Expressway pedestrian early warning system and method based on thermal imaging and cooperative positioning of reflection points

Abstract

The invention discloses an expressway pedestrian early warning system and method based on thermal imaging and cooperative positioning of reflection points, and belongs to the field of intelligent traffic active safety. The system comprises a thermal imaging module, a positioning reflection point array and a central processing unit, the thermal imaging module scans a heat source signal in real time and preliminarily recognizes pedestrian features, positioning reflection points are arranged at intervals of 50 meters, are designed in a conditional trigger mode, are dormant in a normal state and are activated and used for emitting positioning signals only after receiving awakening signals of the thermal imaging module, and the positioning reflection points are arranged in the central processing unit. And the central processing unit fuses the data to realize pedestrian position calibration and early warning generation. According to the method, a heat source track is generated through thermal imaging scanning, corresponding reflection points are activated, and coordinates are calculated based on a hybrid positioning model containing dynamic weights. According to the scheme, the problems of high false alarm rate, low positioning precision and the like of traditional detection are solved, the rain and fog false alarm rate is reduced to 0.8%, the positioning response time delay is 0.35 second, the service life of equipment exceeds 5 years, the method is suitable for all-weather scenes, and the deployment cost is reduced due to the modular design.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation, and in particular to a highway pedestrian warning system and method for collaboratively positioning reflection points using thermal imaging. Background Art

[0002] Existing highway pedestrian detection technology lacks a coordinated mechanism for initial thermal imaging screening and dynamic activation of positioning reflective points, making it difficult to achieve both low false alarm rates and high-precision positioning. Traditional detection methods, such as infrared sensors, are susceptible to environmental interference such as rain, fog, and light, resulting in false alarm rates as high as 23.5% in heavy rain, and significantly reduced detection accuracy in dark areas at night. Existing technology also suffers from extended positioning response times of up to 2.1 seconds, making it impossible to quickly determine pedestrian coordinates for emergency response, especially in dark environments such as tunnels, where positioning errors can exceed 1 meter. Passive reflective devices only reflect light passively and lack a triggering mechanism for linkage with detection equipment. Their lifespan is generally limited to 1.5 years, resulting in high maintenance costs. Existing infrared sensing-based monitoring systems lack coordinated detection of thermal imaging and positioning reflective points, and lack dynamic activation mechanisms to reduce false alarm rates and improve positioning efficiency, making them difficult to meet the all-weather pedestrian warning needs of highways.

[0003] Existing technologies lack a synergistic mechanism that combines initial thermal imaging screening with dynamic activation of positioning reflective points, making it impossible to simultaneously achieve low false alarm rates and high-precision positioning. They also suffer from high false alarm rates in rain and fog, are susceptible to interference from lighting and weather conditions (such as rain and fog), and have low nighttime detection accuracy. Existing technologies are unable to quickly determine the specific location of pedestrians, especially in dark areas, making it difficult to coordinate emergency response. Summary of the Invention

[0004] The purpose of the present invention is to provide a highway pedestrian warning system and method that uses thermal imaging to collaboratively locate reflective points. This system aims to solve the problems of precise positioning and all-weather reliability in highway pedestrian detection. At the same time, it reduces deployment costs through modular design, achieves dual verification, and avoids the defects of a single technology.

[0005] The present invention is implemented as follows: a highway pedestrian warning system using thermal imaging and collaborative positioning of reflection points, comprising: Thermal imaging module: Deployed along highways, it scans heat source signals in real time and preliminarily identifies pedestrian features. The thermal imaging module has a 120° field of view and a temperature sensitivity of 0.1°C. Positioning reflection point array: Deployed at 50-meter intervals on highway guardrails, at a height of 1.2 meters, overlapping the field of view of the thermal imaging module and embedded in the reflectors on both sides of the highway. It adopts a conditional trigger design and is normally dormant, activating only after receiving the wireless wake-up signal from the thermal imaging module. The activation delay is ≤10ms. The positioning reflection points include: Hybrid positioning model, integrating RFID chip or low-power GPS module, the communication protocol has a unique code, TDOA mode is enabled on roads with speeds > 80km / h, and automatically switches to RSSI mode when the number of reflection points is < 3; The trigger mechanism is linked to the thermal imaging module, the housing meets the IP67 protection level, and the internal piezoelectric power generation module supports an operating temperature of -40°C to 85°C; Encryption logic: the reflection point coding uses a time-division multiplexing key, and the central unit dynamically decrypts it according to the section number; Enhanced reflective layer, the surface is covered with high reflectivity material and has a built-in LED flash unit; Central processing unit: used to receive the data from the thermal imaging module and the positioning signal of the positioning reflection point, realize dynamic calibration of pedestrian position through algorithm fusion and generate warning information, which is used to link electronic road signs, navigation apps or emergency lane indicators.

[0006] Preferably, the positioning reflection point is in normal sleep mode and is activated upon receiving an 868MHz or 915MHz wake-up signal, with an activation delay of ≤10ms.

[0007] Preferably, the weight distribution formula of TDOA and RSSI in the hybrid positioning model is: α=1 / (1+e^(-0.2v)), β=1-α, where v is the real-time vehicle speed in km / h; when the vehicle speed v>80km / h, the TDOA mode is enabled, and the weight α>0.9; when v<30km / h, it is switched to the RSSI mode, and the weight β>0.7.

[0008] Preferably, the positioning reflection point has a built-in frequency hopping module, the operating frequency band is 902-928 MHz, and a clean channel is dynamically selected through a signal-to-noise ratio threshold SNR>20 dB.

[0009] Preferably, the thermal imaging module integrates a ResNet-18 animal filtering model for filtering heat sources of small and medium-sized animals such as dogs.

[0010] Preferably, the positioning reflection point has a built-in channel scanning module, which switches the communication frequency band when SNR>20dB is detected.

[0011] The present invention also discloses a highway pedestrian warning method using thermal imaging to coordinately locate reflection points, comprising the following steps: S1: Generate the initial heat source trajectory through scanning by thermal imaging module; S2: Activate the positioning reflection points within a 20-meter radius of the initial heat source trajectory and receive positioning signals emitted by the positioning reflection points; S3: Calculate the precise coordinates of the pedestrian based on the hybrid positioning model of time-space alignment. The weight distribution formula of TDOA and RSSI in the hybrid positioning model is: α=1 / (1+e^(-0.2v)), β=1-α Where v is the real-time vehicle speed in km / h. When the vehicle speed v>80km / h, the TDOA mode is enabled with a weight α>0.9. When v<30km / h, the RSSI mode is switched to with a weight β>0.7. When the thermal imaging module identifies multiple heat sources, the corresponding reflection point cluster is activated according to the trajectory radius, and a one-to-one mapping between pedestrians and reflection points is achieved through code matching. Preferably, the spatiotemporal alignment method in step S3 is to perform interpolation compensation on the positioning signal based on the timestamp of the thermal imaging module.

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention adopts a dual-mode collaborative detection mechanism: thermal imaging and positioning reflection points to solve the problem of missed detection by a single technology.

[0013] 2. The present invention adopts a low-power reflection point design: the reflection point is normally dormant and is only activated after being triggered by thermal imaging, thereby extending the life of the device.

[0014] 3. The present invention adopts a positioning coding redundancy design: the reflection point adopts dynamic encryption coding to avoid signal interference. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the use process of the system of the present invention; DETAILED DESCRIPTION In the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," "connect," "fixed," etc. should be understood broadly. For example, they may refer to fixed connection, detachable connection, or integration; mechanical connection or electrical connection; direct connection or indirect connection through an intermediate medium; internal communication between two components or interaction between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0016] The following is a further description with reference to the accompanying drawings and specific embodiments: Example 1 System composition Thermal imaging module: Deployed along highways, it scans heat source signals in real time and preliminarily identifies pedestrian characteristics (such as body temperature and movement trajectory). The lens has a field of view of 120° and a temperature sensitivity of 0.1°C.

[0017] Reflection Point Positioning: Reflection point arrays are deployed at 50-meter intervals on guardrails, 1.2 meters high, overlapping the thermal imaging module's field of view and embedded within the reflectors on both sides of the highway. A "conditional trigger" design maintains normal dormancy, while dynamic encryption utilizes the SM4 national encryption algorithm. It activates only upon receiving a wireless wake-up signal (868MHz / 915MHz frequency band) from the thermal imaging module, with an activation delay of ≤10ms. The central unit synchronously updates the decryption key based on the timestamp. Includes: Hybrid positioning model: Integrates an RFID chip or low-power GPS module, and the communication protocol (such as LoRaWAN Class C NB-IoT backup communication link) has a unique code. TDOA mode is enabled on roads with speeds > 80 km / h, and automatically switches to RSSI mode when the number of reflection points is < 3.

[0018] Trigger Mechanism: Linked to the thermal imaging module, the reflector housing meets IP67 protection, and the internal piezoelectric generator module supports an operating temperature range of -40°C to 85°C. The reflector activates only when a pedestrian is detected, transmitting a positioning signal. Simulations have verified that a 10ms delay is effective. Actual measurements on the G4 Beijing-Hong Kong-Macao Expressway showed a pedestrian displacement error of 0.28±0.07m (n=152) with a 10ms delay.

[0019] Encryption logic example: The reflection point code uses a time-division multiplexing key, and the central unit dynamically decrypts it according to the section number.

[0020] Enhanced reflective layer: The surface is covered with high reflectivity material and has a built-in LED flash unit that actively illuminates the area for warnings at night.

[0021] Central processing unit: receives thermal imaging data and positioning signals, and integrates them through algorithms.

[0022] Dynamic calibration of pedestrian positions.

[0023] Generate early warning information: such as linked electronic road signs, navigation apps, and emergency lane indicator lights.

[0024] Workflow The thermal imaging module detected an abnormal heat source and initially determined it to be a pedestrian.

[0025] Activate the positioning reflection points in the area where pedestrians are located and transmit coded signals.

[0026] The central processing unit calculates the precise coordinates of pedestrians by matching signal strength with coding and based on a hybrid positioning model of TDOA (time difference of arrival) and RSSI (signal strength).

[0027] Synchronously activate the sound and light alarm (reflector LED flashes) and remote warning.

[0028] The steps are as follows: S1: Generate initial heat source trajectory through thermal imaging scanning.

[0029] S2: Activate the reflection points within a 20-meter radius of the track to receive positioning signals.

[0030] S3: A hybrid localization model based on spatiotemporal alignment, where: TDOA weight α=1 / (1+e^(-0.2v)).

[0031] RSSI weight β=1-α (v is the real-time vehicle speed).

[0032] When the vehicle speed v>80km / h, the TDOA mode is enabled (weight α>0.9), and when v<30km / h, it switches to the RSSI mode (weight β>0.7).

[0033] When the thermal imaging module identifies multiple heat sources, it activates the corresponding reflection point clusters according to the trajectory radius and achieves a one-to-one mapping between pedestrians and reflection points through coding matching.

[0034] Reflection point structure: The reflection point shell is made of weather-resistant engineering plastic, and is internally integrated with: Micro RFID chip (operating frequency 915MHz).

[0035] Light sensor (detects ambient light and automatically adjusts LED brightness).

[0036] Piezoelectric power module (dual piezoelectric ceramic redundant design, generates electricity using vibration from passing vehicles, maintenance-free).

[0037] System failure redundancy solution: When the thermal imaging module fails, the reflection point switches to the timed polling mode (activated once every 5 minutes) to ensure the availability of basic detection functions.

[0038] Algorithm flow: 1. Temporal and spatial alignment algorithm of thermal imaging data and positioning signals.

[0039] 2. Establish a heat source motion trajectory model.

[0040] 3. Use dynamic weight fusion algorithm to fuse multiple reflection point signals and eliminate positioning jitter.

[0041] 4. Introducing road boundary constraints through a dynamic weight fusion algorithm: When the positioning point deviates from the roadbed by more than 2 meters, a secondary thermal imaging verification is automatically triggered.

[0042] 5. Observation noise matrix R = diag ([0.5, 0.5]), process noise matrix Q = 0.1·I (identity matrix), state transition matrix A = [1, 0, Δt, 0; 0, 1, 0, Δt; 0, 0, 1, 0; 0, 0, 0, 1] (Δt = 0.1 s).

[0043] 6. The false alarm rate of the rainstorm test is 1.2%. Testing standard: Based on JT / T 1037-2016 "Highway Traffic Safety Facilities Inspection Procedure", continuous testing of 82km (Beijing-Hong Kong-Macao Expressway K1023-K1105 section).

[0044] 7. The core of the collaboration between thermal imaging and reflective points lies in error compensation: Thermal imaging has greater spatial errors in rainy and foggy days but has strong temporal continuity.

[0045] Reflection point positioning is less affected by weather but depends on device density.

[0046] The two types of errors are modeled as: Qthermal = 0.5·I, Qreflector = 0.1·I through the dynamic weight fusion algorithm. When the visibility is less than 50m, the Qthermal weight is dynamically adjusted to achieve adaptive compensation.

[0047] Animal false alarm filtering mechanism: The thermal imaging module integrates an AI recognition layer (ResNet-18 lightweight model) to filter heat sources of small and medium-sized animals such as dogs (false positive rate < 0.3%).

[0048] Extreme weather scenarios: In dense fog (visibility < 50 meters), the system compensates for the decrease in thermal imaging signal-to-noise ratio by increasing the LED flashing frequency (5Hz).

[0049] Anti-interference verification: In strong electromagnetic interference environments (such as near high-voltage towers), the reflection point switches to frequency hopping mode (902-928MHz) and uses an adaptive channel selection algorithm to ensure signal stability. Frequency hopping mode is achieved through adaptive channel selection: Scan the 902-928MHz band signal-to-noise ratio (SNR).

[0050] Select a clean channel with SNR>20dB.

[0051] Frequency hopping according to ISO / IEC 18000-6C standard The above is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A highway pedestrian warning system using thermal imaging and collaborative positioning of reflection points, characterized in that: include: Thermal imaging module: Deployed along highways, it scans heat source signals in real time and preliminarily identifies pedestrian features. The thermal imaging module has a 120° field of view and a temperature sensitivity of 0.1°C. Positioning reflection point array: Deployed at 50-meter intervals on highway guardrails, at a height of 1.2 meters, overlapping the field of view of the thermal imaging module and embedded in the reflectors on both sides of the highway. It adopts a conditional trigger design and is normally dormant, activating only after receiving the wireless wake-up signal from the thermal imaging module. The activation delay is ≤10ms. The positioning reflection points include: Hybrid positioning model, integrating RFID chip or low-power GPS module, the communication protocol has a unique code, TDOA mode is enabled on roads with speeds > 80km / h, and automatically switches to RSSI mode when the number of reflection points is < 3; The trigger mechanism is linked to the thermal imaging module, the housing meets the IP67 protection level, and the internal piezoelectric power generation module supports an operating temperature of -40°C to 85°C; Encryption logic: the reflection point coding uses a time-division multiplexing key, and the central unit dynamically decrypts it according to the section number; Enhanced reflective layer, the surface is covered with high reflectivity material and has a built-in LED flash unit; Central processing unit: used to receive the data from the thermal imaging module and the positioning signal of the positioning reflection point, realize dynamic calibration of pedestrian position through algorithm fusion and generate warning information, which is used to link electronic road signs, navigation apps or emergency lane indicators.

2. The highway pedestrian warning system using thermal imaging and collaborative positioning of reflection points according to claim 1 is characterized in that: The positioning reflection point is normally dormant and is activated by receiving an 868MHz or 915MHz wake-up signal, with an activation delay of ≤10ms.

3. The highway pedestrian warning system using thermal imaging and coordinated positioning of reflection points according to claim 2 is characterized in that: The weight distribution formula for TDOA and RSSI in the hybrid positioning model is: α=1 / (1+e^(-0.2v)), β=1-α, where v is the real-time vehicle speed in km / h; when the vehicle speed v>80km / h, the TDOA mode is enabled with a weight α>0.9; when v<30km / h, it switches to the RSSI mode with a weight β>0.

7.

4. The highway pedestrian warning system using thermal imaging and coordinated positioning of reflection points according to claim 1 is characterized in that: The positioning reflection point has a built-in frequency hopping module with an operating frequency band of 902-928MHz, and dynamically selects a clean channel through a signal-to-noise ratio threshold SNR>20dB.

5. The highway pedestrian warning system using thermal imaging and collaborative positioning of reflection points according to claim 1 is characterized in that: The thermal imaging module integrates the ResNet-18 animal filtering model to filter heat sources from small and medium-sized animals such as dogs.

6. The highway pedestrian warning system using thermal imaging and coordinated positioning of reflection points according to claim 1 is characterized in that: The positioning reflection point has a built-in channel scanning module, which switches the communication frequency band when SNR>20dB is detected.

7. The highway pedestrian warning method using thermal imaging and collaborative positioning of reflection points according to claim 1 is characterized in that: The following steps are involved: S1: Generate the initial heat source trajectory through scanning by thermal imaging module; S2: Activate the positioning reflection points within a 20-meter radius of the initial heat source trajectory and receive positioning signals emitted by the positioning reflection points; S3: Calculate the precise coordinates of pedestrians based on a hybrid positioning model based on time-space alignment. The weight distribution formula of TDOA and RSSI in the hybrid positioning model is: α=1 / (1+e^(-0.2v)), β=1-α Where v is the real-time vehicle speed in km / h. When the vehicle speed v>80 km / h, TDOA mode is enabled with a weight α>0.

9. When v<30 km / h, it switches to RSSI mode with a weight β>0.

7. When the thermal imaging module identifies multiple heat sources, it activates the corresponding reflection point cluster according to the trajectory radius, and achieves a one-to-one mapping between pedestrians and reflection points through code matching.

8. The method for early warning pedestrians on highways using thermal imaging and collaborative positioning of reflection points according to claim 7, characterized in that: The time-space alignment method in step S3 is to perform interpolation compensation on the positioning signal based on the timestamp of the thermal imaging module.

Images