CN-121980416-A - River basin non-point source pollution real-time monitoring method and system based on multi-source data fusion

CN121980416ACN 121980416 ACN121980416 ACN 121980416ACN-121980416-A

Abstract

The invention relates to the technical field of pollution monitoring, in particular to a river basin non-point source pollution real-time monitoring method and system based on multi-source data fusion. According to the invention, by constructing an empty, ground and water multi-source heterogeneous data fusion and progressive analysis framework and introducing a composite pollution risk index, and combining time sequence feature analysis, autocorrelation analysis and distribution features of high-risk pollution source units, the identification of pollution types can be further refined, accurate treatment and intervention are facilitated, the instantaneity, accuracy and scientificity of river basin surface source pollution monitoring are improved, and the problems that accurate real-time tracing of pollution, poor early warning pertinence and low supervision efficiency cannot be realized due to the fact that the change mixing treatment caused by regular water quality fluctuation and artificial pollution emission under seasonal climate driving are effectively solved.

Inventors

CHEN WEI
XIE XIAOYU
YANG QIANGBIN
TANG YUANZHE
Zhang Miaoda
ZHU LEI

Assignees

重庆文理学院

Dates

Publication Date: 20260505
Application Date: 20260115

Claims (10)

1. The river basin non-point source pollution real-time monitoring method based on multi-source data fusion is characterized by comprising the following steps of: Acquiring the ammonia nitrogen concentration of a water outlet of each unit to be evaluated in a river basin of a hilly agricultural area, the fluorescence intensity of blue algae associated with the water body, the surface flow velocity of the water body and the water level drop of the upstream and downstream sections of the sub-river basin; Identifying a plurality of important focusing units according to the threshold comparison result of the ammonia nitrogen concentration, determining a composite pollution risk index according to the blue algae fluorescence intensity, the water surface flow velocity and the water level drop of each important focusing unit, and determining a plurality of high-risk pollution source units according to the threshold comparison result of the composite pollution risk index; Determining a risk mode according to the blue algae fluorescence intensity, the ammonia nitrogen concentration and the time sequence characteristics of the water surface flow velocity of the high-risk pollution source unit so as to obtain an abnormal fluctuation mode, a seasonal regular mode and a random disturbance mode; acquiring a near infrared band reflectivity average value of the ground surface of the high-risk pollution source unit with the risk mode being the seasonal regular mode; Determining the pollution types of the high-risk pollution source units with the risk modes of the abnormal fluctuation mode and the random disturbance mode, and determining the pollution types according to the near infrared band reflectivity average value and the distribution characteristics of the high-risk pollution source units with the risk modes of the seasonal regular mode; Outputting a pollution early warning comprising the high risk pollution source unit and the pollution type.
2. The method for monitoring river basin surface source pollution in real time based on multi-source data fusion according to claim 1, wherein the steps of identifying a plurality of important attention units according to the threshold comparison result of ammonia nitrogen concentration, determining a composite pollution risk index according to the blue algae fluorescence intensity, the water surface flow velocity and the water level drop of each important attention unit, and determining a plurality of high risk pollution source units according to the threshold comparison result of the composite pollution risk index comprise: When the ammonia nitrogen concentration is greater than a preset concentration threshold value, determining the unit to be evaluated as the important focusing unit; Constructing a three-dimensional vector based on the normalized blue algae fluorescence intensity, the water surface flow velocity and the water level drop to obtain a plurality of current state vectors, and determining the composite pollution risk index according to the deviation degree of the current state vectors; and when the composite pollution risk index is larger than a preset risk threshold value, determining that the important focusing unit is the high-risk pollution source unit.
3. The method for monitoring river basin surface source pollution in real time based on multi-source data fusion according to claim 2, wherein determining a risk pattern according to the blue algae fluorescence intensity, the ammonia nitrogen concentration and the time sequence characteristics of the water surface flow rate of the high risk pollution source unit to obtain an abnormal fluctuation pattern, a seasonal regular pattern and a random disturbance pattern comprises: determining respective time sequence characteristic values according to the time sequence of the blue algae fluorescence intensity, the ammonia nitrogen concentration and the water surface flow velocity; Determining the risk mode of the high-risk pollution source unit according to the time sequence characteristic value to obtain the abnormal fluctuation mode, the seasonal regular mode and the random disturbance mode; the time sequence characteristic value comprises a comprehensive variation coefficient and a comprehensive variation amplitude.
4. The method for monitoring river basin non-point source pollution in real time based on multi-source data fusion according to claim 3, wherein the determining of the respective time sequence characteristic values according to the time sequence of the blue algae fluorescence intensity, the ammonia nitrogen concentration and the water surface flow rate comprises the following steps: respectively determining an intensity variation coefficient, a concentration variation coefficient and a flow velocity variation coefficient according to the relative discrete degree of the blue algae fluorescence intensity, the ammonia nitrogen concentration and the normalization processing result of the water surface flow velocity in a preset monitoring period; respectively determining intensity variation amplitude, concentration variation amplitude and flow rate variation amplitude according to average relative variation rates of the blue algae fluorescence intensity, the ammonia nitrogen concentration and the adjacent moments of the normalization processing result of the water surface flow rate in a preset monitoring period; And carrying out weighted summation on the intensity variation coefficient, the concentration variation coefficient and the flow velocity variation coefficient to obtain the comprehensive variation coefficient, and carrying out weighted summation on the intensity variation amplitude, the concentration variation amplitude and the flow velocity variation amplitude to obtain the comprehensive variation amplitude.
5. The method for real-time monitoring of river basin area source pollution based on multi-source data fusion according to claim 4, wherein the determining the risk pattern of the high risk pollution source unit according to the time sequence eigenvalue to obtain the abnormal fluctuation pattern, the seasonal regular pattern, and the random disturbance pattern comprises: when the comprehensive variation coefficient is larger than or equal to a preset variation threshold value or the comprehensive variation amplitude is smaller than or equal to a preset amplitude threshold value, confirming that the risk mode is the abnormal fluctuation mode; And when the risk mode is not the abnormal fluctuation mode, acquiring the blue algae fluorescence intensity, the ammonia nitrogen concentration and the water surface flow velocity, and determining that the risk mode is the seasonal regular mode or the random disturbance mode according to an autocorrelation analysis result of the acquired result.
6. The method for monitoring river basin area source pollution in real time based on multi-source data fusion according to claim 5, wherein the determining that the risk pattern is the seasonal regular pattern or the random disturbance pattern according to the autocorrelation analysis result of the obtained result comprises: respectively calculating the fluorescence intensity of the blue algae, the ammonia nitrogen concentration and the autocorrelation coefficient of the water surface flow velocity under a preset hysteresis period to obtain an intensity autocorrelation coefficient, a concentration autocorrelation coefficient and a flow velocity autocorrelation coefficient; Respectively determining an intensity confidence interval, a concentration confidence interval and a flow velocity confidence interval according to the respective statistical significance of the intensity autocorrelation coefficient, the concentration autocorrelation coefficient and the flow velocity autocorrelation coefficient; Determining that the risk pattern is the seasonal regular pattern when the intensity confidence interval does not include a preset correlation threshold, the concentration confidence interval does not include a preset correlation threshold, and the flow rate confidence interval does not include a preset correlation threshold; otherwise, determining the risk mode as the random disturbance mode.
7. The method for monitoring river basin area source pollution in real time based on multi-source data fusion according to claim 6, wherein the process of determining the pollution type of the high-risk pollution source unit with the risk mode being the abnormal fluctuation mode and the random disturbance mode and determining the pollution type according to the near infrared band reflectivity average value and the distribution characteristics of the high-risk pollution source unit with the risk mode being the seasonal regular mode comprises the following steps: Determining that the pollution type is exogenous interference when the risk pattern is the abnormal fluctuation pattern or the random disturbance pattern; When the risk mode is the seasonal regular mode, determining that the distribution type of the high-risk pollution source units is aggregation distribution or dispersion distribution according to a threshold screening result of a neighbor distance mean value determined by the spatial proximity characteristics of the high-risk pollution source units; calculating the relative deviation between the highest value and the lowest value of the near infrared band reflectivity mean value in a preset determination period to obtain a buffer attenuation rate; and determining the pollution type according to the aggregation distribution, the dispersion distribution and the coupling relation of the buffer attenuation rate.
8. The method for monitoring river basin area source pollution in real time based on multi-source data fusion according to claim 7, wherein the process of determining that the distribution type of the high-risk pollution source unit is aggregate distribution or scatter distribution according to the threshold screening result of the neighbor distance mean value determined by the spatial proximity feature of the high-risk pollution source unit comprises: when the neighbor distance average value is larger than a preset distance threshold value, determining that the distribution type is the dispersion distribution; and when the neighbor distance average value is smaller than or equal to a preset distance threshold value, determining the distribution type as the aggregation distribution.
9. The method for monitoring river basin area source pollution in real time based on multi-source data fusion according to claim 8, wherein the determining the pollution type according to the coupling relation of the aggregation distribution, the dispersion distribution and the buffer attenuation rate comprises: When the buffer attenuation rate is smaller than a preset attenuation threshold value and the distribution type is the aggregation distribution, determining that the pollution type is aggregation buffer deficiency type; And when the buffer attenuation rate is smaller than a preset attenuation threshold value and the distribution type is the dispersion distribution, determining that the pollution type is dispersion buffer type.
10. The river basin non-point source pollution real-time monitoring system based on multi-source data fusion is constructed based on the river basin non-point source pollution real-time monitoring method based on multi-source data fusion as set forth in any one of claims 1 to 9, and is characterized by comprising the following steps: the acquisition module is used for acquiring the ammonia nitrogen concentration of the water outlet of each unit to be evaluated in the river basin of the hilly agricultural area, the blue algae fluorescence intensity of the associated water body, the water surface flow velocity, the water level drop of the upstream and downstream sections of the sub-river basin, and the near infrared band reflectivity average value of the ground surface; The identification module is used for identifying a plurality of important focusing units according to the threshold value comparison result of the ammonia nitrogen concentration, determining a composite pollution risk index according to the blue algae fluorescence intensity, the water surface flow velocity and the water level drop of each important focusing unit, and determining a plurality of high-risk pollution source units according to the threshold value comparison result of the composite pollution risk index; the mode determining module is used for determining a risk mode according to the blue algae fluorescence intensity, the ammonia nitrogen concentration and the time sequence characteristics of the water surface flow velocity of the high-risk pollution source unit so as to obtain an abnormal fluctuation mode, a seasonal regular mode and a random disturbance mode; the type determining module is used for determining the pollution type of the high-risk pollution source unit with the risk mode being the abnormal fluctuation mode and the random disturbance mode, and determining the pollution type according to the near infrared band reflectivity average value and the distribution characteristics of the high-risk pollution source unit with the risk mode being the seasonal regular mode; And the output module is used for outputting pollution early warning comprising the high-risk pollution source unit and the pollution type.

Description

River basin non-point source pollution real-time monitoring method and system based on multi-source data fusion Technical Field The invention relates to the technical field of pollution monitoring, in particular to a river basin non-point source pollution real-time monitoring method and system based on multi-source data fusion. Background With the continuous improvement of the agricultural development intensity in hilly areas, the agricultural activities such as hilly lands, fruit trees, and scattered livestock and poultry cultivation are deeply coupled with the fragile mountain hydrologic environment, so that increasingly severe pressure is formed on the water environment of the flow field, and the precise and efficient pollution control is becoming urgent management demand. In a typical small agricultural watershed in hills, the hydrologic process is dominated by terrains and rainfall, is complex and changeable, has various pollution sources, and particularly has water quality fluctuation driven by natural factors such as seasonal concentrated rainfall, snow melting and the like, and is mutually overlapped with the agricultural production activity emission, so that the accurate tracing of human pollution sources and the implementation of targeted treatment face significant challenges. At present, how to effectively screen and strip regular natural disturbance signals driven by the rainfall runoff of the terrain in massive and dynamic monitoring data so as to accurately lock real agricultural non-point source pollution contribution in real time is a core problem of improving the effectiveness of the non-point source pollution treatment of hilly areas and guaranteeing the water ecological safety of areas. The Chinese patent application publication No. CN121211104A discloses a method and a system for on-line monitoring of river basin agricultural non-point source pollution, wherein the method comprises the steps of arranging a plurality of monitoring nodes in a river basin, obtaining a monitoring data set through the plurality of monitoring nodes, preprocessing the monitoring data set, establishing a river basin agricultural non-point source pollution monitoring model, wherein the river basin agricultural non-point source pollution monitoring model comprises a network model and a supervision model, the network model is used for outputting an abnormal detection result, the supervision model is used for generating a preliminary abnormal detection feature, integrating the preliminary abnormal detection feature with the abnormal detection result to obtain the output of the river basin agricultural non-point source pollution monitoring model, carrying out real-time early warning on the agricultural non-point source pollution in the river basin according to the established river basin agricultural non-point source pollution monitoring model and a preset early warning threshold, and sending early warning information when the monitored data exceeds the early warning threshold. Therefore, the method and the system for on-line monitoring of the river basin agricultural non-point source pollution have the following problems that the method adopts generalized monitoring parameters and static early warning logic and lacks scene specific design, false or missing report is easily generated due to incapability of adapting to different seasons and hydrologic conditions, regular natural disturbance in time sequence data cannot be effectively analyzed by the method, monitoring conclusion is easy to be inaccurate, the method depends on limited historical experience labeling on abnormal definition, and a novel or unknown pollution emission mode cannot be effectively identified by a model. Disclosure of Invention Therefore, the invention provides a river basin non-point source pollution real-time monitoring method and system based on multi-source data fusion, which are used for solving the problems that in the prior art, accurate real-time tracing of pollution, poor early warning pertinence and low supervision efficiency cannot be realized due to the fact that the change mixing treatment caused by the regular water quality fluctuation driven by seasonal climate and the artificial pollution emission is carried out through multi-source parameter fusion and time sequence analysis-based stripping of seasonal fluctuation. In order to achieve the above object, in one aspect, the present invention provides a method for monitoring river basin non-point source pollution in real time based on multi-source data fusion, comprising: Acquiring the ammonia nitrogen concentration of a water outlet of each unit to be evaluated in a river basin of a hilly agricultural area, the fluorescence intensity of blue algae associated with the water body, the surface flow velocity of the water body and the water level drop of the upstream and downstream sections of the sub-river basin; Identifying a plurality of important focusing units according to the thr