A Highway Monitoring System Based on GIS

Abstract

The invention discloses a highway monitoring system based on GIS, comprising a data acquisition module and a data processing module; the data acquisition module is used for data acquisition in a low-visibility environment of a highway; the data acquisition module comprises: an image acquisition unit, a humidity sensor unit, a satellite remote sensing data receiving unit, and an unmanned aerial vehicle inspection unit; the data processing module is used for processing the data acquired by the data acquisition module, combining the data acquired by the data acquisition module to form a feature space vector X, calculating the distance between the vector X and a pre-stored standard feature vector sample set of agglomerate fog, fog and haze, respectively allocating weights according to the feature vectors, and calculating the weights allocated according to the calculated feature vectors; if the value of D_fog is the smallest, it is fog; if the value of D_haze is the smallest, it is haze; the purpose of the invention is to realize the accurate monitoring and judgment of agglomerate fog, fog and haze on a highway.

Description

Technical Field

[0001] The invention belongs to the technical field of traffic monitoring and relates to a highway monitoring system based on GIS. Background Art

[0002] As an important artery of modern transportation, the safe and efficient operation of highways is of key significance to economic development and social life. However, severe weather such as fog, mist, and haze seriously affect the traffic capacity and traffic safety of highways. To meet this challenge, GIS-based highway monitoring systems have gradually become the focus of research and application, aiming to achieve accurate monitoring and timely warning of these weather phenomena through advanced technical means.

[0003] In actual production and application, the existing highway monitoring system faces many problems. For the monitoring of fog, mist, and haze, data collection is not comprehensive and accurate enough, and it is difficult to capture their subtle features and rapid changes, leading to misjudgment. Delays, interruptions, or data loss are prone to occur during data transmission, affecting the timeliness of early warning. In addition, the existing early warning methods are relatively simple and cannot accurately convey the severity and scope of weather conditions, resulting in drivers being unable to take effective countermeasures, thereby increasing the risk of traffic accidents.

[0004] It was found that there are mainly the following solutions at present: some systems use more advanced sensor technologies, such as lidar, millimeter-wave radar, etc., which improve the accuracy of data collection. At the same time, big data analysis and intelligent algorithms are used to make early warning decisions, which enhances the accuracy of early warnings, but has high requirements on computing resources.

[0005] The advantage of these solutions is that they improve certain aspects of monitoring and early warning performance, but they also have disadvantages, such as high cost, limited scope of application, and high resource consumption. Summary of the invention

[0006] The present invention provides a highway monitoring system based on GIS, which can realize the monitoring and judgment of fog, mist and haze on the highway.

[0007] In order to solve the above problems, the technical solution adopted by the invention is:

[0008] A highway monitoring system based on GIS, including a data acquisition module and a data processing module;

[0009] The data acquisition module is used for data acquisition in low-visibility environments on expressways;

[0010] The data acquisition module comprises:

[0011] An image acquisition unit, used for acquiring image data of the highway environment through cameras arranged along the highway;

[0012] The image data includes texture, color and dynamic change feature information of fog, mist and haze in the image;

[0013] Humidity sensor unit, used to collect humidity data in highway environment;

[0014] A satellite remote sensing data receiving unit, used to receive satellite remote sensing data covering the highway area, wherein the satellite remote sensing data includes spectral characteristics and spatial distribution characteristics of fog, mist and haze;

[0015] A drone inspection unit is used for inspections over highways. The drone inspection unit carries meteorological sensors and particle detection equipment to collect characteristic data of fog, mist and haze;

[0016] The texture feature data includes image grayscale, variance, contrast, and energy; the color features include hue, saturation, and brightness; the shape and dynamic features include visibility, brightness, and contour fuzziness;

[0017] The collected data of the humidity sensor includes humidity threshold and humidity change rate;

[0018] The satellite remote sensing spectral characteristics and spatial distribution characteristics information include visible light reflectance, near infrared wave reflectance, absorptivity of water vapor absorption band, and regional area size and discreteness;

[0019] The characteristic data collected by the drone inspection unit include PM2.5 concentration, PM10 concentration, and water vapor condensation state parameters;

[0020] The data processing module is used to process the data collected by the data acquisition module, combine the data collected by the data acquisition module to form a feature space vector X, and calculate the distance between the feature space vector X collected in real time and the pre-stored standard feature vector sample set {X_fog, X_fog, X_haze} of group fog, fog and haze through the Euclidean distance; the Euclidean distance calculation formula is: for two vectors A=[a1,a2,…,an] and B=[b1,b2,…,bn], the distance d(A,B)= ; Calculate d(X, X_fog bank), d(X, X_fog), and d(X, X_haze) respectively, and judge the similarity according to the calculated distances. The smaller the distance, the higher the similarity. If d(X, X_fog bank) < d(X, X_fog) and d(X, X_fog bank) < d(X, X_haze), then the current monitoring data is more inclined to the fog bank feature; conversely, if d(X, X_fog) is the smallest, then it is more inclined to fog; if d(X, X_haze) is the smallest, then it is more inclined to haze; Assign weights to the feature vectors of the camera, humidity sensor, satellite remote sensing, and drone patrol respectively. Denote the camera feature weight as w1, the humidity sensor feature weight as w2, the satellite remote sensing feature weight as w3, and the drone patrol feature weight as w4, and satisfy w1 + w2 + w3 + w4 = 1. And the specific values of the weights w1, w2, w3, w4 are determined based on the importance and reliability of each data collection method in judging fog bank, fog, and haze through a large number of experiments and data analyses. For the fog bank, calculate the weighted distance D_fog bank = w1 × d(X_camera, X_fog bank_camera) + w2 × d(X_humidity, X_fog bank_humidity) + w3 × d(X_remote sensing, X_fog bank_remote sensing) + w4 × d(X_patrol, X_fog bank_patrol), where X_camera and X_fog bank_camera are the current monitoring feature vector collected by the camera and the standard feature vector of the fog bank respectively; X_humidity and X_fog bank_humidity are the current monitoring humidity feature vector collected by the humidity sensor and the standard humidity feature vector of the fog bank respectively; X_remote sensing and X_fog bank_remote sensing are the current monitoring remote sensing feature vector collected by the satellite remote sensing and the standard remote sensing feature vector of the fog bank respectively; X_patrol and X_fog bank_patrol are the current monitoring patrol feature vector collected by the drone patrol and the standard patrol feature vector of the fog bank respectively. Similarly, calculate D_fog and D_haze, and compare the magnitudes of the calculated weighted distances of D_fog bank, D_fog, and D_haze. When the value of D_fog bank is the smallest, it is judged that the current monitoring situation is fog bank; if the value of D_fog is the smallest, it is fog; if the value of D_haze is the smallest, it is haze.

[0021] The principle of this solution is as follows: By collecting data on fog, fog bank, and haze on the highway and analyzing and processing these data, and then analyzing the texture, color, and dynamic change characteristics of the image data, the numerical values and change trends of the humidity data, the spectral and spatial characteristics of the satellite remote sensing data, and the data characteristics of the drone patrol data, use the Euclidean distance and weighted distance to calculate the magnitude of the result value to judge whether it is fog bank, fog, or haze. When the value of D_fog bank is the smallest, it is judged that the current monitoring situation is fog bank; if the value of D_fog is the smallest, it is fog; if the value of D_haze is the smallest, it is haze.

[0022] The beneficial effects produced by this solution:

[0023] Through image acquisition, humidity sensors, satellite remote sensing and drone inspections, various information related to low-visibility environments on highways can be obtained comprehensively and accurately, which breaks the limitations of traditional single data sources and deeply understands the characteristics of fog, mist and haze from multiple angles; the rich data collected is carefully extracted and processed, such as the analysis of texture, color, shape dynamics and other features, as well as humidity, remote sensing and drone inspection data to improve the accuracy of judging low-visibility weather phenomena; the Euclidean distance and weighted calculation methods are used to judge the weather type, and the similarity between different feature vectors is measured, and the weights are assigned to the acquisition modules, so that the judgment results can more accurately identify whether it is fog, mist or haze, and at the same time can accurately convey the severity and scope of the weather conditions, so that drivers can take effective countermeasures.

[0024] Furthermore, the camera of the image acquisition unit is arranged on a street light pole, a gantry or a special monitoring tower on the highway.

[0025] The humidity sensor unit includes a plurality of humidity sensors, which are arranged at intervals along the highway and are densely arranged at bridges, tunnel entrances and exits, and valleys.

[0026] The satellite remote sensing data receiving unit includes a data receiving antenna and a data processing subunit. The data receiving antenna can receive multi-spectral satellite remote sensing data. The data processing subunit can perform preliminary processing on the received data, filter out data related to the current monitoring task through preliminary processing, and classify them according to different characteristics or bands.

[0027] The meteorological sensors carried by the drone include temperature sensors, humidity sensors, and particle detection equipment.

[0028] Furthermore, the data processing module includes:

[0029] The data receiving submodule is used to receive the image data, humidity data, satellite remote sensing data and drone inspection data collected by the collection module.

[0030] The data preprocessing submodule is used to perform format conversion, noise removal and data normalization on the received data.

[0031] The feature extraction submodule is used to extract the texture features, color features, humidity features, spectral features, spatial distribution features, microscopic physical features and vertical structure features of fog, mist and haze from the preprocessed data.

[0032] The identification and judgment submodule is used to determine whether the collected data corresponds to fog, mist or haze based on the extracted features and using a preset recognition algorithm.

[0033] Furthermore, the data receiving submodule is compatible with data transmission protocols of different types of acquisition devices, including wired transmission protocols and wireless transmission protocols.

[0034] Furthermore, the noise removal method adopted by the data preprocessing submodule includes median filtering, mean filtering or wavelet denoising method.

[0035] Furthermore, the feature extraction submodule adopts a gray level co-occurrence matrix method, a local binary pattern method or a Gabor filter method when extracting texture features.

[0036] Furthermore, the color features in the data acquisition module adopt a method of analyzing hue, saturation and brightness in the HSV color space.

[0037] Furthermore, the data acquisition module utilizes multi-spectral analysis and spatial analysis methods of satellite remote sensing data when extracting spectral features and spatial distribution features. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 System flow chart of the present invention. DETAILED DESCRIPTION

[0039] The embodiment is basically as shown in the attached Figure 1 As shown, a highway monitoring system based on GIS includes a data acquisition module and a data processing module.

[0040] The data acquisition module is used for data acquisition in low visibility environments on expressways; the data acquisition module includes:

[0041] The data acquisition module comprises:

[0042] An image acquisition unit is used to collect image data of the highway environment through cameras arranged along the highway; the image data includes texture, color and dynamic change feature information of fog, mist and haze in the image;

[0043] Humidity sensor unit, used to collect humidity data in highway environment;

[0044] A satellite remote sensing data receiving unit, used for receiving satellite remote sensing data covering the highway area;

[0045] The satellite remote sensing data includes spectral characteristics and spatial distribution characteristics of fog, mist and haze;

[0046] The unmanned aerial vehicle inspection unit is used for inspection over the highway. The unmanned aerial vehicle inspection unit carries a meteorological sensor and a particle detection device to collect characteristic data of fog, mist and haze; the texture feature data includes image grayscale, variance, contrast and energy; the color features include hue, saturation and brightness value; the shape and dynamic features include visibility, brightness and contour fuzziness; the collected data of the humidity sensor includes humidity threshold and humidity change rate; the satellite remote sensing features include visible light reflectance, near infrared wave reflectance, water vapor absorption band absorptivity, as well as regional area size and discreteness; the data collected by the unmanned aerial vehicle inspection unit includes PM2.5 concentration, PM10 concentration and water vapor condensation state parameters;

[0047] The data processing module is used to process the data collected by the data collection module, and combine the data collected by the data collection module to form a feature space vector. Through the Euclidean distance, for the feature vector X collected in real time, the distance between it and the pre-stored standard feature vector sample set {X_fog, X_fog, X_haze} of group fog, fog and haze is calculated. The Euclidean distance calculation formula is: for two vectors A=[a1,a2,an] and B=[b1,b2,…,bn], the distance d(A,B)= Calculate d(X, X_fog bank), d(X, X_fog), and d(X, X_haze) respectively, and judge the similarity according to the calculated distances. The smaller the distance, the higher the similarity. If d(X, X_fog bank) < d(X, X_fog) and d(X, X_fog bank) < d(X, X_haze), then the current monitoring data is more inclined to the fog bank characteristics; conversely, if d(X, X_fog) is the smallest, then it is more inclined to fog; if d(X, X_haze) is the smallest, then it is more inclined to haze. Assign weights to the feature vectors of cameras, humidity sensors, satellite remote sensing, and drone inspections respectively. Denote the camera feature weight as w1, the humidity sensor feature weight as w2, the satellite remote sensing feature weight as w3, and the drone inspection feature weight as w4, and satisfy w1 + w2 + w3 + w4 = 1. And the specific values of the weights w1, w2, w3, w4 are determined through a large number of experiments and data analyses according to the importance and reliability of each data collection method in judging fog bank, fog, and haze. For the fog bank, calculate the weighted distance D_fog bank = w1 × d(X_camera, X_fog bank_camera) + w2 × d(X_humidity, X_fog bank_humidity) + w3 × d(X_remote sensing, X_fog bank_remote sensing) + w4 × d(X_inspection, X_fog bank_inspection), where X_camera and X_fog bank_camera are the current monitoring feature vector collected by the camera and the standard feature vector of the fog bank respectively; X_humidity and X_fog bank_humidity are the current monitoring humidity feature vector collected by the humidity sensor and the standard humidity feature vector of the fog bank respectively; X_remote sensing and X_fog bank_remote sensing are the current monitoring remote sensing feature vector collected by satellite remote sensing and the standard remote sensing feature vector of the fog bank respectively; X_inspection and X_fog bank_inspection are the current monitoring inspection feature vector collected by drone inspection and the standard inspection feature vector of the fog bank respectively. Similarly, calculate D_fog and D_haze, and compare the magnitudes of the calculated weighted distances of D_fog bank, D_fog, and D_haze. When the value of D_fog bank is the smallest, then judge that the current monitoring situation is a fog bank; if the value of D_fog is the smallest, then it is fog; if the value of D_haze is the smallest, then it is haze.

[0048] By collecting data of fog, fog bank, and haze on highways, analyzing and processing these data, and then analyzing the texture, color, and dynamic change characteristics of image data, the numerical values and change trends of humidity data, the spectral and spatial characteristics of satellite remote sensing data, and the data characteristics of drone inspection data, use the Euclidean distance to judge whether the collected data corresponds to a fog bank, fog, or haze.

[0049] Through image acquisition, humidity sensors, satellite remote sensing and drone inspections, various information related to low-visibility environments on highways can be obtained comprehensively and accurately, which breaks the limitations of traditional single data sources and deeply understands the characteristics of fog, mist and haze from multiple angles; the rich data collected is carefully extracted and processed, such as the analysis of texture, color, shape dynamics and other features, as well as humidity, remote sensing and drone inspection data to improve the accuracy of judging low-visibility weather phenomena; the Euclidean distance and weighted calculation method are used to judge the weather type, which not only considers the similarity between different feature vectors, but also assigns weights to the acquisition modules, so that the judgment result can more accurately identify whether the current weather condition is fog, mist or haze.

[0050] The camera of the image acquisition unit is arranged on a street light pole, a gantry or a special monitoring tower of the expressway.

[0051] The humidity sensor unit includes a plurality of humidity sensors, which are arranged at intervals along the highway and are densely arranged at special terrain locations such as bridges, tunnel entrances and exits, and valleys where clusters of fog and mist are easily formed.

[0052] The satellite remote sensing data receiving unit comprises a data receiving antenna and a data processing subunit. The data receiving antenna can receive multi-spectral satellite remote sensing data, and the data processing subunit can perform preliminary processing on the received data.

[0053] The meteorological sensors carried by the drone include temperature sensors, humidity sensors, and particle detection equipment.

[0054] The data processing module comprises:

[0055] A data receiving submodule, used to receive the image data, humidity data, satellite remote sensing data and drone inspection data collected by the acquisition module;

[0056] The data preprocessing submodule is used to convert the format of the received data, remove noise and perform data normalization;

[0057] The feature extraction submodule is used to extract the texture features, color features, humidity features, spectral features, spatial distribution features, microscopic physical features and vertical structure features of fog, mist and haze from the preprocessed data;

[0058] The identification and judgment submodule is used to determine whether the collected data corresponds to fog, mist or haze based on the extracted features and using a preset recognition algorithm.

[0059] The data receiving submodule is compatible with data transmission protocols of different types of acquisition devices, including wired transmission protocols and wireless transmission protocols.

[0060] The noise removal method adopted by the data preprocessing submodule includes median filtering, mean filtering or wavelet denoising method.

[0061] The feature extraction submodule adopts a gray level co-occurrence matrix method, a local binary pattern method or a Gabor filter method when extracting texture features.

[0062] The color features in the data acquisition module are analyzed by the method of hue, saturation and brightness in the HSV color space.

[0063] The data acquisition module uses multi-spectral analysis and spatial analysis methods of satellite remote sensing data to extract spectral features and spatial distribution features. Specific embodiment 2

[0065] 1 Collect camera image feature vector

[0066] 1.1 For texture features, the contrast, correlation, energy and entropy values ​​are combined into a sub-vector. The image grayscale is set to A, the variance value is set to B, the contrast is set to C, the energy is set to D, and the texture feature sub-vector is set to [A, B, C, D].

[0067] 1.2 In terms of color features, the hue, saturation and lightness values ​​in the HSV color space are combined into a sub-vector. The hue value is E, the saturation value is F, and the lightness value is G. The color feature sub-vector is [E, F, G].

[0068] In 1.3 shape and dynamic features, the geometric parameters of the cloud shape and the dynamically changing speed parameters are combined into a sub-vector. Assuming that the visibility is H, the brightness is I, and the contour blur is J, the shape and dynamic feature sub-vector is [H, I, J].

[0069] 2. Collect humidity sensor feature vector

[0070] 2.1 Combine the humidity threshold feature RH and the humidity change rate feature to form the humidity sensor feature vector [RH, ΔRH].

[0071] 3. Collect satellite remote sensing feature vectors

[0072] 3.1 In terms of spectral characteristics, the reflectivity N in the visible light band, the reflectivity O in the near-infrared band, and the absorptivity P in the water vapor absorption band constitute the spectral characteristic sub-vector [N, O, P].

[0073] In 3.2 spatial distribution characteristics, the area size (R) and the degree of dispersion S of the fog, mist or haze coverage area are combined into a sub-vector [R, S].

[0074] 3.3 The satellite remote sensing feature vector is [N, O, P, R, S].

[0075] 4. Collect UAV inspection feature vectors

[0076] In 4.1 microscopic physical characteristics, the particle matter concentration PM2.5 concentration T, PM10 concentration U, and water vapor condensation state parameter V are combined into a sub-vector T, U, V].

[0077] 4.3 The feature vector of drone inspection is [T,U,V,]

[0078] All feature vectors are combined to form a feature space vector. This feature space contains all the features extracted by cameras, humidity sensors, satellite remote sensing and drone inspections. The total feature space vector is in the form of [A, B, C, D, E, F, G, H, I, J, RH, ΔRH, N, O, P, Q, R, S, T, U, V]. Through the Euclidean distance, for the feature vector X collected in real time, the pre-stored group fog feature vector is X-fog = , X-fog = , X-haze = ; Calculate its distance from the pre-stored standard feature vector sample set of cluster fog, fog and haze {X_cluster fog, X_fog, X_haze}; The real-time vector X = [A, B, C, D, E, F, G, H, I, J, RH, ΔRH, N, O, P, Q, R, S, T, U, V], and the Euclidean distance d (X, X-cluster fog) from the cluster fog feature vector = =

[0079] Similarly, d(X, X_fog) and d(X, X_haze) can be obtained. The similarity is judged based on the calculated distances. The smaller the distance, the higher the similarity. If d(X, X_radiation fog) < d(X, X_fog) and d(X, X_radiation fog) < d(X, X_haze), then the current monitoring data is more inclined to the characteristics of radiation fog; conversely, if d(X, X_fog) is the smallest, it is more inclined to fog; if d(X, X_haze) is the smallest, it is more inclined to haze. Weights are assigned to the feature vectors of the camera, humidity sensor, satellite remote sensing, and drone patrol respectively. Denote the weight of the camera feature as w1, the weight of the humidity sensor feature as w2, the weight of the satellite remote sensing feature as w3, and the weight of the drone patrol feature as w4, and satisfy w1 + w2 + w3 + w4 = 1. The specific values of the weights w1, w2, w3, and w4 are determined based on the importance and reliability of each data collection method in judging radiation fog, fog, and haze through a large number of experiments and data analyses. For radiation fog, calculate the weighted distance D_radiation fog = w1 × d(X_camera, X_radiation fog_camera) + w2 × d(X_humidity, X_radiation fog_humidity) + w3 × d(X_remote sensing, X_radiation fog_remote sensing) + w4 × d(X_patrol, X_radiation fog_patrol), where X_camera and X_radiation fog_camera are the current monitoring feature vector collected by the camera and the standard feature vector of radiation fog respectively; X_humidity and X_radiation fog_humidity are the current monitoring humidity feature vector collected by the humidity sensor and the standard humidity feature vector of radiation fog respectively; X_remote sensing and X_radiation fog_remote sensing are the current monitoring remote sensing feature vector collected by the satellite remote sensing and the standard remote sensing feature vector of radiation fog respectively; X_patrol and X_radiation fog_patrol are the current monitoring patrol feature vector collected by the drone patrol and the standard patrol feature vector of radiation fog respectively. Similarly, calculate D_fog and D_haze, and compare the magnitudes of D_radiation fog, D_fog, and D_haze. When the value of D_radiation fog is the smallest, it is judged that the current monitoring situation is radiation fog; if the value of D_fog is the smallest, it is fog; if the value of D_haze is the smallest, it is haze.

[0080] The above are only embodiments of the present invention. Common knowledge such as the known specific structures and characteristics in the scheme is not described in detail here. Ordinary technicians in the relevant field are aware of all the common technical knowledge in the technical field to which the invention belongs before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply conventional experimental means before that date. Ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, several deformations and improvements can be made without departing from the structure of the present invention, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A highway monitoring system based on GIS, characterized in that: Including data acquisition module and data processing module; The data acquisition module is used for data acquisition in low-visibility environments on expressways; The data acquisition module comprises: An image acquisition unit, used for acquiring image data of the highway environment through cameras arranged along the highway; The image data includes texture, color and dynamic change feature information of fog, mist and haze in the image; Humidity sensor unit, used to collect humidity data in highway environment; A satellite remote sensing data receiving unit, used to receive satellite remote sensing data covering the highway area, wherein the satellite remote sensing data includes spectral characteristics and spatial distribution characteristics of fog, mist and haze; A drone inspection unit is used for inspections over highways. The drone inspection unit carries meteorological sensors and particle detection equipment to collect characteristic data of fog, mist and haze; The texture feature data includes image grayscale, variance, contrast, and energy; the color features include hue, saturation, and brightness; the shape and dynamic features include visibility, brightness, and contour fuzziness; The collected data of the humidity sensor includes humidity threshold and humidity change rate; The satellite remote sensing spectral characteristics and spatial distribution characteristics information include visible light reflectance, near infrared wave reflectance, absorptivity of water vapor absorption band, and regional area size and discreteness; The characteristic data collected by the drone inspection unit include PM2.5 concentration, PM10 concentration, and water vapor condensation state parameters; The data processing module is used to process the data collected by the data acquisition module, combine the data collected by the data acquisition module to form a feature space vector X, and calculate the distance between the feature space vector X collected in real time and the standard feature vector sample sets {X_fog bank, X_fog, X_haze} of fog bank, fog and haze through the Euclidean distance. The Euclidean distance calculation formula is: for two vectors A = [a1, a2,..., an] and B = [b1, b2,..., bn], the distance d(A, B) = ; calculate d(X, X_fog bank), d(X, X_fog) and d(X, X_haze) respectively, and judge the similarity according to the calculated distances. The smaller the distance, the higher the similarity. If d(X, X_fog bank) < d(X, X_fog) and d(X, X_fog bank) < d(X, X_haze), then the current monitoring data is more inclined to the fog bank feature; conversely, if d(X, X_fog) is the smallest, then it is more inclined to fog; if d(X, X_haze) is the smallest, then it is more inclined to haze; assign weights to the feature vectors of the camera, humidity sensor, satellite remote sensing and drone inspection respectively. Denote the camera feature weight as w1, the humidity sensor feature weight as w2, the satellite remote sensing feature weight as w3, and the drone inspection feature weight as w4, and satisfy w1 + w2 + w3 + w4 = 1. And the specific values of the weights w1, w2, w3, w4 are determined through a large number of experiments and data analyses according to the importance and reliability of each data acquisition method in judging fog bank, fog and haze. For fog bank, calculate the weighted distance D_fog bank = w1 × d(X_camera, X_fog bank_camera) + w2 × d(X_humidity, X_fog bank_humidity) + w3 × d(X_remote sensing, X_fog bank_remote sensing) + w4 × d(X_inspection, X_fog bank_inspection), where X_camera and X_fog bank_camera are the current monitoring feature vector collected by the camera and the standard feature vector of fog bank respectively; X_humidity and X_fog bank_humidity are the current monitoring humidity feature vector collected by the humidity sensor and the standard humidity feature vector of fog bank respectively; X_remote sensing and X_fog bank_remote sensing are the current monitoring remote sensing feature vector collected by satellite remote sensing and the standard remote sensing feature vector of fog bank respectively; X_inspection and X_fog bank_inspection are the current monitoring inspection feature vector collected by drone inspection and the standard inspection feature vector of fog bank respectively. Similarly, calculate D_fog and D_haze, and compare the magnitudes of the calculated weighted distances D_fog bank, D_fog and D_haze. When the value of D_fog bank is the smallest, then judge that the current monitoring situation is fog bank; if the value of D_fog is the smallest, then it is fog; if the value of D_haze is the smallest, then it is haze.

2. The GIS-based highway monitoring system according to claim 1 is characterized in that: The camera of the image acquisition unit is arranged on a street light pole, a gantry or a special monitoring tower of the highway; The humidity sensor unit includes a plurality of humidity sensors, which are arranged at intervals along the highway and are densely arranged at bridges, tunnel entrances and exits, and valley locations; The satellite remote sensing data receiving unit includes a data receiving antenna and a data processing subunit, wherein the data receiving antenna can receive multi-spectral satellite remote sensing data, and the data processing subunit can perform preliminary processing on the received data, filter out data related to the current monitoring task through preliminary processing, and classify the data according to different characteristics or bands; The meteorological sensors carried by the drone include temperature sensors, humidity sensors, and particle detection equipment.

3. The GIS-based highway monitoring system according to claim 1 is characterized in that: The data processing module comprises: A data receiving submodule, used to receive the image data, humidity data, satellite remote sensing data and drone inspection data collected by the acquisition module; The data preprocessing submodule is used to convert the format of the received data, remove noise and perform data normalization; The feature extraction submodule is used to extract the texture features, color features, humidity features, spectral features, spatial distribution features, microscopic physical features and vertical structure features of fog, mist and haze from the preprocessed data; The identification and judgment submodule is used to determine whether the collected data corresponds to fog, mist or haze based on the extracted features and using a preset recognition algorithm.

4. The GIS-based highway monitoring system according to claim 3 is characterized in that: The data receiving submodule is compatible with data transmission protocols of different types of acquisition devices, including wired transmission protocols and wireless transmission protocols.

5. The GIS-based highway monitoring system according to claim 3 is characterized in that: The noise removal method adopted by the data preprocessing submodule includes median filtering, mean filtering or wavelet denoising method.

6. The GIS-based highway monitoring system according to claim 3 is characterized in that: The feature extraction submodule adopts a gray level co-occurrence matrix method, a local binary pattern method or a Gabor filter method when extracting texture features.

7. The GIS-based highway monitoring system according to claim 1 is characterized in that: The data acquisition module extracts color features by analyzing hue, saturation and brightness in the HSV color space.

8. The GIS-based highway monitoring system according to claim 1 is characterized in that: The data acquisition module uses multi-spectral analysis and spatial analysis methods of satellite remote sensing data to extract spectral features and spatial distribution features.

Images