Method for evaluating sensory quality of urban water body supplemented by reclaimed water

Abstract

The present application relates to a kind of evaluation methods of the sensory quality of urban water body supplemented by reclaimed water.The present application belongs to the field of water resources, which integrates multispectral imaging, electronic nose, electronic tongue and basic water quality sensing and other multiple technologies to realize the synchronous collection and analysis of water color, odor, taste and physicochemical parameters, and establishes a label database combined with subjective evaluation of city. Through data fusion and intelligent analysis model, the sensory quality of water body can be quickly and objectively evaluated, abnormal changes can be identified in time, and the risk of abnormality can be traced and optimized to provide effective technical support for urban water environment management and safe use of reclaimed water.

Description

Technical Field

[0001] This invention belongs to the field of water resources, and more specifically relates to a method for evaluating the sensory quality of urban water bodies replenished by reclaimed water. Background Technology

[0002] With the accelerating pace of urbanization, the number and area of ​​urban water bodies are increasing daily, undertaking multiple functions such as ecological regulation, landscape beautification, and environmental purification. In recent years, due to water scarcity pressures, urban rivers, lakes, and landscape water bodies have increasingly relied on reclaimed water for replenishment. However, due to the complex sources and varying treatment processes of reclaimed water, sensory quality issues in urban water bodies after replenishment have become increasingly prominent, such as abnormal color, unpleasant odor, and decreased transparency. These issues not only affect the urban landscape and residents' experience but may also impact urban ecological security and public perception. Therefore, how to scientifically, objectively, and in real-time evaluate the sensory quality of urban water bodies after reclaimed water replenishment has become a critical issue that urgently needs to be addressed in the field of urban water environment management. Currently, monitoring and evaluation methods for urban water quality are mostly based on physicochemical indicators. Sensory attributes such as color, odor, and comfort often rely solely on manual inspections and subjective assessments, lacking a systematic, information-based, and automated sensory evaluation system. This makes it difficult to achieve timely early warning and source analysis of abnormal changes and is also detrimental to the scientific management of sensory risks and public participation. Especially when faced with new risk sources and complex changes brought about by reclaimed water replenishment, traditional water quality monitoring methods exhibit numerous shortcomings, such as poor timeliness, incomplete sensory data collection, and a disconnect between subjective evaluation and objective analysis. Therefore, there is an urgent need to develop a new method for evaluating the sensory quality of urban water bodies replenished by reclaimed water. This method integrates multi-dimensional sensory and physicochemical parameter intelligent synchronous collection, subjective and objective label correlation modeling, rapid anomaly identification, and source tracing optimization to achieve transparent, scientific, and intelligent management of urban water body sensory quality. Summary of the Invention

[0003] This invention aims to address the technical problems of traditional evaluation methods in the process of replenishing urban water bodies, such as incomplete monitoring of the sensory quality of water bodies after reclaimed water replenishment, strong subjectivity, poor real-time performance, difficulty in timely identification and tracing of abnormal changes, and lack of intelligent decision support. It aims to achieve multi-dimensional, synchronous, and objective monitoring and evaluation of sensory attributes such as color, odor, and transparency of urban water bodies, as well as basic physicochemical parameters, thereby improving the intelligent early warning, tracing, and scientific management capabilities for sensory anomalies.

[0004] To achieve the above objectives, the present invention employs the following technical solution: The method includes: A multi-sensory-physical parameter synchronous acquisition platform is constructed to collect high-resolution, real-time, and synchronous data on water color, gloss, surface suspended matter, odor, taste, temperature, pH, dissolved oxygen, and conductivity. Multiple rounds of double-blind subjective sensory evaluations were conducted on urban water bodies at sampling points, including subjective scoring and free description of dimensions such as color, transparency, odor, and overall comfort. Subjective sensory scoring data and descriptive information were collected and summarized. Anomaly detection algorithms are used to analyze subjective sensory data, and sliding time windows are used to detect local anomaly distribution, abrupt changes in sensory scores, and abnormal clustering phenomena. By combining time series analysis methods, we compare and analyze the multidimensional sensor data, subjective and objective scores, and historical water body evolution sequences corresponding to the abnormal monitoring results; Based on multi-parameter retrospective analysis, the water replenishment records, water source characteristics, environmental factors and trace changes at the abnormal mutation points are integrated and analyzed.

[0005] In one embodiment, the multispectral imaging optical camera accurately captures changes in the color, gloss, and surface suspended matter of the water body; the electronic nose uses multiple specific gas sensing units to detect odor information released by the water body; and the electronic tongue uses multiple types of ion sensing units to detect taste molecules in the water sample.

[0006] In one scheme, the physical parameter module collects basic water quality parameters, including temperature, pH, dissolved oxygen and conductivity, in real time. All the collected multidimensional data are integrated by the main control circuit in time synchronization, and after eliminating information timing deviations, they are uploaded to the urban water body management cloud platform via wireless network.

[0007] In one approach, the subjective sensory evaluation is completed by a group of citizen representatives of different ages, genders and occupations. The evaluation is conducted in multiple rounds of double-blind scoring, and the intuitive feelings are described in free text. The resulting subjective data is used for subjective-objective correlation modeling.

[0008] In one approach, the local outlier factor algorithm is used to detect anomalies in multidimensional sensory rating data. Based on a sliding time window mechanism, sampling points throughout the city are continuously monitored. When abrupt changes or abnormal distributions in sensory ratings are detected, they are promptly captured and marked as potential anomalies.

[0009] In one approach, time-series modeling techniques are applied to capture trends in sensory scores, sensor data, and historical water replenishment sequences before and after an anomaly occurs. By analyzing historical evolution trajectories, the specific time period and pattern of the anomaly outbreak are determined, enabling the time-series localization of the anomaly.

[0010] One approach employs multi-parameter backtracking analysis to comprehensively integrate water replenishment records at mutation points, characteristics of water replenishment sources, environmental factors, and trends in trace changes. Based on Bayesian networks or causal relationship modeling, the approach infers the anomaly induction mechanism, enabling intelligent attribution of responsibility for the anomaly source and intelligent recommendation of optimized water replenishment decision-making schemes.

[0011] In one approach, the method supports platform-based and modular deployment of multi-dimensional intelligent acquisition arrays, which can be flexibly applied to water replenishment scenarios such as urban rivers, lakes, and artificial water bodies, enabling full-dimensional, real-time synchronous monitoring and recording of the sensory and physicochemical states of water bodies.

[0012] Beneficial effects of this invention: This invention integrates multispectral imaging, electronic nose, electronic tongue, and basic water quality sensing technologies to achieve high-resolution, simultaneous acquisition of sensory attributes and physicochemical parameters of urban water bodies, significantly improving the objectivity and comprehensiveness of sensory quality evaluation after reclaimed water replenishment. The method effectively combines subjective citizen sensory evaluation with multidimensional sensor data analysis, enabling timely detection and location of sensory anomalies, intelligent source tracing, and optimized replenishment recommendations. Compared to traditional monitoring methods relying on manual inspections and single physicochemical indicators, this invention offers advantages such as high automation, wide monitoring range, fast response speed, accurate anomaly identification, and scientific management decision-making. It greatly enhances the continuous monitoring and risk prevention capabilities of urban water environment sensory quality, providing strong technical support for smart city water management and the safe utilization of reclaimed water. Attached Figure Description

[0013] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0014] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Typical embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0015] Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. To facilitate understanding, the invention will now be described more fully with reference to the accompanying drawings. Typical embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to make the disclosure of the invention more thorough and complete.

[0016] like Figure 1 As shown, a method for evaluating the sensory quality of urban water bodies replenished by reclaimed water is implemented as follows: Step 1. Construct a multi-dimensional sensory-physical parameter synchronous acquisition platform In constructing a multi-dimensional sensory-physical parameter synchronous acquisition platform, the first step is to comprehensively analyze the core evaluation indicators of urban water quality, including sensory information such as color, odor, and transparency, as well as key physicochemical parameters such as temperature, pH, and dissolved oxygen. Based on this, an innovative multi-dimensional intelligent acquisition array was designed. This array integrates a high-sensitivity optical camera, an electronic nose (E-nose), an electronic tongue (E-tongue), and traditional water quality sensors onto the same hardware platform. The optical camera employs multispectral imaging technology, enabling precise capture of minute changes in water color, gloss, and surface suspended matter. The electronic nose and electronic tongue utilize various specific gas and ion sensing units, respectively, and through dynamic airflow and an automatic water sample collection device, perform high-resolution detection of the odor emitted by the water and trace taste molecules in the solution, achieving digital characterization of odor and taste.

[0017] Meanwhile, the physical parameter module collects basic water quality information such as temperature, pH, dissolved oxygen, and conductivity in real time. All data is synchronized and integrated by a dedicated acquisition and control circuit to avoid information timing deviations. The entire platform adopts a modular design and can be flexibly deployed at various urban water body banks or floating monitoring stations. The collected multidimensional data is uploaded to the urban water body management cloud platform in real time via wireless network, realizing full-dimensional, real-time, and synchronous monitoring and recording of the water's sensory and physicochemical states, laying a solid data foundation for subsequent intelligent sensory quality assessment.

[0018] Step 2. Conduct intelligent subjective-objective sensory label association modeling In the process of implementing intelligent subjective-objective sensory label association modeling, a group of citizen representatives of different ages, genders, and occupations was first organized to conduct multiple rounds of double-blind sensory evaluations of urban water bodies at various sampling points. Participants, unaware of the specific source of the water, subjectively scored dimensions such as color, transparency, odor, and overall comfort, while also expressing their intuitive feelings about the water in a free-form description. After data collection, advanced natural language processing (NLP) algorithms were used to process the citizens' descriptions. First, word segmentation and word embedding (such as Word2Vec or BERT models) were used to convert the text into vector representations; then, clustering algorithms such as K-means or hierarchical clustering were used to classify the topic keywords, and the main sensory labels were extracted through similarity calculation.

[0019] The quantification of sensory labels can be achieved using the TF-IDF (Term Frequency-Inverse Document Frequency) method, with the following formula: in, This represents the weight of the i-th label in the j-th evaluation text. Let N be the frequency of label i in text j, and N be the total number of evaluation texts. Let i be the number of texts containing the label i.

[0020] Next, to achieve deep coupling between subjective sensory labels and objective multidimensional sensory-physical parameters, a convolutional neural network (CNN) model is used to establish a mapping relationship. Let the multidimensional input data matrix be... ,in Let n be the set of real numbers, n be the number of sampling points, and m be the multidimensional sensor parameters for each sampling point; the sensory label vector is... , where k is the total number of extracted sensory labels. The CNN network uses multiple layers of convolution and pooling to... Feature extraction is performed, and the final output is... The corresponding predicted label probability. The training objective of the model can be measured by the mean squared error loss function (MSE), as follows: in, Let i be the predicted label value for the i-th sample. These are the true label values. By continuously optimizing the CNN network weights through backpropagation and gradient descent algorithms, efficient and automated mapping of multi-dimensional sensor data to subjective sensory labels is achieved. This establishes an intelligent association model between subjective and objective sensory data, improving the scientific rigor and interpretability of urban water body sensory quality assessment.

[0021] Step 3. Establish a dynamic weight adjustment mechanism suitable for the characteristics of reclaimed water. Based on the multidimensional sensory-physical parameter acquisition and subjective-objective mapping model (steps 1 and 2), a dynamic weight adjustment mechanism suitable for the characteristics of reclaimed water is established. First, key trace substances that easily induce sensory changes during reclaimed water replenishment, such as specific trace organic matter and disinfection byproducts, need to be extracted. The concentrations of these trace substances are monitored in real time by a multidimensional sensor array, and components that may cause significant sensory changes are identified using thresholding or principal component analysis (PCA). To this end, a weight adjustment factor is constructed. Where j represents different sensory dimensions (such as color, taste, turbidity, etc.), and the weighting factor of each dimension is based on the concentration of its related trace substances. and sensory sensitivity coefficient Dynamic calculation can be expressed by the following formula: Where p is the total number of traces. Let be the sensitivity coefficient of the i-th trace to the j-th sensory index. Let i be the current measured concentration of the i-th trace substance. and These represent the mean and standard deviation of the trace element in historical data, respectively. When the concentration of a trace element deviates significantly from the normal range, the weight of the corresponding sensory indicator will automatically increase or decrease, achieving dynamic and sensitive adjustment of indicators based on the characteristics of reclaimed water replenishment.

[0022] Furthermore, by incorporating environmental variables such as season (S), climate (W), and water replenishment frequency (F) of the urban water body, a decision tree model is used to provide criteria for weight adjustment. The decision tree uses these environmental factors as node features, learns from historical sensory data, and automatically summarizes rules for weight adjustment under different conditions. The decision path T(S, W, F) outputs preliminary weight correction suggestions for each sensory indicator. Simultaneously, Bayesian inference is used to probabilistically optimize the weight results. Let... This is the final weight vector. Let E be the posterior probability of each weight combination under given environmental condition E, then we have: in, A set of environmental factors In weighted combination The likelihood probability of observing environmental condition E. The prior distribution of the weights can be derived from historical sensory quality events. By weighted fusion of the weight suggestions output by the decision tree and the Bayesian posterior probability distribution, a real-time, dynamic, and adaptively adjustable sensory evaluation weight system is formed, thereby significantly improving the sensitivity and accuracy of sensory quality evaluation of urban water bodies replenished by reclaimed water under changing environments.

[0023] Step 4. Conduct grading, quantification, and visual feedback of water body sensory quality. In the stage of quantification and visualization feedback for water body sensory quality grading, the comprehensive sensory score obtained from the aforementioned multidimensional sensory data and dynamic weight adjustment model is first used to classify the water quality at each sampling point in the city using an adaptive clustering algorithm. To achieve automatic grading, K-means++ starts with an intelligent selection method for initial centers, processing the comprehensive sensory feature vectors of all sampling points... (in The weighted combination of multidimensional sensory and physical attributes of the i-th sampling point is used as the clustering input. The algorithm aims to divide the n sampling points into K level categories (e.g., excellent, good, average, poor) and minimize the within-class variance SSE (Sum of Squared Errors). in, Let k represent the set of all sampled points in the k-th class. Let K be the centroid of the k-th class. The value of K can be automatically optimized and determined by the silhouette coefficient, etc., to ensure the scientific nature and discriminative power of the classification results.

[0024] After clustering, the system sets sensory quality level standards based on the centroid features of each cluster and assigns all sampling points to the corresponding levels. Based on the grading results, a sensory quality map and heat map of urban water bodies are further developed. First, the geographical location information of each sampling point is spatially mapped to its grading results, using colors to distinguish different sensory levels, for example, using gradient colors such as blue, green, yellow, and red to represent levels from excellent to poor. The heat map uses weighted kernel density estimation (KDE) to smooth the distribution of comprehensive sensory scores in each region, and calculates the overall sensory quality distribution hotspots of urban water bodies. The formula is: in, For location The density estimate is given at each location, where n is the number of sampling points, d is the spatial dimension, h is the smoothing bandwidth, and K is the kernel function (such as a Gaussian kernel). Finally, the system publishes the sensory grading map and heat map in real time via a web-based map service or mobile client. This facilitates management departments in monitoring the health status of water bodies and quickly locating abnormal water areas. Simultaneously, it provides the public with an intuitive and interactive reference to the sensory status of water bodies, achieving efficient information feedback and collaborative governance.

[0025] Step 5. Intelligent Early Warning and Tracing of Sensory Quality Anomalies In implementing an intelligent early warning and tracing mechanism for abnormal sensory quality in urban water bodies, a real-time monitoring system for multi-dimensional sensory data streams is first constructed, using the sensory scores and physicochemical indicators of each sampling point over time as input. To efficiently identify sensory anomalies, intelligent detection algorithms based on Isolation Forest and Local Outlier Factor (LOF) are employed to achieve real-time detection of anomalies or mutations in large-scale data. Taking LOF as an example, for each sampling point i at any given time, its sensory feature vector is... The LOF algorithm first calculates the reachable distance of its k nearest neighbors. Then, the locally reachable density (LRD) is calculated: in, for The set of k nearest neighbors. The LOF score is then: when When the value is significantly greater than 1, the sampling point is determined. This identifies potential sensory anomalies. By continuously monitoring citywide data through a sliding time window, the system can promptly detect abrupt changes, abnormal distributions, or local clustering of sensory scores.

[0026] Following the discovery of anomalies, source tracing was conducted by combining time-series analysis and historical evolution trajectories. This involved integrating multi-dimensional sensor data, subjective and objective scores, and historical water body evolution sequences corresponding to the anomaly points. By comparison, time-series modeling methods such as ARIMA (Autoregressive Moving Average) are used to capture the changing trends before and after the occurrence of anomalies. The basic expression of the ARIMA model is: in, Let be the sensory rating or key indicator at time t, c be a constant term, and p and q be the orders of the autoregression and moving average, respectively. , These are the model coefficients. The noise is white noise. By comparing ARIMA predictions with actual data, the system can pinpoint the specific time period and evolution pattern of abnormal outbreaks.

[0027] To achieve attribution of responsibility for anomalies and optimized decision-making, the system integrates water replenishment records, source characteristics, environmental factors, and trace changes at the point of abrupt change based on multi-parameter backtracking analysis. It then uses Bayesian networks or causal relationship modeling to infer the most likely anomaly triggering mechanism. Simultaneously, based on the source tracing results and historical responses of urban water bodies, it recommends targeted optimized water replenishment schemes. For example, if an abnormal increase in a certain type of trace organic matter is found to be highly correlated with a specific water replenishment scheme, it suggests dynamically adjusting the water replenishment frequency, optimizing disinfection processes, or adding temporary purification steps. In this way, the system not only achieves intelligent and automated detection and response to sensory anomalies in water bodies but also provides scientific decision support tools for urban water environment management.

[0028] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0029] It should be understood that the above detailed description of the technical solutions of the present invention with reference to preferred embodiments is illustrative and not restrictive. Those skilled in the art can modify the technical solutions described in the embodiments or make equivalent substitutions for some of the technical features based on reading this specification; however, these modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for evaluating the sensory quality of urban water bodies replenished by reclaimed water, characterized in that: The method includes: A multi-sensory-physical parameter synchronous acquisition platform was constructed to collect high-resolution, real-time, and synchronous data on water color, gloss, surface suspended matter, odor, taste, temperature, pH, dissolved oxygen, and conductivity. The collected data are physicochemical index data. Multiple rounds of double-blind subjective sensory evaluations were conducted on urban water bodies at sampling points. This included subjective scoring and free description of dimensions such as color, transparency, odor, and overall comfort. Subjective sensory scoring data and descriptive information were collected and summarized. The results of the double-blind subjective sensory evaluations were the subjective sensory scoring data. The physicochemical index data and the subjective sensory rating data are combined to form multidimensional sensory rating data. The local outlier factor algorithm is used to analyze multidimensional sensory rating data, and local abnormal distribution, sensory rating mutations and abnormal clustering are detected by sliding time window. By combining time series analysis methods, we compare and analyze the multidimensional sensor data, subjective and objective scores, and historical water body evolution sequences corresponding to the abnormal monitoring results; Based on multi-parameter retrospective analysis, the water replenishment records, water source characteristics, environmental factors and trace changes at the abnormal mutation points are integrated and analyzed.

2. The method for evaluating the sensory quality of urban water bodies replenished by reclaimed water according to claim 1, characterized in that: The multi-dimensional sensory-physical parameter synchronous acquisition platform captures changes in the color, gloss, and surface suspended matter of the water body. The electronic nose uses multiple specific gas sensing units to detect odor information released by the water body, and the electronic tongue uses multiple ion sensing units to detect taste molecules in the water sample.

3. The method for evaluating the sensory quality of urban water bodies replenished by reclaimed water according to claim 1, characterized in that: The multi-dimensional sensory-physical parameter synchronous acquisition platform collects basic water quality parameters, including temperature, pH, dissolved oxygen, and conductivity, in real time. All the collected multi-dimensional data are integrated by the main control circuit in time synchronization, and after eliminating information timing deviations, they are uploaded to the urban water body management cloud platform via wireless network.

4. The method for evaluating the sensory quality of urban water bodies replenished by reclaimed water according to claim 1, characterized in that: The subjective sensory evaluations were completed by representatives of different ages, genders, and occupations, and were scored in multiple rounds using a double-blind method. The subjective information was described in free text, and the resulting subjective data was used for subjective-objective correlation modeling.

5. The method for evaluating the sensory quality of urban water bodies replenished by reclaimed water according to claim 1, characterized in that: The local outlier factor algorithm is used to detect anomalies in multidimensional sensory rating data. Based on the sliding time window mechanism, sampling points throughout the city are continuously monitored. When abrupt changes or abnormal distributions in sensory ratings are detected, they are captured and marked as potential anomalies in a timely manner.

6. The method for evaluating the sensory quality of urban water bodies replenished by reclaimed water according to claim 1, characterized in that: By applying time-series modeling techniques to capture trends in sensory scores, sensor data, and historical water replenishment sequences before and after an anomaly occurs, and analyzing the specific time period and pattern of the anomaly outbreak through historical evolution trajectory analysis, the anomaly can be located in time sequence.

7. The method for evaluating the sensory quality of urban water bodies replenished by reclaimed water according to claim 1, characterized in that: Multi-parameter retrospective analysis was used to comprehensively integrate the water replenishment records of the mutation point, the characteristics of the water source, the environmental factors and the trend of trace changes, and the abnormal induction mechanism was inferred based on Bayesian network or causal relationship modeling.

8. The method according to any one of claims 1 to 7, characterized in that: The method described supports platform-based and modular deployment of multi-dimensional intelligent acquisition arrays, enabling full-dimensional, real-time synchronous monitoring and recording of the sensory and physicochemical states of water bodies.

Images