High-pole lamp lighting intelligent control system and method

Abstract

The application provides a high-pole lamp lighting intelligent control system and method, which comprises an intelligent lighting control cloud box and a lighting control management system, and the intelligent lighting control cloud box is in wireless communication connection with the lighting control management system. The application can realize intelligent monitoring control and fine management of high-pole lamp lighting by operating a lighting management platform, a mobile phone APP and the intelligent lighting control cloud box to send instructions for separately controlling, batch controlling and plan controlling the on-site lamps, and also can control the on-site lamps by time period and area. Meanwhile, the application can realize functions such as a map function, equipment inspection, lighting control, power over-reading, gateway inspection, equipment time correction, electricity price clock, batch parameter adjustment, energy efficiency management, alarm and the like, and traffic flow monitoring, safety monitoring, light intensity and coverage control functions based on image recognition and neural network technology, which has important significance for improving lighting efficiency, reducing energy consumption and realizing intelligent management of urban traffic.

Description

Technical Field

[0001] The present invention relates to the field of intelligent lighting technology, specifically to an intelligent control system and method for high pole lamp lighting. Background Art

[0002] With global energy issues becoming increasingly prominent, traditional lighting control systems are no longer able to meet the demands of refined management and energy conservation and emission reduction. Lighting energy-saving control is a key area of ​​energy conservation. The most common approach is to use a large number of lamps on high poles for extended periods of time. This approach often results in fixed brightness and height settings, which presents numerous challenges.

[0003] The switch control method of lighting equipment is single, the equipment control parameters are difficult to modify according to actual conditions, the operating parameters of the lighting equipment such as voltage, current, and electric energy cannot be monitored, lighting failures cannot be promptly alarmed and handled, and the operating data of lighting equipment cannot be centrally statistically managed.

[0004] On the one hand, the lighting intensity cannot be adjusted, and the illumination and coverage cannot be adjusted according to the number of vehicles and personnel or the operating time. On the other hand, the lighting cannot be realized according to the actual situation of vehicles and pedestrians. In other words, it cannot realize dynamic intelligent lighting control based on the changing situation of vehicles and pedestrians, turning on the lights when people come and turning off when they leave. It is also impossible to realize dynamic intelligent lighting control based on the distribution of vehicles and pedestrians in different areas of the square. These defects not only easily cause power loss and light pollution, but also have a direct impact on the service life of the lamps, increasing various maintenance and upkeep costs.

[0005] High pole lamps are located in industrial operation areas and various traffic roads in cities, and their unique location is not fully utilized to play their other functions besides lighting, such as traffic flow monitoring and safety monitoring.

[0006] In existing technologies, this fixed lighting model, with its limited flexibility and high lighting power, wastes significant amounts of electricity and public resources. Furthermore, this fixed lighting model leads to serious light pollution, which not only harms people's health but also has irreversible impacts on the ecological environment. Therefore, developing an intelligent control system and method for high-mast lamps with a rational design, ease of use, and simple structure is of great significance for improving lighting efficiency, reducing energy consumption, and achieving intelligent lighting management. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a high pole lamp lighting intelligent control system and method and control method thereof, so as to solve the deficiencies of the prior art mentioned in the above background technology.

[0008] To solve the above technical problems, the embodiments of the present invention provide the following technical solutions:

[0009] A high-pole lamp lighting intelligent control system includes multiple smart lighting cloud boxes, a lighting management cloud platform, a lighting control background, a data monitoring screen, a mobile terminal and a database server. The multiple smart lighting cloud boxes are electrically connected to the corresponding circuit loops of multiple lamps respectively. The smart lighting cloud boxes are wirelessly connected to the lighting management cloud platform. The lighting management cloud platform is communicatively connected to the lighting control background, the data monitoring screen and the database server. The lighting management cloud platform is wirelessly connected to the mobile terminal.

[0010] The smart lighting cloud box includes a cloud box control mainboard and a liquid crystal display electrically connected to the cloud box control mainboard, a GPS and 4G / 5G Internet of Things integrated module, a power module, a high-power relay integrated module, a local on / off light button, a local / remote selection button, a 4-position aviation plug socket and an image acquisition module, wherein the GPS and 4G / 5G Internet of Things integrated module is wirelessly connected to the lighting management cloud platform;

[0011] The power supply module is electrically connected to the leakage protection switch, the GPS and 4G / 5G Internet of Things integrated module is electrically connected to the GPS positioning antenna interface and the 4G / 5G Internet of Things antenna interface, and the high-power relay integrated module is electrically connected to the input / output terminal block;

[0012] Among them, the lighting management cloud platform and lighting control background are servers or industrial computers.

[0013] The present invention also provides a control method for the above-mentioned high pole lamp lighting intelligent control system, comprising:

[0014] The real-time operating parameters of the circuits monitored by the monitoring modules installed on the corresponding circuit loops of multiple lamps are sent to the cloud box control mainboard of the intelligent lighting cloud box. The cloud box control mainboard wirelessly sends the real-time operating parameter data to the lighting management cloud platform through GPS and 4G / 5G IoT integrated modules. The lighting management cloud platform then forwards the real-time operating parameter data to the lighting control backend, data monitoring screen, mobile terminal and database server;

[0015] The lighting management cloud platform is linked to the smart lighting cloud box. The GPS and 4G / 5G IoT integrated modules of the smart lighting cloud box automatically obtain longitude and latitude information. The cloud box control mainboard controls the GPS and 4G / 5G IoT integrated modules to wirelessly report progress data to the lighting management cloud platform. The real-time location of the device is retrieved through Amap or Baidu Maps, and the operating status, power consumption data, circuit working status, circuit lighting status, real-time current, voltage, and power information, and historical usage data of the high pole lamp are displayed. The high pole lamp can also be turned on and off and monitored.

[0016] The image acquisition module captures environmental video images at a predetermined distance, extracts and stores motion trajectory information of moving vehicles and pedestrians, and establishes a motion trajectory database. At the same time, it detects specific targets based on the environmental video images and, if any dangerous factors are detected, issues remote sound and light warnings to vehicles and pedestrians. The motion trajectory database includes the image acquisition time, image location, and weather conditions at the time of image acquisition. The specific targets include, but are not limited to, obstacles, subsidence areas, and accident scenes.

[0017] Based on the motion trajectory information of the moving vehicles and pedestrians in the motion trajectory database, the total number of moving vehicles and pedestrians within a predetermined distance range of the target high pole light is determined, and combined with the information in the motion trajectory database, a neural network training sample library is formed; and based on the neural network training sample library, a pedestrian and vehicle flow prediction model based on a neural network is constructed;

[0018] Based on the pedestrian and vehicle flow prediction model, the total number of traveling vehicles and pedestrians is predicted, and the brightness value and the light coverage value of the target high pole lamp are determined; at the same time, the pedestrian and vehicle flow information is wirelessly pushed to the GIS map system as the broadcast data for subsequent vehicle road conditions.

[0019] Optionally, the switching control and operation monitoring of the high pole lamp specifically includes:

[0020] Calculate the time between dusk and dawn in the current time period, automatically turn on lights at night and automatically turn off lights at dawn, adjust light on / off times in advance, manually correct longitude and latitude, synchronize device time, and set different on / off times for lighting circuits; and automatically switch the device to local mode when offline, executing preset offline longitude and latitude automatic lighting operations by determining whether it is communicating normally with the cloud platform server;

[0021] By remotely opening and closing all circuits or a single circuit of the device, monitoring the working status of the relay, and monitoring the device voltage and current values ​​through the circuit monitoring module, it is determined whether the lighting circuit is actually open or closed; by setting the current, voltage, power and normal working range values ​​of the lighting equipment of different circuits, the fault of the lighting and circuit is judged. The database server records the fault and notifies the data monitoring screen and the lighting control background. It automatically generates a real-time alarm and a lighting fault work order based on the configured trigger alarm conditions. The work order is automatically distributed to the preset operation and maintenance personnel, and the operation and maintenance personnel are notified through mobile terminal message reminders and SMS text messages to promptly handle the fault.

[0022] Among them, the control of automatically turning on the lights at night and automatically turning off the lights at dawn includes two control methods: local control and remote control. The local control is through the local control button switch of the lighting control cloud box to control the loop relay to open and close; the remote control is through the lighting control background or mobile phone terminal to issue switch instructions, and use the MQTT communication protocol attributes to issue attributes. Finally, the cloud box control mainboard receives the instructions, controls the high-power relay integrated module to open or close, and then controls the switch of the light.

[0023] Optionally, it also includes: the cloud box main control board receives the real-time working parameters of the circuit monitored by the monitoring module set on the corresponding circuit loop of the lamp, records the power current, voltage, and power information of the circuit and calculates the power consumption of the corresponding high pole lamp, and wirelessly reports the power consumption to the lighting management cloud platform through GPS and 4G / 5G Internet of Things integrated modules, and stores the power consumption data of the device through the database server; through the electricity fee configuration module of the lighting control background, the electricity price information is input to automatically calculate the electricity cost, and support annual, monthly, weekly and daily electricity consumption statistics query.

[0024] Optionally, after the lighting control background receives the real-time working parameter data, it also includes statistical processing of the data, statistics on the overall power consumption, daily power consumption, year-on-year power consumption, month-on-month power consumption, and time-sharing power consumption of single equipment, single circuit, multiple equipment, projects, and regional equipment, and displays the equipment energy efficiency information in real time through the data monitoring screen.

[0025] Optionally, it also includes selecting multiple smart lighting cloud boxes that need to be adjusted through the lighting control background when the smart lighting cloud box is in an online communication state, adjusting the switch control instructions or device parameters, and sending the attributes to different cloud box control main boards through the MQTT protocol, and the cloud box control main board executes the corresponding parameter data.

[0026] Optionally, it also includes sending the server time to the cloud box control mainboard of the smart lighting cloud box through the lighting control background, the cloud box control mainboard automatically synchronizes the server time, and manually adjusts the device time when the synchronization fails.

[0027] Optionally, the detecting of a specific target based on an environmental video image is specifically:

[0028] Collect specific target images of the road to build a feature database, select representative pictures from the feature database as training and test sets, and annotate the photos in the training set with dangerous features, such as obstacles, collapsed areas, and accident scenes;

[0029] Perform image preprocessing on the training set sample images; the preprocessing includes image scaling, cropping, brightness adjustment, and data augmentation, including rotation, translation, and flipping;

[0030] The pre-processed training set sample images are processed by the Yolov5 algorithm, and the target obstacles are selected using the labelimg software to start training the specific target image dataset;

[0031] A specific target image dataset is used to train the basic YOLOv5 network model to obtain a trained YOLOv5 network model;

[0032] The trained YOLOv5 network model is used to identify specific target images, and the output results of the trained YOLOv5 network model are used as the recognition results to complete the detection of specific target images; among them, the basic YOLOv5 network model is the YOLOv5s network model, and the loss function in the training process is the WIoU loss function.

[0033] Optionally, the neural network-based pedestrian and vehicle flow prediction model is constructed based on the neural network training sample library, specifically:

[0034] Dividing the neural network training sample library into training samples and test samples, using the learning samples as BP neural network input vectors, and using the total number of traveling vehicles and pedestrians as target output;

[0035] Set the initial weights and initial thresholds of the training samples, set the minimum expected error of the BP neural network, the learning rate and the number of cycles, and calculate the output of the hidden layer and output layer of the BP neural network as well as the deviation between the target value and the actual output;

[0036] Determine whether the deviation is within the allowable range. If not, optimize the output layer and hidden layer weights using a swarm particle algorithm until the error is within the allowable range, thereby completing the training of the pedestrian and vehicle flow prediction model based on the neural network.

[0037] The test samples were input into the constructed pedestrian and vehicle flow prediction model based on neural network, and the total number of moving vehicles and pedestrians was predicted to verify the pedestrian and vehicle flow prediction model;

[0038] The relevant parameters of the total number of moving vehicles and pedestrians to be predicted are input into the verified pedestrian and vehicle flow prediction model to predict the total number of moving vehicles and pedestrians in the future time series.

[0039] Optionally, the total number of vehicles and pedestrians is predicted based on the pedestrian and vehicle flow prediction model, and the target high pole lamp brightness value and light coverage value are determined, specifically:

[0040] Based on the pedestrian and vehicle flow prediction model, the number x of vehicles and pedestrians within the predetermined distance range of the target high pole lamp is obtained, and the corresponding target high pole lamp required brightness value is used as the required value L(x), and the current brightness value of the target high pole lamp is used as the comparison value L. The target high pole lamp required brightness value L(x) is calculated according to the formula:

[0041]

[0042] Among them, L max represents the maximum brightness of the target high pole lamp, λ represents the brightness attenuation parameter of the target high pole lamp, x represents the total number of moving vehicles and pedestrians, and ξ represents the preset reference value of the total number of moving vehicles and pedestrians;

[0043] Comparing the demand value L(x) with the comparison value L: when L(x)>L, the brightness of the target high pole lamp is increased; when L(x)<L, the brightness of the target high pole lamp is reduced; when L(x)=L, the brightness of the target high pole lamp is not changed;

[0044] Calculate the traffic flow density value ρ based on the number x of vehicles and pedestrians within the predetermined distance range of the target high pole lamp, where ρ = x / S, where S = πL2, and L is the distance within the predetermined distance range;

[0045] Based on the traffic flow density value ρ, the light coverage value is adjusted, where the light coverage radius R and the traffic flow density value ρ satisfy R=K*ρ, where K is a constant; the specific methods for adjusting the light coverage value include adjusting the light pitch angle, adjusting the light beam angle, and adjusting the light height from the ground.

[0046] The beneficial effects of the above technical solution of the present invention are as follows:

[0047] This invention uses commands issued by operating a lighting management platform, a mobile phone app, and a lighting intelligent control cloud box to control on-site lighting individually, in batches, and by schedule. It can also control it by time period or region, thus achieving intelligent monitoring, control, and refined management of high-pole lighting. It also enables map functions, equipment inspections, lighting control, power overload readings, gateway inspections, equipment time calibration, electricity price clocks, batch parameter adjustment, energy efficiency management, and alarms. Furthermore, it can also implement traffic flow monitoring, safety monitoring, and light illumination and coverage control based on image recognition and neural network technology. This is of great significance for improving lighting efficiency, reducing energy consumption, and achieving intelligent urban traffic management. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is the framework diagram of the high pole lamp intelligent lighting control system described in the present invention.

[0049] Figure 2 This is a hierarchical structure diagram of the high pole lamp intelligent lighting control system described in the present invention.

[0050] Figure 3 This is the principle block diagram of the smart control cloud box of the present invention. DETAILED DESCRIPTION

[0051] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0052] Example 1

[0053] like Figure 1 and Figure 3 As shown, the present invention proposes an intelligent control system for high pole lamp lighting, including multiple intelligent lighting cloud boxes 1, a lighting management cloud platform 2, a lighting control background 3, a data monitoring screen 4, a mobile terminal 5 and a database server 6. The multiple intelligent lighting cloud boxes 1 are electrically connected to the corresponding circuit loops 7 of multiple lamps 8 respectively, the intelligent lighting cloud box 1 is wirelessly connected to the lighting management cloud platform 2, the lighting management cloud platform 2 is communicatively connected to the lighting control background 3, the data monitoring screen 4, and the database server 6, and the lighting management cloud platform 2 and the mobile terminal 5 are wirelessly connected.

[0054] The intelligent lighting cloud box 1 includes a cloud box control mainboard 11 and a liquid crystal display 12 electrically connected to the cloud box control mainboard 11, a GPS and 4G / 5G Internet of Things integrated module 13, a power supply module 14, a high-power relay integrated module 15, a local on / off light button 16, a local / remote selection button 17, a 4-position aviation socket 111 and an image acquisition module 113, wherein the GPS and 4G / 5G Internet of Things integrated module 13 is wirelessly connected to the lighting management cloud platform 2.

[0055] The power supply module 14 is electrically connected to the leakage protection switch 18, the GPS and 4G / 5G Internet of Things integrated module 13 is electrically connected to the GPS positioning antenna interface 19 and the 4G / 5G Internet of Things antenna interface 110, and the high-power relay integrated module 15 is electrically connected to the input / output terminal block 112.

[0056] Among them, the lighting management cloud platform 2 and the lighting control background 3 are servers or industrial computers.

[0057] The lighting circuit is connected to the lighting fixture, the relay controls the circuit switch to be energized, the circuit monitoring module monitors the real-time working parameters of the circuit and sends them to the main control board. The main control board realizes automatic control through real-time communication between the communication module and the control management system composed of the lighting management cloud platform 2, the lighting control background 3, the data monitoring screen 4, the mobile terminal 5 and the database server 6. The LCD display locally checks the working status of the circuit; the GPS positioning module provides geographic location information positioning.

[0058] By linking the control system with the equipment, the equipment positioning module automatically obtains the longitude and latitude information, calculates the dark and dawn times of the current time period, and realizes automatic light-on at night and automatic light-off at dawn. In addition, the light-on and light-off times can be adjusted in advance, the longitude and latitude can be manually corrected, the equipment time can be corrected synchronously, and different switch times of the lamp circuits can be set separately. At the same time, by judging whether the communication with the cloud platform server is normal, the equipment can be automatically switched to the local mode when it is offline to execute the preset offline longitude and latitude automatic light-on and light-off mode.

[0059] By remotely opening and closing all circuits of the device or a single circuit, the switch is turned on to monitor the working status of the relay. At the same time, the circuit detection device is used to monitor the voltage and current values ​​of the device to determine whether the lighting circuit is actually turned on or off.

[0060] By monitoring the working status of lamps in real time and reporting it to the management platform, the management platform can determine the faults of lamps and circuits by setting the current, voltage, power and normal working range values ​​of lamps in different circuits. The background system records the fault notification to the implementation large screen and management background. The background system automatically generates a fault work order for the lamp equipment, and the system automatically distributes the work order to the preset operation and maintenance management personnel. The operation and maintenance personnel are notified through mobile phone APP messages and SMS to deal with the fault in time. The work order is automatically completed and filed in the work order history record by the fault handling system.

[0061] The GPS positioning module of the lighting control cloud box collects the current equipment latitude and longitude information and sends it to the lighting management platform. At the same time, if the positioning correction fails, the latitude and longitude correction positioning information can be entered. After the positioning is completed, the positioning information is automatically synchronized to the main control board.

[0062] The present invention adopts high pole lamp multi-circuit different mode lighting control, the specific contents are as follows:

[0063] Remote automatic control: The high pole lamp main control board receives control instructions from the remote computer and controls the opening or closing of multiple groups of high pole lamps according to these instructions; the high pole lamp main control board maintains communication connection with the remote computer through wireless means (4G / 5G) or optical cable, controls the opening or closing of the lamps according to the instructions of the remote computer, and uploads the collected data to the remote computer.

[0064] Local automatic control function: When the communication is interrupted or fails, the high pole lamp main control board automatically recognizes that the communication is in an abnormal state. In the case of abnormal communication, the high pole lamp control automatically switches to the local automatic control state. Each high pole lamp main control board can automatically calculate the light on and off time according to the local latitude and longitude, and automatically control the light to be turned on or off according to the calculated time.

[0065] Local manual control: The high pole lamp main control board has an input interface for receiving local status feedback. When the local signal is turned on, the lamp is lit. When the local signal is disconnected, the lamp is turned off. After a delay of 10 seconds, the remote automatic control mode is restored.

[0066] Remote time control: When the main control board of the high pole lamp is in remote working mode, different lighting times are set for different circuits of the high pole lamp through the background. During the lighting working time, the background automatically issues switch commands to the circuit. Multiple working time periods can be set. Different circuits of the same lamp pole can be set separately without affecting each other. Multiple different lighting control strategies can be set in advance.

[0067] Cluster batch control: The above control mode can be used to control high pole lamp cluster batches.

[0068] In addition, the loop power consumption is monitored and transmitted to the cloud management platform through the loop monitoring module of the control cloud box. The management platform can set the electricity price, accurately calculate the power consumption and generate statistical results. It can realize annual time-sharing electricity consumption statistics, fixed electricity price statistics, independent statistical queries for different loops, and export result reports. The system consists of a device access module on the device side and the platform automatically obtains the device address. The device identification, server communication address and key are set in advance during device production. When the device is installed, the device identification and key are entered through the platform device access interface to automatically access the management platform. The system automatically connects to the device, obtains the working status and data of the device, and is ready for use.

[0069] Example 2

[0070] The present invention also provides a control method for the above-mentioned high pole lamp lighting intelligent control system, comprising the following steps:

[0071] S1. The real-time working parameters of the circuits monitored by the monitoring modules installed on the corresponding circuit loops of multiple lamps are sent to the cloud box control mainboard of the intelligent lighting cloud box. The cloud box control mainboard wirelessly sends the real-time working parameter data to the lighting management cloud platform through the GPS and 4G / 5G Internet of Things integrated module. The lighting management cloud platform then forwards the real-time working parameter data to the lighting control background, data monitoring screen, mobile terminal and database server.

[0072] S2. The lighting management cloud platform is linked to the smart lighting cloud box. The GPS and 4G / 5G IoT integrated modules of the smart lighting cloud box automatically obtain longitude and latitude information. The cloud box control mainboard controls the GPS and 4G / 5G IoT integrated modules to wirelessly report progress data to the lighting management cloud platform. The real-time location of the device is retrieved through Gaode Map or Baidu Map, and the operating status, power consumption data, circuit working status, circuit lighting status, real-time current, voltage, and power information, and historical usage data of the high pole lamp are displayed. The high pole lamp can also be switched on and off and monitored.

[0073] S3. Capture environmental video images at a predetermined distance, extract and store motion trajectory information of moving vehicles and pedestrians, and establish a motion trajectory database. Specific targets are detected based on the environmental video images. If a dangerous factor is detected, remote audio and visual warnings are issued to vehicles and pedestrians. For example, if a road collapse occurs 50-100 meters ahead of a vehicle, the high-mast light can be adjusted to high-beam flash and an audio alarm can be sounded to prevent subsequent vehicles from unknowingly entering the danger zone. Specifically, the video images at a predetermined distance centered on the target high-mast light can be decomposed frame by frame into continuous image frames. Each image frame represents a time point in the video. Decomposing the video into image frames provides a continuous image data stream for subsequent target detection and analysis, with each frame representing the situation at a specific time point. Furthermore, specific time points can be controlled according to the time granularity required by the system, for example, by dividing the time interval into 50ms granularity to obtain image data at a specific time point. The motion trajectory database also includes the image acquisition time, image location, and weather conditions at the time of image acquisition, all of which affect the total number of moving vehicles and pedestrians. Specific targets include but are not limited to obstacles, collapse areas, and accident scenes. These factors are all factors that affect traffic safety, and continuous monitoring can be achieved using high pole lights as a carrier.

[0074] S4. Based on the motion trajectory information of vehicles and pedestrians in the motion trajectory database, the total number of vehicles and pedestrians within a predetermined distance range of the target high pole light is determined. Combined with the motion trajectory database information, a neural network training sample library is formed. Based on the neural network training sample library, a neural network-based pedestrian and vehicle flow prediction model is constructed. A schematic diagram of the training sample library is shown in Table 1.

[0075] Table 1: Training sample library

[0076]

[0077] S5. Based on the pedestrian and vehicle flow prediction model, the total number of vehicles and pedestrians is predicted, and the target high-pole light brightness and light coverage values ​​are determined. Simultaneously, pedestrian and vehicle flow information is wirelessly pushed to the GIS mapping system as data for subsequent vehicle traffic condition reports.

[0078] In this embodiment, the on / off control and operation monitoring of the high pole lamp specifically include:

[0079] Calculates the time between dusk and dawn in the current time period, automatically turning lights on at night and off at dawn, adjusting light on / off times in advance, manually correcting longitude and latitude, synchronizing device time, and setting different on / off times for lighting circuits. Furthermore, by determining whether communication with the cloud platform server is working properly, the device automatically switches to local mode when offline and automatically turns lights on and off at preset longitude and latitude.

[0080] By remotely starting and stopping the device all circuits or single circuit, monitoring the relay working state, and monitoring the device voltage, current value through the circuit monitoring module, it is judged whether the lamp circuit is truly opened or closed. By setting the current, voltage, power, normal working interval value of different circuit lamp equipment, the fault of the lamp and the circuit is judged, the database server records the fault and notifies the data monitoring large screen and the lighting control background, automatically combines the configured trigger alarm condition to realize real-time alarm and generate the fault work order of the lamp, automatically distributes the work order to the preset operation and maintenance personnel, and timely handles the fault through the mobile terminal message reminder and the SMS notification of the operation and maintenance personnel.

[0081] Among them, the control of automatically turning on the light at night and automatically turning off the light at sunrise includes two control modes of local control and remote control. The local control switches the circuit relay on and off through the lighting control cloud box local control button switch. The remote control issues the on-off instruction through the lighting control background or the mobile terminal, and issues the attribute through the MQTT communication protocol attribute. Finally, the cloud box control mainboard receives the instruction to control the high-power relay integrated module to turn on or off, thereby controlling the switch of the lamp.

[0082] In the embodiment, the cloud box main control board also receives the real-time working parameters of the circuit monitored by the monitoring module arranged on the corresponding circuit of the lamp, records the power consumption of the corresponding high-pole lamp, and calculates the power consumption of the corresponding high-pole lamp. The power consumption is reported to the lighting management cloud platform through the GPS and 4G / 5G Internet of Things integrated module, and the power consumption data of the equipment is stored through the database server. The electricity cost configuration module of the lighting control background is inputted, the electricity cost is automatically calculated, and the annual, monthly, weekly, and daily electricity consumption statistical query is supported.

[0083] In the embodiment, after the lighting control background receives the real-time working parameter data, the data is statistically processed, the overall power consumption of single equipment, single circuit, multiple equipment, project, and regional equipment, single-day power consumption, same-period power consumption, and ring-by power consumption are statistically processed, and the equipment energy efficiency information is displayed in real time through the data monitoring large screen.

[0084] In the embodiment, when the intelligent lighting cloud box is in the communication online state, the lighting control background selects multiple intelligent lighting cloud boxes that need to be adjusted, adjusts the switch control instruction or equipment parameter, and issues the attribute to the different cloud box control mainboards through the MQTT protocol. The cloud box control mainboard executes the corresponding parameter data.

[0085] In the embodiment, the lighting control background issues the server time to the cloud box control mainboard of the intelligent lighting cloud box, and the cloud box control mainboard automatically synchronizes the server time. When the synchronization fails, the equipment time is manually adjusted.

[0086] In this embodiment, the specific target is detected based on the environmental video image, specifically:

[0087] Specific target images of the road are collected to build a feature database. Representative pictures in the feature database are selected as training sets and test sets. Dangerous features are annotated on the photos in the training set. Dangerous features include obstacles, collapsed areas, and accident scenes.

[0088] Perform image preprocessing on the training set sample images; preprocessing includes image scaling, cropping, brightness adjustment, and data enhancement, including rotation, translation, and flipping.

[0089] The preprocessed training set sample images are processed by the Yolov5 algorithm, and the target obstacles are selected using the labelimg software to start training the specific target image dataset.

[0090] A specific target image dataset is used to train the basic YOLOv5 network model to obtain a trained YOLOv5 network model.

[0091] The trained YOLOv5 network model is used to identify specific target images, and the output results of the trained YOLOv5 network model are used as the recognition results to complete the detection of specific target images; among them, the basic YOLOv5 network model is the YOLOv5s network model, and the loss function in the training process is the WIoU loss function.

[0092] In this embodiment, a neural network-based pedestrian and vehicle flow prediction model is constructed based on the neural network training sample library, specifically:

[0093] The neural network training sample library is divided into training samples and test samples, the learning samples are used as BP neural network input vectors, and the total number of moving vehicles and pedestrians is used as the target output; the data samples of a period of time in the neural network training sample library (including image acquisition time, image location, and weather at the time of image acquisition) are divided into two parts, the first 75% are used as training samples for training the network, and the last 25% are used as test samples for testing the network.

[0094] Set the initial weights and initial thresholds of the training samples, set the minimum expected error of the BP neural network, the learning rate and the number of cycles, and calculate the output of the hidden layer and output layer of the BP neural network as well as the deviation between the target value and the actual output.

[0095] If the deviation is not within the allowable range, the output layer and the hidden layer weights are optimized by the swarm particle algorithm until the error is within the allowable range, and the training of the vehicle and pedestrian flow prediction model based on the neural network is completed. If the deviation is not within the allowable range, the error signal is back-propagated along the original path, and the network connection weights of the neurons of each layer are gradually adjusted until the error is less than the specified accuracy. At this time, the first group of learning is completed, and the next group of learning is entered. Until the connection weights have a prediction error within the specified range for all training groups, the optimal weights at this time are output. The more training samples, the more sufficient the network learning, the greater the network experience value, and the higher the prediction accuracy. When the error is less than 10%, the training is stopped, and the prediction is started.

[0096] The test sample is input into the vehicle and pedestrian flow prediction model based on the neural network constructed in step 33, and the total number of driving vehicles and pedestrians is predicted to verify the vehicle and pedestrian flow prediction model. For example, after the network training is completed, the network is tested by using another 25% of the data samples to check whether the model meets the requirements. The total number of driving vehicles and pedestrians corresponding to another 25% of the data samples is predicted by using the BP neural network obtained by training, and the error between the model prediction value and the actual observation value is compared. When the prediction error of the BP neural network model for each group of test data is lower than the specified level, the test is passed, and the total number of driving vehicles and pedestrians can be predicted. At this time, the correlation coefficient between the model prediction value and the actual measurement value is 0.90, the maximum prediction error of the model is 10%, and the test is passed.

[0097] The related parameters of the total number of driving vehicles and pedestrians to be predicted are input into the verified vehicle and pedestrian flow prediction model, and the total number of driving vehicles and pedestrians in the future time sequence is predicted.

[0098] In this embodiment, based on the vehicle and pedestrian flow prediction model, the total number of driving vehicles and pedestrians is predicted, and the target high-pole lamp brightness value and the light near-far coverage value are determined. Specifically,

[0099] Based on the vehicle and pedestrian flow prediction model, the number x of driving vehicles and pedestrians in the target high-pole lamp predetermined distance range is obtained, and the corresponding target high-pole lamp demand brightness value is taken as the demand value L(x). The current brightness value of the target high-pole lamp is taken as the comparison value L, and the target high-pole lamp demand brightness value L(x) is calculated according to the formula:

[0100]

[0101] L(x) = Lmax x e-λx max Lmax represents the maximum brightness value of the target high-pole lamp, λ represents the brightness attenuation parameter of the target high-pole lamp, x represents the total number of driving vehicles and pedestrians, and ξ represents the preset reference value of the total number of driving vehicles and pedestrians.

[0102] Compare the demand value L(x) with the comparison value L: when L(x)>L, the brightness of the target high pole lamp is increased; when L(x)<L, the brightness of the target high pole lamp is reduced; when L(x)=L, the brightness of the target high pole lamp is not changed.

[0103] Based on the number x of vehicles and pedestrians within the predetermined distance range of the target high pole lamp, the traffic flow density value ρ is calculated, where ρ = x / S, where S = πL2, and L is the distance within the predetermined distance range.

[0104] Based on the traffic flow density value ρ, the light coverage value is adjusted, where the light coverage radius R and the traffic flow density value ρ satisfy R=K*ρ, where K is a constant; the specific methods for adjusting the light coverage value include adjusting the light pitch angle, adjusting the light beam angle, and adjusting the light height from the ground.

[0105] In summary, the present invention issues commands through the operation of the lighting management platform, mobile phone APP, and lighting intelligent control cloud box to control on-site lamps individually, in batches, and in plans. It can also control them by time period and area, thus realizing intelligent monitoring and control and refined management of high-pole lighting. At the same time, it can realize map functions, equipment inspections, lighting control, power over-reading, gateway inspections, equipment time calibration, electricity price clocks, batch parameter adjustment, energy efficiency management, alarms and other functions, as well as traffic flow monitoring, safety monitoring, and light illumination and coverage control functions based on image recognition and neural network technology. It is of great significance to improve lighting efficiency, reduce energy consumption, and realize intelligent management of urban traffic.

[0106] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A control method for a high pole lamp lighting intelligent control system, characterized in that: include: The real-time operating parameters of the circuits monitored by the monitoring modules installed on the corresponding circuit loops of multiple lamps are sent to the cloud box control mainboard of the intelligent lighting cloud box. The cloud box control mainboard wirelessly sends the real-time operating parameter data to the lighting management cloud platform through GPS and 4G / 5G IoT integrated modules. The lighting management cloud platform then forwards the real-time operating parameter data to the lighting control backend, data monitoring screen, mobile terminal and database server; The lighting management cloud platform is linked to the smart lighting cloud box. The GPS and 4G / 5G IoT integrated modules of the smart lighting cloud box automatically obtain longitude and latitude information. The cloud box control mainboard controls the GPS and 4G / 5G IoT integrated modules to wirelessly report progress data to the lighting management cloud platform. The real-time location of the device is retrieved through Amap or Baidu Maps, and the operating status, power consumption data, circuit working status, circuit lighting status, real-time current, voltage, and power information, and historical usage data of the high pole lamp are displayed. The high pole lamp can also be turned on and off and monitored. The image acquisition module captures environmental video images at a predetermined distance, extracts and stores motion trajectory information of moving vehicles and pedestrians, and establishes a motion trajectory database. At the same time, it detects specific targets based on the environmental video images and, if any dangerous factors are detected, issues remote sound and light warnings to vehicles and pedestrians. The motion trajectory database includes the image acquisition time, image location, and weather conditions at the time of image acquisition. The specific targets include, but are not limited to, obstacles, subsidence areas, and accident scenes. Based on the motion trajectory information of the moving vehicles and pedestrians in the motion trajectory database, the total number of moving vehicles and pedestrians within a predetermined distance range of the target high pole light is determined, and combined with the information in the motion trajectory database, a neural network training sample library is formed; and based on the neural network training sample library, a pedestrian and vehicle flow prediction model based on a neural network is constructed; Based on the pedestrian and vehicle flow prediction model, the total number of traveling vehicles and pedestrians is predicted, and the brightness value and the distance coverage value of the target high pole lamp are determined; at the same time, the pedestrian and vehicle flow information is wirelessly pushed to the GIS map system as the subsequent vehicle road condition broadcast data; The total number of vehicles and pedestrians is predicted based on the pedestrian and vehicle flow prediction model, and the target high pole lamp brightness value and light coverage value are determined, specifically: Based on the pedestrian and vehicle flow prediction model, the number x of vehicles and pedestrians within the predetermined distance range of the target high pole lamp is obtained, and the corresponding target high pole lamp required brightness value is used as the required value L(x), and the current brightness value of the target high pole lamp is used as the comparison value L. The target high pole lamp required brightness value L(x) is calculated according to the formula: Among them, L max represents the maximum brightness of the target high pole lamp, λ represents the brightness attenuation parameter of the target high pole lamp, x represents the total number of moving vehicles and pedestrians, and ξ represents the preset reference value of the total number of moving vehicles and pedestrians; Comparing the demand value L(x) with the comparison value L: when L(x)>L, the brightness of the target high pole lamp is increased; when L(x)<L, the brightness of the target high pole lamp is reduced; when L(x)=L, the brightness of the target high pole lamp is not changed; Based on the number x of vehicles and pedestrians within the predetermined distance range of the target high pole lamp, the traffic flow density value ρ is calculated, where ρ = x / S, where S = πL 2 , L is the distance within the predetermined distance range; Based on the traffic flow density value ρ, the light coverage value is adjusted, where the light coverage radius R and the traffic flow density value ρ satisfy R=K*ρ, where K is a constant; the specific methods for adjusting the light coverage value include adjusting the light pitch angle, adjusting the light beam angle, and adjusting the light height from the ground.

2. The intelligent control method for high pole lamp lighting according to claim 1, characterized in that: The switching control and operation monitoring of the high pole lamp specifically includes: Calculate the time between dusk and dawn in the current time period, automatically turn on lights at night and automatically turn off lights at dawn, adjust light on / off times in advance, manually correct longitude and latitude, synchronize device time, and set different on / off times for lighting circuits; and automatically switch the device to local mode when offline, executing preset offline longitude and latitude automatic lighting operations by determining whether it is communicating normally with the cloud platform server; By remotely opening and closing all circuits or a single circuit of the device, monitoring the working status of the relay, and monitoring the device voltage and current values ​​through the circuit monitoring module, it is determined whether the lighting circuit is actually open or closed; by setting the current, voltage, power and normal working range values ​​of the lighting equipment of different circuits, the fault of the lighting and circuit is judged. The database server records the fault and notifies the data monitoring screen and the lighting control background. It automatically generates a real-time alarm and a lighting fault work order based on the configured trigger alarm conditions. The work order is automatically distributed to the preset operation and maintenance personnel, and the operation and maintenance personnel are notified through mobile terminal message reminders and SMS text messages to promptly handle the fault. Among them, the control of automatically turning on the lights at night and automatically turning off the lights at dawn includes two control methods: local control and remote control. The local control is through the local control button switch of the lighting control cloud box to control the loop relay to open and close; the remote control is through the lighting control background or mobile phone terminal to issue switch instructions, and use the MQTT communication protocol attributes to issue attributes. Finally, the cloud box control mainboard receives the instructions, controls the high-power relay integrated module to open or close, and then controls the switch of the light.

3. The intelligent control method for high pole lamp lighting according to claim 1, characterized in that: Also includes: The cloud box main control board receives the real-time working parameters of the circuit monitored by the monitoring module set on the corresponding circuit loop of the lamp, records the current, voltage, and power information of the circuit and calculates the power consumption of the corresponding high pole lamp, and wirelessly reports the power consumption to the lighting management cloud platform through GPS and 4G / 5G Internet of Things integrated module, and stores the device power consumption data through the database server; through the electricity fee configuration module of the lighting control background, the electricity price information is input to automatically calculate the electricity cost, and support annual, monthly, weekly and daily electricity consumption statistics query.

4. The intelligent control method for high pole lamp lighting according to claim 1, characterized in that: It also includes statistical processing of data after the lighting control background receives real-time working parameter data, and statistics on the overall power consumption, daily power consumption, year-on-year power consumption, month-on-month power consumption, and time-sharing power consumption of single equipment, single circuit, multiple equipment, projects, and regional equipment, and displays the equipment energy efficiency information in real time through the data monitoring screen.

5. The intelligent control method for high pole lamp lighting according to claim 1, characterized in that: It also includes when the smart lighting cloud box is in an online communication state, selecting multiple smart lighting cloud boxes that need to be adjusted through the lighting control background, adjusting the switch control instructions or device parameters, and sending the attributes to different cloud box control main boards through the MQTT protocol, and the cloud box control main boards execute the corresponding parameter data.

6. The intelligent control method for high pole lamp lighting according to claim 1, characterized in that: It also includes sending the server time to the cloud box control mainboard of the smart lighting cloud box through the lighting control background. The cloud box control mainboard automatically synchronizes the server time and manually adjusts the device time when the synchronization fails.

7. The intelligent control method for high pole lamp lighting according to claim 1, characterized in that: The specific target is detected based on the environmental video image, specifically: Collect specific target images of the road to build a feature database, select representative pictures from the feature database as training and test sets, and annotate the photos in the training set with dangerous features, such as obstacles, collapsed areas, and accident scenes; Perform image preprocessing on the training set sample images; The preprocessing includes image scaling, cropping, brightness adjustment, and data enhancement, including rotation, translation, and flipping; The pre-processed training set sample images are processed by the Yolov5 algorithm, and the target obstacles are selected using the labelimg software to start training the specific target image dataset; Use a specific target image dataset to train the basic YOLOv5 network model to obtain a trained YOLOv5 network model; The trained YOLOv5 network model is used to identify specific target images, and the output results of the trained YOLOv5 network model are used as the recognition results to complete the detection of specific target images; among them, the basic YOLOv5 network model is the YOLOv5s network model, and the loss function in the training process is the WIoU loss function.

8. The intelligent control method for high pole lamp lighting according to claim 1, characterized in that: The pedestrian and vehicle flow prediction model based on the neural network is constructed according to the neural network training sample library, specifically: Dividing the neural network training sample library into training samples and test samples, using the learning samples as BP neural network input vectors, and using the total number of traveling vehicles and pedestrians as target output; Set the initial weights and initial thresholds of the training samples, set the minimum expected error of the BP neural network, the learning rate and the number of cycles, and calculate the output of the hidden layer and output layer of the BP neural network as well as the deviation between the target value and the actual output; Determine whether the deviation is within the allowable range. If not, optimize the output layer and hidden layer weights using a swarm particle algorithm until the error is within the allowable range, thereby completing the training of the pedestrian and vehicle flow prediction model based on the neural network. The test samples were input into the constructed pedestrian and vehicle flow prediction model based on neural network, and the total number of moving vehicles and pedestrians was predicted to verify the pedestrian and vehicle flow prediction model; The relevant parameters of the total number of moving vehicles and pedestrians to be predicted are input into the verified pedestrian and vehicle flow prediction model to predict the total number of moving vehicles and pedestrians in the future time series.

Images