitoring data due to excessive endurance current. The area between the upload board 11 and the download board 12 is used to install the flight controller, the power control system and the communication system.

[0024] An intelligent traffic information collection system 2 is installed on the top of the upper board 11. The intelligent traffic information collection system 2 can use an existing intelligent traffic monitoring system, such as the traffic data collection equipment ATN 910 series, the traffic information collection system Sitraffic Conduct+ or the road traffic monitoring equipment ZTE TrafficEye, etc., and mainly monitors the following road traffic information. Traffic flow: the number of vehicles passing a certain point per unit time, which can be divided into lanes, directions and types; vehicle speed: the speed of a single vehicle, the average speed; vehicle classification: classification and statistics based on vehicle type (such as cars, trucks, buses, etc.); traffic density: the number of vehicles on a unit length of road; vehicle spacing: the time interval or distance between adjacent vehicles; lane occupancy rate: the proportion of time that a lane is occupied by vehicles; traffic events: monitoring and recording of events such as traffic accidents, vehicle failures, and illegal parking. Time distribution of traffic flow: traffic flow distribution in different time periods (such as peak hours and off-peak hours); road environment information: factors affecting traffic flow such as weather conditions, road conditions, and road construction; signal light status: the real-time status of each signal light and its impact on traffic flow. The monitoring of this information can be achieved through a variety of technical means, such as video surveillance, sensors (such as geomagnetic coils, radars, lidars, etc.), GPS data, mobile device data, etc. The collection and analysis of this data can help traffic management departments optimize traffic signal control, improve road design, and improve traffic safety and efficiency.

[0025] The intelligent traffic information collection system 2 and the drone 1 can be directly fixed or installed in an adjustable manner. For example, an auxiliary leveling mechanism is installed between the upper plate 11 and the bottom of the intelligent traffic information collection system 2 to control the left and right flipping adjustment of the intelligent traffic information collection system 2. Since the intelligent traffic information collection system 2 will not undergo obvious state changes when the drone 1 is in a hovering state, lifting or rotating, but when the drone 1 moves laterally, the entire fuselage of the drone 1 will automatically tilt, causing the intelligent traffic information collection system 2 to tilt, thereby affecting its monitoring effect. The tilt can be corrected by software to achieve the purpose of precise quantitative control.

[0026] A rotation adjustment mechanism 3 is installed at the bottom of the drone 1. Figure 6 and Fig. 9As shown. The rotation adjustment mechanism 3 includes a fixed gear ring 31, a fixing hole 32, a motor base plate 33, a driving gear 34, a driving motor 35, a track wheel 36, a radial rod 37 and a positioning sleeve 38. A support is provided at the bottom of the download plate 12, and each support is fixed to the fixed gear ring 31 through the fixing hole 32. A driving gear 34 and a track wheel 36 are installed on the upper side of each motor base plate 33, and a stop cap is fixed on the top of the rotating shaft of each driving gear 34 and track wheel 36. When the motor base plate 33 is assembled with the fixed gear ring 31, the driving gear 34 is meshed with the outer edge teeth of the fixed gear ring 31, the track wheel 36 is fitted with the inner edge of the fixed gear ring 31, and each stop cap is located on the upper side of the inner and outer edges of the fixed gear ring 31. A driving motor 35 is fixed to the bottom of each motor base plate 33, and the rotating shaft of each driving motor 35 is installed together with the corresponding driving gear 34, wherein the driving motor 35 has a self-locking function. A radial rod 37 is fixed to the inner end of each motor base plate 33, and a positioning sleeve 38 is fixed to the inner end of each radial rod 37. At the same time, a central column is fixed to the bottom center of the download plate 12, and each positioning sleeve 38 is respectively sleeved on the outer side of the central column and can rotate around the central column.

[0027] The outer ends of the motor base plates 33 of the rotation adjustment mechanism 3 are respectively fixed with support mechanisms 4, which include a cross brace 41 and a support foot 48. When the controller controls each drive motor 35 to rotate, each motor base plate 33 can be moved to a relatively centralized position, such as Figure 2 and Figure 6 As shown, each support mechanism 4 can be brought closer to each other. The rotating sleeve 46 in the middle is mounted on the inner side of the guardrail, and the rotating sleeves 46 on both sides are supported on the outer side of the guardrail, and the forces on both sides are balanced. Through the rotation pressure of each driving gear 34, the inner and outer rotating sleeves 46 are urged to squeeze and support the inner and outer sides of the guardrail respectively. In order to improve the support strength and stability, an elastic top pressure mechanism can be added in each cross brace rod 41. Figure 6 As shown, the elastic pressing mechanism can be an inner sleeve rod 42 mounted inside the cross brace rod 41, a spring seat 43 is fixed to the inner end of the inner sleeve rod 42, a fixed stop is provided in the inner cavity of the cross brace rod 41, and a thrust spring 44 is mounted between the spring seat 43 and the fixed stop. At the same time, an inner stop 49 is fixed on the cross brace rod 41 to limit the spring seat 43 from moving outward. A central axis 45 is vertically fixed to the outer end of the inner sleeve rod 42, and outer stops 47 are respectively fixed above and below the central axis 45. A rotating sleeve 46 is mounted on the central axis 45 between the upper and lower outer stops 47, and the rotating sleeve 46 can rotate along the central axis 45. A supporting foot 48 is vertically fixed to the bottom of the central axis 45. The supporting foot 48 can be located on a horizontal ground, and the rotating sleeve 46 can be supported on the outside of the guardrail.

[0028] Retention and rotation monitoring: By controlling the support mechanisms 4 to be evenly distributed along the circumference, Figure 1 and Figure 3 As shown, each support leg 48 is supported on the bottom surface of both sides of the road or the built foundation. When each driving motor 35 is controlled to rotate, the drone 1 and the intelligent traffic information collection system 2 can be rotated along the circumference as a whole based on the support mechanism 4.

[0029] The above monitoring machine can be applied to various road monitoring, and can be controlled to be placed on the guardrails on both sides of the road or on the ground. It is particularly suitable for moving and staying at any point of the highway guardrail to perform monitoring tasks. The above monitoring machine can also be equipped with special charging facilities to improve its endurance. For example, charging poles are set up in each section of the highway, and an auxiliary charging mechanism 5 is installed on the top of the charging pole. The monitoring machine can stay on the guardrail or on the upper side of the mounting plate 52 of the auxiliary charging mechanism 5 for charging.

[0030] like Figure 7 In the figure, the auxiliary charging mechanism 5 includes a fixed seat 51, a hanging plate 52, a movable pressure plate 53, an inner plate 54, a guide column 55, a reset spring 56, an elastic conductor 57 and a cylindrical conductor 58. A fixed seat 51 is fixed on the top of the charging column, and the hanging plate 52 is fixed to the fixed seat 51 through a connecting rod. A perforation is provided on the rear side of the hanging plate 52 and a guide column 55 is matched and installed. A movable pressure plate 53 is fixed at the rear end of each guide column 55, and an inner plate 54 is fixed on the front side of each guide column 55. A reset spring 56 is sleeved on the rear side of each guide column 55. In a natural state, each reset spring 56 extends backward, so that each guide column 55 and the movable pressure plate 53 are in the rear position. Elastic conductors 57 are respectively fixed on the front sides of the left and right inner plates 54, and each elastic conductor 57 is respectively connected to the power cord. A cylindrical conductor 58 is fixed in the middle of the central axis 45, and the two cylindrical conductors 58 are respectively connected to the input of the charging controller through the power cord. As shown Figure 7 In the figure, after the rotating sleeve 46 in the middle is mounted on the inner side of the movable pressure plate 53, the rotating sleeves 46 on the two side edges are supported on the front side of the hanging plate 52. As the supporting mechanisms 4 on both sides shrink inward, the movable pressure plate 53 is squeezed toward the hanging plate 52, thereby making the elastic conductor 57 and the columnar conductor 58 in close contact to maintain a good connection state.

[0031] Based on the auxiliary charging mechanism 5, in order to improve the stability of the detection machine charging process, an automatic locking mechanism 6 can be added to avoid the instability of the charging state caused by severe weather such as strong winds. Figure 8As shown, a form of automatic locking mechanism 6 includes an annular groove 61, a drawstring 62, an upper left slider 63, an upper right slider 64, a lower left slider 65 and a lower right slider 66. The annular groove 61 includes an upper T-shaped groove, left and right T-shaped grooves and a front T-shaped groove. The upper left slider 63, the upper right slider 64, the lower left slider 65 and the lower right slider 66 are respectively T-shaped and are matched and fitted in corresponding grooves. A right drawstring 62 is connected between the upper left slider 63 and the lower right slider 66, a left drawstring 62 is connected between the upper right slider 64 and the lower left slider 65, and tension springs are connected to the middle or end portions of the drawstrings 62 on the left and right sides. An inner top spring is connected between the upper left slider 63 and the upper right slider 64. When the rotating sleeve 46 in the middle is mounted on the rear side wall of the movable pressing plate 53, the rotating sleeves 46 on the left and right sides are symmetrically supported on the front side wall of the hanging plate 52. As the rotating sleeves 46 on the left and right sides roll inward respectively, the lower left slider 65 and the lower right slider 66 are continuously driven to move inward, thereby pulling the upper right slider 64 and the upper left slider 63 to move inward. Finally, the upper left slider 63 and the upper right slider 64 are clamped on both sides of the middle cross brace 41 to form a stable clamping relationship, thereby improving the retention stability during the charging process.

[0032] Based on the above-mentioned deep learning-based intelligent traffic flow monitoring machine solution, the specific usage steps are as follows: 1. Equipment startup and inspection: Turn on the power of the drone, check the working status of each rotor motor, cantilever battery and intelligent traffic information collection system to ensure the normal operation of the equipment. Ensure that the drone and its subsystems are in good working condition to provide basic guarantee for subsequent monitoring.

[0033] 2. Flight control and positioning: Use the remote control or preset program to control the drone to take off, and use GPS or other positioning systems to move the drone to the designated monitoring location. The drone is accurately located in the road area that needs to be monitored and is ready to start data collection.

[0034] 3. Monitoring data collection: Start the intelligent traffic information collection system and start collecting data such as traffic volume, speed, vehicle classification, traffic density, vehicle spacing, lane occupancy, traffic events, traffic flow time distribution, road environment information and signal light status in real time. Comprehensively collect and record road traffic information to provide real-time and accurate data support. Real-time video data collection: Integrate high-definition cameras in the intelligent traffic information collection system of drones to collect road video data in real time. The camera covers a wide area and provides high-resolution video streams. High-definition real-time video data provides high-quality input for deep learning models to ensure the accuracy and real-time nature of monitoring. Preprocess the video data through edge computing devices to extract key frames, remove noise, and compress the data volume. The preprocessed data is transmitted to the deep learning model processing unit through the wireless communication system. Preprocessing improves data transmission efficiency, reduces latency, and provides high-quality data input for deep learning models.

[0035] 4. Data transmission and processing: The collected traffic data is transmitted to the traffic management center or cloud server in real time through the wireless communication system for processing and analysis. The traffic management department can monitor the road conditions in real time, make timely traffic management decisions, and improve the efficiency of road traffic management. Collect a large amount of historical traffic data and video surveillance data, and use deep learning algorithms (such as convolutional neural network CNN and recurrent neural network RNN) for model training. The training model can identify different types of vehicles, predict changes in traffic flow, and detect traffic events (such as accidents, illegal parking). Through large amounts of data training, the model has a high degree of accuracy and can effectively identify and predict complex traffic conditions.

[0036] 5. Move and reposition: According to traffic management needs, control the drone to move between different monitoring points and repeat the above leveling and data collection process. This enables comprehensive monitoring of different road areas and flexible response to dynamically changing traffic conditions.

[0037] 6. Stay monitoring: At key points where long-term monitoring is required, the drone is controlled to land and stay firmly on the road guardrail or the ground through the support mechanism to continue monitoring. Long-term stable monitoring is achieved at specific locations to provide continuous traffic data support. When the drive motors are controlled to rotate simultaneously, the drone and the intelligent traffic information collection system can rotate in a circle as a whole based on the support mechanism.

[0038] 7. Mid-trip charging: When the drone is low on battery, it will fly to the charging column on the highway and automatically dock and charge through the auxiliary charging mechanism. The automatic locking mechanism will be activated to ensure the stability of the charging process. This will improve the drone's endurance and meet the needs of long-term, continuous monitoring.

[0039] 8. Data analysis and feedback: The traffic management center analyzes the real-time transmitted data, generates traffic flow reports, accident reports, congestion forecasts, etc., and provides feedback to adjust traffic lights, traffic diversion and other management measures. Optimize traffic signal control, improve road design, improve traffic safety and efficiency, and reduce congestion and accident rates. Input the pre-processed video data into the pre-trained deep learning model for real-time analysis. The model automatically identifies vehicle types, counts traffic flow, calculates vehicle speeds, detects traffic incidents, etc. The deep learning model can quickly and accurately analyze traffic conditions, generate detailed traffic data and reports, and improve monitoring efficiency and accuracy. The deep learning model can not only identify regular traffic information, but also detect abnormal traffic incidents such as traffic accidents, vehicle failures, illegal parking, etc. The model automatically marks and alarms abnormal events. Real-time detection and alarm of traffic incidents are achieved, and traffic management departments can respond quickly to improve road safety.

[0040] 9. Equipment maintenance and upgrade: Regularly inspect and maintain the drone and its various subsystems, and update software and hardware to ensure long-term stable operation and functional improvement of the equipment. Extend the service life of the equipment, improve equipment performance and monitoring effects, and adapt to the ever-changing traffic monitoring needs.

[0041] Through the above usage steps, the deep learning-based intelligent traffic flow monitoring machine can realize efficient, flexible and accurate road traffic monitoring, provide important data support for traffic management departments, optimize traffic management, and improve road safety and traffic efficiency.

[0042] The above specific embodiments of the present invention are only used to illustrate or explain the principles of the present invention, and do not constitute a limitation of the present invention. Therefore, any modification, equivalent substitution, improvement, etc. made without departing from the spirit and scope of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent traffic flow monitoring machine based on deep learning, comprising a drone (1), an intelligent traffic information collection system (2) and a supporting mechanism (4), characterized in that: A rotation adjustment mechanism (3) is installed at the bottom of the drone (1), and the rotation adjustment mechanism (3) comprises a fixed gear ring (31), a motor base plate (33) and a drive motor (35). The gear ring (31) is fixed to the download plate (12) at the bottom of the drone (1), and a driving gear (34) and a track wheel (36) are installed on the upper side of each motor base plate (33), and a stop cap is fixed to the top of the rotating shaft of each driving gear (34) and track wheel (36); when the motor base plate (33) and the fixed gear ring (31) are assembled, the driving gear (34) and the outer edge teeth of the fixed gear ring (31) are meshed, the track wheel (36) and the inner edge of the fixed gear ring (31) are abutted, and each stop cap is located at The upper sides of the inner and outer edges of the fixed gear ring (31); a driving motor (35) is fixed at the bottom of each motor base plate (33), and the rotating shaft of each driving motor (35) is installed together with the corresponding driving gear (34), wherein the driving motor (35) has a self-locking function; a radial rod (37) is fixed at the inner end of each motor base plate (33), and a positioning sleeve (38) is fixed at the inner end of each radial rod (37), and a central column is fixed at the bottom center of the download plate (12), and each positioning sleeve (38) is respectively sleeved on the outer side of the central column and can rotate around the central column; and a supporting mechanism (4) is respectively fixed at the outer end of the motor base plate (33) of the rotation adjustment mechanism (3).

2. The deep learning-based intelligent traffic flow monitoring device according to claim 1 is characterized in that: The support mechanism (4) comprises a cross brace (41), a rotating sleeve (46) and a supporting foot (48), wherein the rotating sleeve (46) is located at the end of the cross brace (41), and the supporting foot (48) is located at the lower end of the rotating sleeve (46). When the controller controls each driving motor (35) to rotate, each motor seat plate (33) can be moved to a relatively concentrated position, thereby enabling each supporting mechanism (4) to be closer to each other. The rotating sleeve (46) located in the middle is mounted on the inner side of the guardrail, and the rotating sleeves (46) located on both sides are supported on the outer side of the guardrail. The forces on both sides are balanced, and the rotation pressure of each driving gear (34) causes the inner and outer rotating sleeves (46) to squeeze and support the inner and outer sides of the guardrail respectively.

3. The deep learning-based intelligent traffic flow monitoring device according to claim 2 is characterized in that: An elastic pressing mechanism is added in each cross brace (41), wherein an inner sleeve rod (42) is sleeved inside the cross brace (41), a spring seat (43) is fixed to the inner end of the inner sleeve rod (42), a fixed stop is provided in the inner cavity of the cross brace (41), a thrust spring (44) is sleeved between the spring seat (43) and the fixed stop, and an inner stop (49) is fixed on the cross brace (41) for limiting the spring seat (43) and preventing it from moving outward.

4. The deep learning-based intelligent traffic flow monitoring device according to claim 3 is characterized in that: A central shaft (45) is vertically fixed to the outer end of the inner sleeve rod (42); outer baffles (47) are respectively fixed to the upper and lower ends of the central shaft (45); the rotating sleeve (46) is sleeved on the central shaft (45) between the upper and lower outer baffles (47); and the supporting foot (48) is vertically fixed to the bottom of the central shaft (45).

5. The deep learning-based intelligent traffic flow monitoring device according to claim 1 is characterized in that: The intelligent traffic information collection system (2) is fixed to the top of an upper loading plate (11) on the upper part of the drone (1), or is movably assembled on the top of the upper loading plate (11) on the upper part of the drone (1).

6. The deep learning-based intelligent traffic flow monitoring device according to claim 5 is characterized in that: A plurality of cantilever fixing seats (13) are evenly distributed and fixed around the periphery between the upper loading plate (11) and the downloading plate (12); a cantilever (14) is fixed to the outer end of each cantilever fixing seat (13); a rotor motor (15) is fixed to the outer end of each cantilever (14); a rotor blade (16) is mounted on the upper side of the rotating shaft of each rotor motor (15); and an independent battery (17) is fixed at the root of each cantilever (14); each independent battery (17) supplies power to the corresponding rotor motor (15); and the area between the upper loading plate (11) and the downloading plate (12) is used to install a flight controller, a power control system, and a communication system.

7. The deep learning-based intelligent traffic flow monitoring device according to claim 1 is characterized in that: It also includes an auxiliary charging mechanism (5), wherein a fixing seat (51) is fixed on the top of the charging column, a hanging plate (52) is fixed to the fixing seat (51) via a connecting rod, two left and right inner plates (54) are fixed on the front side of the hanging plate (52), elastic conductors (57) are respectively fixed on the front sides of the left and right inner plates (54), each elastic conductor (57) is respectively connected to a power line, a columnar conductor (58) is fixed in the middle of the central axis (45), and the two columnar conductors (58) are respectively connected to the input of the charging controller via the power line.

Images