Vehicle identification method, street lamp fault judgment method and intelligent street lamp energy-saving control system

Abstract

According to the vehicle identification and street lamp fault judgment method and the intelligent street lamp energy-saving control system, the improved European-style clustering algorithm is adopted to identify vehicles running at a high speed, and static accumulated snow and dynamic vehicles are distinguished through four stages of point cloud preprocessing, improved European-style clustering, vehicle verification and time sequence tracking, so that the false detection rate caused by rain and snow noise is reduced. A street lamp sectional type control scheme is adopted, a certain road section is divided into a plurality of lamp control subsystems, and a main control module in each lamp control subsystem controls a plurality of street lamps in the lamp control subsystem to achieve the purpose of energy conservation. Real-time monitoring of street lamp faults is realized by adopting a dynamic threshold adjustment mechanism, an alarm threshold is dynamically adjusted according to historical data, and the problem that a fixed threshold cannot adapt to load change is avoided, so that the problem of false alarm caused by power grid interference is reduced. The system can effectively deal with rain and snow weather, and can control a plurality of street lamps through one master control module, thereby greatly reducing the reconstruction cost.

Description

Technical Field

[0001] The present invention belongs to the technical field of road lighting, and specifically relates to a vehicle identification and street lamp fault judgment method and an intelligent street lamp energy-saving control system. Background Art

[0002] With the acceleration of urbanization, the scale of public lighting construction is increasing. The number of urban street lighting in China has increased from 21.995 million in 2013 to 36.145 million in 2023, and is expected to reach 37.9577 million in 2024. This has formed a widespread urban street lighting network with huge power consumption. Therefore, how to effectively implement energy conservation has attracted much attention. Determining the appropriate illumination level is also a prerequisite for energy conservation. Illumination below the national standard cannot achieve good lighting conditions. Average illumination above the national standard not only wastes energy but also easily causes light interference and reduces visual visibility for motorists and pedestrians.

[0003] To save energy, a variety of smart streetlight management solutions are currently available. For example, Chinese patent CN117042234A discloses an intelligent streetlight lighting system based on dynamic traffic flow. In this system, a lidar sensor and an ultrasonic radar sensor are installed on each streetlight body to detect vehicles and control the video acquisition of the camera on the streetlight body.

[0004] Chinese patent CN115623642A discloses a smart streetlight management system and control method. Each streetlight controller is installed on a smart streetlight. Each smart streetlight controller includes a main control unit, a photosensor, a power meter, a radar sensor, a dimming unit, and a communication unit. The provided smart streetlight management system and method enable different control methods to be used at different time periods. In the smart energy-saving control mode, the brightness of streetlights within a preset distance threshold can be dynamically and stepwise adjusted.

[0005] The street lighting system disclosed in the above patent has the following deficiencies:

[0006] (1) Patent CN117042234A uses laser radar and ultrasonic radar sensors to sense vehicles: 1) When the system is operating in rainy or snowy weather, the sensors may cause detection failure; 2) The control scheme uses a street lamp integrated controller installed on each street lamp, including laser radar, camera, ultrasonic radar, etc., which greatly increases the cost of street lamp modification;

[0007] (2) The streetlight lighting system disclosed in patent CN117042234A: 1) During the third preset time period, if the radar sensor does not sense any vehicle driving, the streetlight is in the off state. When a vehicle driving is detected, the brightness of the streetlight is calculated based on the vehicle speed and the streetlight is turned on. This control scheme causes the streetlight to be in a frequent on-off state, which will greatly reduce the service life of the streetlight; 2) When the streetlight is turned from off to on, the driver's eyes will experience a very obvious "light adaptation" phenomenon due to the change in light from dark to bright, which may cause the driver to make errors in judging the road conditions and increase the potential safety hazards of driving. Summary of the Invention

[0008] The present invention aims to provide a vehicle identification and streetlight fault diagnosis method, as well as an intelligent streetlight energy-saving control system, that effectively handles rainy and snowy weather. This system controls multiple streetlights through a single master control module, significantly reducing retrofit costs. Furthermore, a dynamic energy consumption threshold adjustment mechanism is employed to detect streetlight faults, reducing false alarms caused by power grid disturbances.

[0009] To achieve the above object, the technical solution adopted by the present invention is:

[0010] In a first aspect, the present invention provides a vehicle recognition method based on an improved Euclidean clustering algorithm for identifying high-speed vehicles, comprising four stages: point cloud preprocessing, improved Euclidean clustering, vehicle verification, and time series tracking. The vehicle recognition method includes the following:

[0011] Get the current precipitation intensity and define the weather intensity coefficient ω to be calculated in real time using the following formula:

[0012]

[0013] Where γ is the nonlinear adjustment factor; the benchmark threshold is calculated by the following formula, where 30 is the heavy rain benchmark threshold and 5 is the heavy snow benchmark threshold;

[0014]

[0015] Dynamically update the intensity filter threshold I according to the weather intensity coefficient th , with the dynamically updated intensity filter threshold I th The original point cloud data of the collected vehicle is denoised by two methods: using the sliding window denoising method and the sliding window denoising method to obtain the denoised point cloud data.

[0016] Neighborhood search radius d th Dynamic adjustment with weather:

[0017] d th =d base +k d ·ω

[0018] Where: d base : sunny basic threshold; k d : distance adjustment coefficient; ω: weather intensity coefficient, ω∈[0,1], 0 is sunny, 1 is extreme rain and snow;

[0019] Dynamic minimum number of cluster points N min Dynamic adjustment with weather:

[0020] N min =N base +k n ·ω

[0021] Where: N base : Minimum number of points on sunny days; k n : Point adjustment coefficient;

[0022] The denoised point cloud Establish a KD tree and search the neighborhood with radius d th and the dynamic minimum number of cluster points N min Perform region growing and clustering for constraints;

[0023] Clusters that satisfy the vehicle size constraint are retained, and the vehicle position is predicted and updated using Kalman filtering, correlating the detection results of consecutive frames and outputting the vehicle state.

[0024] Furthermore, the intensity filtering threshold I th The relationship between I and the weather intensity coefficient ω is: th =40+60ω; if the point cloud intensity I i Less than the intensity filtering threshold I th , then remove it.

[0025] Furthermore, the sliding window denoising process is: for 5 consecutive points {p i-2 , p i-1 , p i , p i+1 , p i+2 If the distance between two adjacent points is greater than 0.5m, the point in the middle is considered noise and is removed. That is, if the following conditions are met:

[0026] |p i -p i-1 |>0.5m and |p i -p i+1 |>0.5m

[0027] Then p i It is determined to be noise and removed.

[0028] In a second aspect, the present invention provides a method for determining a street lamp fault, the method comprising the following steps:

[0029] 1) Collect the voltage and current data of street lamps in chronological order and eliminate abnormal data based on the gross error judgment criterion;

[0030] 2) Set a sliding window and calculate the moving average of the power within the window based on the collected voltage and current data, that is, the baseline value, and simultaneously calculate the standard deviation σ of the power data within the window;

[0031] 3) Set the range of dynamic energy consumption threshold based on the moving average and standard deviation:

[0032] The upper limit of the dynamic threshold of energy consumption = moving average + kσ;

[0033] The lower limit of the energy consumption dynamic threshold = moving average - kσ;

[0034] Among them, the k value is the adjustment parameter;

[0035] 4) Check whether the power of the new data point exceeds the range of the dynamic energy consumption threshold. If three consecutive points exceed the range of the dynamic energy consumption threshold, an alarm is triggered;

[0036] 5) After receiving m new data points, the window moves forward and removes the earliest m data points; the new m data points will enter the window, and the moving average and standard deviation σ will be recalculated to adaptively update the range of the dynamic energy consumption threshold.

[0037] In a third aspect, the present invention provides a smart street lamp energy-saving control system, which includes a number of light control subsystems and a street lamp management platform. Each light control subsystem controls a number of light control modules, and each light control module controls one street lamp. The main control module is not installed on a specific street lamp. The street lamp management platform includes a cloud platform server, a mobile terminal, and a web terminal.

[0038] Each main control module includes a main control unit, a distance measurement module m2 for identifying vehicles, a photosensitive module m3 for detecting light intensity, a time perception module m4 for obtaining date and time, a photovoltaic panel m5, a controller m6 for managing battery charge and discharge, a battery m7, a remote transmission transceiver module m8, and a rainfall and temperature monitoring module m9. The main control unit m1 obtains data from the distance measurement module m2, the photosensitive module m3, the time perception module m4, and the rainfall and temperature monitoring module m9, and communicates with the cloud platform server through the remote transmission transceiver module m8. The cloud platform server performs data exchange and recording, energy consumption monitoring, and fault alarm. The mobile phone terminal and the web terminal perform two-way communication with the cloud platform server.

[0039] Each lighting control module includes a lighting control unit mn-1, a lighting control sensing module mn-2 for real-time detection of pedestrians around the street light, a driver module mn-3 for dimming the street light, and a fault detection module mn-4.

[0040] After receiving the control information transmitted to the main control unit m1 via wireless communication, the light control unit mn-1 controls the operation mode of the street lamp by controlling the driving module mn-3 according to the set street lamp operation mode. The street lamp operation modes include energy-saving operation mode, independent operation mode and full lighting operation mode, ultimately achieving the control purpose of energy saving.

[0041] Furthermore, the energy-saving operation mode includes the operation mode during the morning and evening peak hours and the operation mode during the non-morning and evening peak hours.

[0042] The operation mode during the morning and evening peak hours is as follows: the main control unit m1 reads the date and time in real time through the time sensing module m4. If the detection is between 6:00-9:00 and 16:00-19:00 on weekdays, and the light intensity is lower than the threshold value detected in real time by the photosensitive module m3, the morning and evening peak hours operation mode will be entered. That is, the main control unit m1 sends a control instruction to each lighting control unit to put each street lamp into high-brightness lighting mode.

[0043] If the detection is not during the working hours of 6:00-9:00 and 16:00-19:00, and the light intensity is lower than the threshold value detected by the photosensitive module m3 in real time, the main control unit m1 will enter the non-peak period operation mode. The non-peak period operation mode includes intelligent control and time control.

[0044] In intelligent control, the street lights are put into low-brightness lighting mode and simultaneously detect in real time whether there are vehicles passing within the sensing range. The vehicle detection function is activated. When a vehicle is identified by the vehicle recognition algorithm, the street lights on the road section controlled by the current light control subsystem will increase the lighting brightness in sequence and enter high-brightness lighting mode. If no vehicle is detected within the set time, the street lights on the road section controlled by the light control subsystem will gradually reduce the brightness and return to low-brightness lighting mode.

[0045] The time control process is as follows: first, the local latitude and longitude of the street lamp is input through the street lamp management platform, and the energy-saving on and off time is set. The system will calculate the sunset time of the day based on the historical curve of the sun's altitude. From sunset time to the energy-saving on time, the street lamp operates in high-brightness lighting mode. From the energy-saving on time point to the energy-saving off time, the street lamp operates in low-brightness lighting mode.

[0046] Furthermore, the vehicle identification algorithm adopts the above-mentioned vehicle identification method.

[0047] Furthermore, the independent operation mode always exists when the street light is operating normally and does not need to be set separately. When a pedestrian is captured, the street light operates in a high-brightness lighting mode;

[0048] The full lighting operation mode causes the street lamp to enter a high-brightness lighting mode. In this operation mode, energy-saving control of the street lamp cannot be achieved.

[0049] Furthermore, the fault detection module detects the working status and energy consumption data of each street lamp in real time. When the street lamp is in the lighting state, the fault detection module starts the fault detection link and collects the street lamp energy consumption data in real time. The street lamp fault judgment method is loaded in the lighting control unit to judge whether a fault occurs. If a fault occurs, the lighting control unit transmits the street lamp alarm information to the main control unit, where the alarm information includes the road section name, road section number and street lamp number; the main control unit then transmits the alarm information to the street lamp management platform. After receiving the information, the street lamp management platform pushes the information to the mobile phone or Web terminal in real time. After receiving the information, the operation and maintenance personnel arrange maintenance in time.

[0050] Furthermore, the street light management platform can display the operating information of each street light in real time, including the street light switch status, fault status, and energy consumption information; it can also remotely set the working status of the lighting control subsystem, including longitude and latitude, date and time, weekday time period, and the operating time of the street light high-brightness lighting mode.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. In rainy and snowy weather, the point clouds collected by vehicle detection sensors such as lidar are scattered by rain and snow particles, generating a large number of noise points, which may lead to problems such as failure to identify vehicles. The present invention adopts an improved Euclidean clustering algorithm to identify high-speed vehicles. Through four stages of point cloud preprocessing, improved Euclidean clustering, vehicle verification, and time series tracking, it distinguishes between static snow and dynamic vehicles, thereby reducing the false detection rate caused by rain and snow noise.

[0053] 2. A segmented street light control scheme is adopted to divide a certain road section into multiple lighting control subsystems. Each lighting control subsystem includes a main control module, which contains a variety of sensors, communication modules and other equipment, thereby controlling multiple street lights in the lighting control subsystem to achieve energy saving. This control scheme does not require the installation of hardware equipment such as detection sensors on each street light, greatly reducing the cost of street light renovation. In addition, the main control module in each lighting control subsystem is solar-powered. When renovating the street lights, there is no need to modify the original underground street light power cables. Wireless communication is used between the main control module and the lighting control modules of the lighting control subsystem, and there is no need to lay separate wired lines. This makes street light renovation simpler, thereby greatly reducing the cost of street light renovation.

[0054] 3. A dynamic threshold adjustment mechanism enables real-time monitoring of streetlight faults. Alarm thresholds are dynamically adjusted based on historical data to prevent fixed thresholds from being unable to adapt to load changes, thereby reducing false alarms caused by grid interference. When a streetlight fault is detected, the main control unit transmits the fault status via the remote transmission transceiver module to the streetlight management platform. Operations and maintenance personnel can view the streetlight's operating status in real time through the streetlight management platform and address the fault promptly, reducing the workload of manual inspections and lowering streetlight maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a schematic diagram of the system structure block diagram of the smart street lamp energy-saving control system.

[0056] Figure 2 This is the hardware block diagram of the lighting control subsystem m.

[0057] Figure 3 This is a schematic diagram of the layout of the lighting control subsystem m. DETAILED DESCRIPTION

[0058] The present invention is described in detail below with reference to the embodiments and drawings, but they are not intended to limit the scope of protection of the present application.

[0059] Figure 1 This is the system block diagram of the smart street light energy-saving control system. The smart street light energy-saving control system includes a number of light control subsystems and a street light management platform. The mth light control subsystem is denoted as light control subsystem m. Each light control subsystem controls a number of light control modules. One light control module controls one street light (e.g. Figure 3 As shown), the nth lighting control module in the mth lighting control subsystem is recorded as lighting control module mn.

[0060] The lighting control subsystem m comprises a master control module m and a number of lighting control modules mn. m represents the number of lighting control subsystems in a particular road section and can be set based on the total length of the road section. m can be 1, 2, 3, etc.; n represents the number of lighting control modules (i.e., the number of streetlights) included in the lighting control subsystem m and can be set based on the road environment. n can be 1, 2, 3, etc. The master control module is not attached to a specific streetlight; instead, one master control module can control a number of streetlights, resolving the high cost of traditional streetlights requiring a master control module, significantly reducing costs.

[0061] The street light management platform is composed of a cloud platform server, a mobile phone terminal and a Web terminal. The cloud platform server and the remote transmission and reception module m8 in the main control module m of the lighting control subsystem m realize data interaction, data recording, energy consumption monitoring, fault alarm and other functions. The mobile phone terminal and the Web terminal serve as terminal devices for data monitoring and control. The two access the cloud platform server through wireless or wired communication to realize the monitoring and control functions of the street lights.

[0062] Figure 2 This is the hardware block diagram of the lighting control subsystem m. The lighting control subsystem m includes a main control module m and several lighting control modules mn. The main control module m includes a main control unit m1, a distance measurement module m2, a photosensor module m3, a time sensing module m4, a photovoltaic panel m5, a controller m6, a battery m7, a remote transmission and reception module m8, and a rainfall and temperature monitoring module m9.

[0063] The light control module mn includes a light control unit mn-1, a light control sensing module mn-2, a driving module mn-3 and a fault detection module mn-4.

[0064] The main control unit m1 takes EFR32 as an example. EFR32 wireless SoC is built for real low-power applications. The core, its network characteristics comply with the ZIGBEE 3.0 technical standard, and provide a complete application integration solution based on the IEEE 802.15.4 standard ISM band. Support Low energy technology (BLE), Thread and proprietary wireless connections, support serial port transparent transmission mode, integrate quick and easy-to-use self-organizing network functions, and provide multiple configurable AD, IO, and PWM interfaces.

[0065] The distance measuring module m2, photosensor module m3, and time sensing module m4 are exemplified here by a laser distance measuring sensor, photosensor, and clock module. The distance measuring module m2 is used to identify high-speed vehicles, the photosensor module m3 is used to detect light intensity, and the time sensing module m4 is used to obtain the date and time. The photovoltaic panel m5 is a solar power generation module, and the controller m6 is a charge and discharge management module for the battery m7. The battery m7 is the power supply module for the main control module m. Using solar power, the existing underground streetlight power supply circuit does not need to be modified during streetlight renovation, simplifying the renovation and reducing the cost. The main control unit m1 calculates and processes the information collected by the distance measuring module m2, photosensor module m3, and time sensing module m4, performs corresponding logical control according to the set streetlight operating mode, and wirelessly transmits the control instructions to the light control unit of the light control module. The present invention also allows different sensing components to be selected as expansion modules according to different needs. The expansion modules communicate with the main control unit to obtain the relevant information collected by the expansion modules.

[0066] After receiving the control information transmitted to the main control unit M1 via wireless communication, the lighting control unit Mn-1 controls the operation mode of the street light by controlling the driver module Mn-3 according to the set street light operation mode. The street light operation mode includes energy-saving operation mode, independent operation mode and full lighting operation mode, ultimately achieving the control purpose of energy saving.

[0067] The driver module mn-3 is the power driver module of the street lamp, which can adjust the brightness of the street lamp in real time and realize the intelligent dimming function of the street lamp.

[0068] The fault detection module mn-4 is a street light fault detection unit. It uses a dynamic threshold adjustment mechanism to achieve real-time monitoring of street light faults. When a street light fault is detected, the fault signal is transmitted to the light control unit mn-1. The light control unit mn-1 transmits the fault information to the main control unit m1 via wireless communication. The main control unit m1 transmits the fault information to the street light management platform via the remote transmission transceiver module m8. After receiving the information, the street light management platform pushes the information to the mobile phone or web terminal in real time. After receiving the information, the operator will arrange for maintenance personnel to repair it in time.

[0069] The light control sensing module mn-2 detects the presence of pedestrians around the street lamp in real time. When a pedestrian is detected and the light intensity is below a set threshold (the threshold is configurable), the light control unit mn-1 controls the driver module mn-3 to operate the street lamp in high-brightness lighting mode. When no pedestrians are detected, the street lamp will continue to operate in energy-saving mode to prevent the street lamp from being in low-light illumination when pedestrians stop under the street lamp.

[0070] The ranging module m2 takes the RPLIDARA3 laser radar as an example. It is suitable for indoor and outdoor environment measurements, has a short response time, a measurement distance of up to 25m, a small size, and a sampling frequency of up to 16,000 times / S. In outdoor mode, RPLIDARA3 has a more reliable anti-sunlight interference capability and can achieve 360-degree all-round scanning and ranging detection of the surrounding environment. It complies with Class 1 laser safety standards and reaches the human eye safety level.

[0071] The photosensitive module m3 takes a 5516 photoresistor as an example, which has high sensitivity, fast response speed, high temperature resistance, an ambient temperature range of -30°C to +70°C, a bright resistance of 5-10kΩ, and a dark resistance of 0.5MΩ.

[0072] The time sensing module m4 uses the DS1302 clock as an example. The DS1302 is a low-power real-time clock chip from Dell Technologies that features trickle-current charging. It can count years, months, days, weeks, hours, minutes, and seconds, and includes features such as leap year compensation. It also contains a 31×8 RAM register for temporary data storage. It communicates with the main control unit m1 via a simple serial interface, storing the current time in RAM. The clock can operate in either a 12-hour or 24-hour format.

[0073] The rainfall and temperature monitoring module m9 is a combination module of a rain gauge and a thermometer. The rain gauge adopts an optical tipping bucket rain sensor, which is small in size, highly reliable, and easy to maintain. It can measure rainfall data in real time, with a maximum instantaneous rainfall of 24 mm / min and a resolution of 0.1 mm. The thermometer has a measurement range of -40°C to 85°C and a measurement accuracy of ±0.3°C.

[0074] Taking the FarEasTone transceiver module m8 as an example, E840-DTU is a new generation of high-performance 4G DTU with wide network coverage, low transmission delay, and support for Cat.1 network access of the three major operators. It also supports multiple configuration methods, supports online management of web platforms and mini-programs, and supports TCP / UDP / MQTT / HTTP protocols.

[0075] The lighting control unit mn-1 and the main control unit m1 use the same hardware EFR32 chip. The lighting control unit mn-1 and the main control unit m1 communicate wirelessly to achieve data interaction, mainly including but not limited to the following functions: 1) uploading the operating status, fault status and energy consumption data of the street light; 2) receiving street light control instructions from the main control unit m1.

[0076] The HC-SR501 is used as an example of the light control sensor module mn-2. The HC-SR501 is a human body sensing module. This automatic control module uses infrared technology and features high sensitivity, strong reliability, ultra-low voltage operation, a sensing angle of less than 100 degrees, temperature compensation, and two selectable triggering modes. When a human body is detected, the output voltage changes between high and low levels, thereby determining the presence of pedestrians near the streetlight and controlling the light's output status.

[0077] The fault detection module mn-4 is a street lamp energy consumption sampling module, which can transmit the energy consumption information of the street lamp to the light control unit mn-1 as a basis for street lamp energy consumption statistics and fault judgment.

[0078] The driving module mn-3 is an energy regulation module, which realizes energy conversion and regulation through power electronic devices, and adjusts the brightness of the street lamp in real time, thereby achieving energy-saving control of the street lamp.

[0079] The control method of the smart street lamp energy-saving control system includes the following steps:

[0080] First, a highway is divided into m sections. The distance of each section can be set freely, but the distance of each section is not more than 1.3 km. A lighting control subsystem m is installed in each section. Each lighting control subsystem is independent of each other. The layout diagram of lighting control subsystem m is as follows: Figure 3 The lighting control subsystem m includes three operating modes: energy-saving operation, independent operation, and full lighting operation. At the same time, each operating mode can adopt a corresponding high-brightness lighting mode or low-brightness lighting mode. The control process of the three operating modes is detailed below:

[0081] 1. Energy-saving operation mode

[0082] The system can be set to energy-saving operation mode through the mobile or web version of the street light management platform. The energy-saving operation mode includes operation during peak hours in the morning and evening and during non-peak hours in the morning and evening.

[0083] 1) Operation mode during morning and evening peak hours:

[0084] Due to the large volume of vehicles and pedestrians during peak hours in the morning and evening, in order to avoid the frequent dimming of street lights, this system is designed to operate during peak hours in the morning and evening. That is, the main control unit m1 reads the date and time in real time through the time perception module m4. When the detection is between 6:00-9:00 and 16:00-19:00 on weekdays (the date and time can be set), and the light intensity is lower than the threshold (that is, a dark environment, where the threshold can be set) in real time through the photosensitive module m3, if both are met at the same time, the system will enter the peak hours operation mode, that is, the main control unit m1 sends a control instruction to each lighting control unit mn-1, so that each street light is in high-brightness lighting mode.

[0085] 2) Operation mode during non-peak hours:

[0086] The main control unit m1 detects the light intensity in real time through the photosensitive module m3. When the light intensity is lower than the preset light intensity threshold (i.e., dark environment, where the threshold is configurable, and the photosensitive module m3 can determine whether it is daytime or nighttime? Or sunny or cloudy?), and it is not during the period of 6:00-9:00 and 16:00-19:00 on weekdays (date and time can be set), the main control unit m1 will enter the non-peak period operation mode. The specific steps are as follows:

[0087] The operation modes during non-peak hours include intelligent control and time control, which can be set through the street light management platform.

[0088] The intelligent control process is as follows: the main control unit m1 detects the light intensity in real time through the photosensitive module m3. When the light intensity is lower than the light intensity threshold (i.e., a dark environment, where the threshold can be set), the main control unit m1 transmits the control instruction to the light control unit mn-1 through wireless communication. After receiving the control instruction, the light control unit mn-1 controls the drive module mn-3 to make the street light enter the low-brightness lighting mode. At the same time, the ranging module m2 detects in real time whether there are vehicles passing within the sensing range, and activates the vehicle detection function.

[0089] When a vehicle is identified by the vehicle recognition algorithm, the street lights on the road section controlled by the current lighting control subsystem m will increase the lighting brightness in sequence, and the street lights will enter the high-brightness lighting mode. If no vehicle is detected within the set time, the street lights on the road section controlled by the lighting control subsystem m will gradually reduce the brightness and return to the low-brightness lighting mode.

[0090] The time control process is as follows: first, input the local latitude and longitude of the street light through the street light management platform, and set the energy-saving on and off time. The system will calculate the sunset time of the day based on the historical curve of the sun's altitude. From sunset time to the energy-saving on time, the street light will operate in high-brightness lighting mode. From the energy-saving on time point to the energy-saving off time, the street light will operate in low-brightness lighting mode.

[0091] 2. Independent operation mode

[0092] The independent operation mode of street lights does not require configuration from the street light management platform. When the system is running, the lighting control unit mn-1 uses the lighting control sensing module mn-2 to detect the presence of pedestrians in the vicinity of the street light in real time. When a pedestrian is detected and the light intensity is below a threshold (i.e., a dark environment, where the threshold is configurable), the street light corresponding to the lighting control unit mn-1 operates in high-brightness lighting mode. When no pedestrians are detected, the street light will continue to operate in energy-saving mode to prevent the street light from being in low illumination when pedestrians stop under the street light. Independent operation mode is always in place during normal operation of the street light and does not require separate configuration. When a pedestrian is detected, the street light illuminates at high brightness to prevent the street light from being in low illumination when someone lingers near the street light.

[0093] 3. Full lighting operation mode

[0094] The street light management platform can be used to set the street light to full lighting mode. The main control unit m1 reads the light intensity in real time through the photosensitive module m3. When the light intensity is lower than the threshold (i.e., a dark environment, where the threshold is configurable), the main control unit m1 transmits the control command to the light control unit mn-1 via wireless communication. After receiving the control command, the light control unit mn-1 controls the drive module mn-3 to make the street light enter high-brightness lighting mode. This operating mode cannot achieve energy-saving control of the street light.

[0095] Secondly, the lighting control unit mn-1 uses the fault detection module mn-4 to detect the working status and energy consumption data of each street lamp in real time. The specific steps are as follows:

[0096] 1. When the street lamp is in the lighting state, the fault detection module mn-4 starts the fault detection link, collects the street lamp energy consumption data in real time, and transmits it to the lighting control unit mn-1;

[0097] 2. Light control unit MN-1 first processes and stores the street lamp energy consumption data, compares it with the dynamic energy consumption threshold, and performs fault diagnosis to determine whether the current energy consumption data is within the dynamic energy consumption threshold range. If not, light control unit MN-1 transmits an alarm message to main control unit M1 via wireless communication. The alarm message includes the road section name, road section number, street lamp number, etc. The dynamic energy consumption threshold is an adaptive baseline established by historical data to avoid false alarms caused by the fixed threshold being unable to adapt to load changes.

[0098] 3. After receiving the street light alarm information, the main control unit m1 transmits the alarm information to the remote transmission transceiver module m8, and uploads it to the street light management platform via wireless communication. After receiving the information, the street light management platform pushes the information to the mobile phone or Web terminal in real time. After receiving the information, the operation and maintenance personnel arrange maintenance in time.

[0099] Finally, the street light management platform can display the operating information of each street light in real time through mobile phones or web terminals via wireless communication, such as street light switch status, fault status, energy consumption information, etc.; it can also remotely set the working status of the lighting control subsystem m, such as latitude and longitude, date and time, weekday time period, and the operating time of the street light high-brightness lighting mode.

[0100] A method for determining street lamp faults, the specific process is:

[0101] Step 1: Data collection and processing

[0102] 1) Collect electrical data such as voltage and current of street lamps in chronological order;

[0103] 2) Eliminate abnormal data based on the gross error judgment criteria.

[0104] Step 2: Dynamic baseline calculation

[0105] 1) Using a sliding window (the window size is the number of days of historical data), the moving average of the power within the window is calculated based on the collected voltage and current data, i.e., the baseline value, and the standard deviation σ of the power data within the window is calculated at the same time;

[0106] 2) Set the range of dynamic energy consumption threshold:

[0107] The upper limit of the dynamic threshold of energy consumption = moving average + kσ (k is an adjustment parameter that can be adjusted according to actual conditions, and σ is the standard deviation of the power data in the window);

[0108] The lower limit of the energy consumption dynamic threshold = moving average - kσ;

[0109] Step 3: Fault detection and judgment

[0110] 1) Check whether the power of the new data point exceeds the range of the dynamic energy consumption threshold;

[0111] 2) If three consecutive points exceed the energy consumption dynamic threshold, an alarm will be triggered;

[0112] Step 4: Adaptive update of dynamic energy consumption threshold

[0113] 1) After receiving m new data points, the window moves forward and removes the earliest m data points;

[0114] 2) The new m data points will enter the window, and the moving average and standard deviation σ will be recalculated to adaptively update the range of the dynamic energy consumption threshold.

[0115] To address the problem of vehicle sensor failure in existing streetlight energy-saving systems during rainy and snowy weather, this paper proposes a vehicle recognition method based on an improved Euclidean clustering algorithm to identify high-speed vehicles. The method consists of four stages: point cloud preprocessing, improved Euclidean clustering, vehicle verification, and time series tracking. The specific steps are as follows:

[0116] Step 1: Point cloud preprocessing

[0117] 1. The ranging module m2 collects the ranging point cloud P: P = {(r i ,θ i ,I i )}, (distance r i , angle θ i , Strength I i );

[0118] 2. Coordinate transformation:

[0119] x i =r i cosθ i

[0120] y i =r i sinθ i

[0121] Among them, (x i ,y i ) is the coordinate of the point cloud in the rectangular coordinate system;

[0122] 3. Dynamic intensity filtering: Set the intensity filter threshold I according to the weather intensity coefficient as follows: th , if the point cloud intensity I i If the intensity is less than the intensity filter threshold, it will be eliminated.

[0123] I th =40+60ω(weather intensity coefficient ω∈[0,1])

[0124] Such as I i ≥I th , retain, otherwise, remove

[0125] 4. Sliding window denoising: for 5 consecutive points {p i-2 , p i-1 , p i , p i+1 , p i+2 If the distance between two adjacent points is greater than 0.5m, the point in the middle is considered noise and is removed. That is, if the following conditions are met:

[0126] |p i -p i-1 |>0.5m and |pi -p i+1 |>0.5m

[0127] Then p i Determine it as noise and remove it;

[0128] The original point cloud is obtained by the above dynamic intensity filtering and sliding window denoising.

[0129] Step 2: Improved Euclidean clustering

[0130] 1. Dynamic parameter calculation:

[0131] 1) Dynamic distance threshold formula:

[0132] Neighborhood search radius d for Euclidean clustering th Dynamic adjustment with weather:

[0133] d th =d base +k d ·ω

[0134] Where: d base : Sunny day basic threshold, k d : distance adjustment coefficient, ω: weather intensity coefficient ω∈[0,1] (0 is sunny, 1 is extreme rain and snow).

[0135] The ω value is calculated in real time using the following formula:

[0136]

[0137] Where, precipitation intensity is obtained in real time by the rainfall and temperature monitoring module m9, γ is a nonlinear adjustment factor, and the benchmark threshold is calculated by the following formula, where 30 is the heavy rain benchmark threshold and 5 is the heavy snow benchmark threshold.

[0138]

[0139] 2) Dynamic minimum number of clustering points N min formula:

[0140] N min =N base +k n ·ω

[0141] Where: N base : Minimum number of points on sunny days; k n : Point adjustment coefficient

[0142] 2. Construct KD tree: for the denoised point cloud Build a KD tree to speed up neighborhood search.

[0143] 3. Region growing clustering:

[0144] 1) Based on the KD tree, the set U does not contain all points when initialized, that is, By gradually removing the visited points, traversing the unvisited points U, any p∈U is taken as the seed point, with p as the center, d th is the radius, and the search neighborhood point N(p) is selected.

[0145] 2) If the neighborhood point N(p)≥N min , then a new cluster C is generated j .

[0146] 3) Iteratively expand the cluster until no new points are added.

[0147] Step 3: Vehicle Verification

[0148] For each cluster C j , calculate its minimum bounding rectangle and retain the clusters that meet the vehicle size constraints:

[0149] Minimum bounding rectangle (MBR) calculation:

[0150]

[0151] Among them, (x i ,y i ) is the coordinate of the i-th point cloud in the rectangular coordinate system;

[0152] Step 4: Timing Tracking

[0153] Use Kalman filtering to predict and update the vehicle position, associate the detection results of consecutive frames, and output the vehicle status.

[0154]

[0155] Where: is the state vector, P t is the covariance matrix, K t is the Kalman gain; u t is the output of Kalman filter; H t is the observation matrix; Q t is the process noise covariance matrix; x t-1 is the posterior state estimation vector at time t-1; B t is the control input matrix; F t is the state transition matrix.

[0156] When a vehicle is identified by the above-mentioned vehicle recognition method, the street lights on the road section controlled by the lighting control subsystem m will increase the lighting brightness in sequence, and the street lights will enter the high-brightness lighting mode. If no vehicle is detected within the set time, the street lights on the road section controlled by the lighting control subsystem m will gradually reduce the brightness and return to the low-brightness lighting mode.

[0157] Any matters not described in the present invention are applicable to the prior art.

Claims

1. A vehicle recognition method based on an improved Euclidean clustering algorithm is used to identify high-speed vehicles. It includes four stages: point cloud preprocessing, improved Euclidean clustering, vehicle verification, and time series tracking. The characteristics are: The vehicle identification method includes the following contents: Get the current precipitation intensity and define the weather intensity coefficient ω to be calculated in real time using the following formula: Where γ is the nonlinear adjustment factor; the benchmark threshold is calculated by the following formula, where 30 is the heavy rain benchmark threshold and 5 is the heavy snow benchmark threshold; Dynamically update the intensity filter threshold I according to the weather intensity coefficient th , with the dynamically updated intensity filter threshold I th The original point cloud data of the collected vehicle is denoised by two methods: using the sliding window denoising method and the sliding window denoising method to obtain the denoised point cloud data. Neighborhood search radius d th Dynamic adjustment with weather: d th =d base +k d ·oh Where: d base : sunny basic threshold; k d : distance adjustment coefficient; ω: weather intensity coefficient, ω∈[0,1], 0 for sunny days and 1 for extreme rain and snow; Dynamic minimum number of cluster points N min Dynamic adjustment with weather: N min =N base +k n ·oh Where: N base : Minimum number of points on sunny days; k n : Point adjustment coefficient; The denoised point cloud Establish a KD tree and search the neighborhood with radius d th and the dynamic minimum number of cluster points N min Perform region growing and clustering for constraints; Clusters that satisfy the vehicle size constraint are retained, and the vehicle position is predicted and updated using Kalman filtering, correlating the detection results of consecutive frames and outputting the vehicle state.

2. The vehicle identification method according to claim 1, characterized in that: The intensity filtering threshold I th The relationship between I and the weather intensity coefficient ω is: th =40+60ω; If the point cloud intensity I i Less than the intensity filtering threshold I th , then remove it.

3. The vehicle identification method according to claim 1, characterized in that: The process of sliding window denoising is as follows: for 5 consecutive points {p i-2 , p i-1 , p i , p i+1 , p i+2 If the distance between two adjacent points is greater than 0.5m, the point in the middle is considered noise and is removed. That is, if the following conditions are met: |p i -p i-1 |>0.5m and |p i -p i+1 |>0.5m Then p i It is determined to be noise and removed.

4. A method for determining a street lamp fault, characterized in that: The determination method comprises the following steps: 1) Collect the voltage and current data of street lamps in chronological order and eliminate abnormal data based on the gross error judgment criterion; 2) Set a sliding window and calculate the moving average of the power within the window based on the collected voltage and current data, that is, the baseline value, and simultaneously calculate the standard deviation σ of the power data within the window; 3) Set the range of dynamic energy consumption threshold based on the moving average and standard deviation: The upper limit of the dynamic threshold of energy consumption = moving average + kσ; The lower limit of the energy consumption dynamic threshold = moving average - kσ; Among them, the k value is the adjustment parameter; 4) Check whether the power of the new data point exceeds the range of the dynamic energy consumption threshold. If three consecutive points exceed the range of the dynamic energy consumption threshold, an alarm is triggered; 5) After receiving m new data points, the window moves forward and removes the earliest m data points; the new m data points will enter the window, and the moving average and standard deviation σ will be recalculated to adaptively update the range of the dynamic energy consumption threshold.

5. A smart street lamp energy-saving control system, characterized in that: The smart street lamp energy-saving control system includes a number of light control subsystems and a street lamp management platform. Each light control subsystem controls a number of light control modules, and one light control module controls one street lamp. The main control module is not set on a specific street lamp. The street lamp management platform includes a cloud platform server, a mobile terminal and a web terminal. Each main control module includes a main control unit, a distance measurement module m2 for identifying vehicles, a photosensitive module m3 for detecting light intensity, a time perception module m4 for obtaining date and time, a photovoltaic panel m5, a controller m6 for managing battery charge and discharge, a battery m7, a remote transmission transceiver module m8, and a rainfall and temperature monitoring module m9. The main control unit m1 obtains data from the distance measurement module m2, the photosensitive module m3, the time perception module m4, and the rainfall and temperature monitoring module m9, and communicates with the cloud platform server through the remote transmission transceiver module m8. The cloud platform server performs data exchange and recording, energy consumption monitoring, and fault alarm. The mobile phone terminal and the web terminal perform two-way communication with the cloud platform server. Each lighting control module includes a lighting control unit mn-1, a lighting control sensing module mn-2 for real-time detection of pedestrians around the street light, a driver module mn-3 for dimming the street light, and a fault detection module mn-4. After receiving the control information transmitted to the main control unit m1 via wireless communication, the light control unit mn-1 controls the operation mode of the street lamp by controlling the driving module mn-3 according to the set street lamp operation mode. The street lamp operation modes include energy-saving operation mode, independent operation mode and full lighting operation mode, ultimately achieving the control purpose of energy saving.

6. The control system according to claim 5, characterized in that: The energy-saving operation mode includes the operation mode during the morning and evening peak hours and the operation mode during the non-morning and evening peak hours. The operation mode during the morning and evening peak hours is as follows: the main control unit m1 reads the date and time in real time through the time sensing module m4. If the detection is between 6:00-9:00 and 16:00-19:00 on weekdays, and the light intensity is lower than the threshold value detected in real time by the photosensitive module m3, the morning and evening peak hours operation mode will be entered. That is, the main control unit m1 sends a control instruction to each lighting control unit to put each street lamp into high-brightness lighting mode. If the detection is not during the working hours of 6:00-9:00 and 16:00-19:00, and the light intensity is lower than the threshold value detected by the photosensitive module m3 in real time, the main control unit m1 will enter the non-peak period operation mode. The non-peak period operation mode includes intelligent control and time control. In intelligent control, the street lights are put into low-brightness lighting mode and simultaneously detect in real time whether there are vehicles passing within the sensing range. The vehicle detection function is activated. When a vehicle is identified by the vehicle recognition algorithm, the street lights on the road section controlled by the current light control subsystem will increase the lighting brightness in sequence and enter high-brightness lighting mode. If no vehicle is detected within the set time, the street lights on the road section controlled by the light control subsystem will gradually reduce the brightness and return to low-brightness lighting mode. The time control process is as follows: first, the local latitude and longitude of the street lamp is input through the street lamp management platform, and the energy-saving on and off time is set. The system will calculate the sunset time of the day based on the historical curve of the sun's altitude. From sunset time to the energy-saving on time, the street lamp operates in high-brightness lighting mode. From the energy-saving on time point to the energy-saving off time, the street lamp operates in low-brightness lighting mode.

7. The control system according to claim 6, characterized in that: The vehicle identification algorithm adopts the vehicle identification method described in any one of claims 1-3.

8. The control system according to claim 5, characterized in that: The independent operation mode always exists when the street light is operating normally and does not need to be set separately. When a pedestrian is detected, the street light will operate in high-brightness lighting mode; The full lighting operation mode causes the street lamp to enter a high-brightness lighting mode. In this operation mode, energy-saving control of the street lamp cannot be achieved.

9. The control system according to claim 5, characterized in that: The fault detection module detects the working status and energy consumption data of each street lamp in real time. When the street lamp is in the lighting state, the fault detection module starts the fault detection link and collects the street lamp energy consumption data in real time. The street lamp fault judgment method according to claim 4 is loaded into the light control unit to judge whether a fault occurs. If a fault occurs, the light control unit transmits the street lamp alarm information to the main control unit, wherein the alarm information includes the road section name, road section number and street lamp number; The main control unit then transmits the alarm information to the street light management platform. After receiving the information, the street light management platform pushes the information to the mobile phone or web terminal in real time. After receiving the information, the operation and maintenance personnel arrange maintenance in a timely manner.

10. The control system according to claim 6, characterized in that: The streetlight management platform can display the operating information of each streetlight in real time, including the streetlight switch status, fault status, and energy consumption information; it can also remotely set the working status of the lighting control subsystem, including latitude and longitude, date and time, weekday time period, and the operating time of the streetlight high-brightness lighting mode.

Images