A method, system, medium, and apparatus for water concentration perception based on scattering features

Abstract

The application discloses a water body concentration perception method and system based on scattering characteristics, a medium and equipment, and belongs to the technical field of optical sensing and water quality monitoring. The method comprises the following steps: acquiring the angle-resolved scattering intensity curve of a water body sample to be measured, and discretizing the angle-resolved scattering intensity curve into a scattering intensity sequence; extracting a multi-dimensional scattering feature vector comprising overall energy and angle-weighted statistical features, amplitude statistical morphological features, angle region energy distribution and specific gravity features, local angle distribution change rate features and global frequency domain morphological features; performing water body concentration grade analysis and judgment and outputting; under the condition that only qualitative classification or imperfect scattering data labeling is available, the multi-dimensional scattering feature vector is extracted to deeply mine the implicit structural information of the scattering signal, and the water body concentration grade perception with high discrimination and high reliability is realized.

Description

Technical Field

[0001] This application relates to the field of optical sensing and water quality monitoring technology, and in particular to a method, system, medium and device for sensing water concentration based on scattering characteristics. Background Technology

[0002] Currently, existing optical concentration measurement methods are mainly divided into two categories: one is based on the Lambert-Beer law, which estimates concentration by detecting the absorption characteristics of a substance at a specific wavelength of light. The measurement range of such devices usually varies depending on the type of medium, with a common measurement range of 0–25 mg / mL. The other category is based on light scattering mechanisms. For example, existing portable optical turbidimeters are used to test the concentration of chlorophyll a in water, with a measurement range of 1–30 μg / mL. Some methods also combine absorption and scattering information to improve measurement accuracy.

[0003] However, existing seawater scattering databases are mostly labeled based on water body type or environmental analogy. For example, Petzold's data only provides qualitative classifications such as clear, coast, and turbid, without providing concentration values ​​that can be directly used for quantitative calculations. At the same time, existing experimental data only cover a limited number of discrete concentration conditions, which is insufficient to support high-precision regression modeling based on continuous labels.

[0004] Therefore, how to extract multidimensional features and mine deep structures from scattering signals under the condition of only qualitative classification or incomplete labeling, and thereby realize a water concentration level sensing method with high discriminative power and high reliability, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] Therefore, it is necessary to address the aforementioned technical problems by providing a water concentration sensing method, system, medium, and device based on scattering characteristics.

[0006] The following technical solution is adopted in this specification: This specification provides a water concentration sensing method based on scattering characteristics, specifically including: Obtain the angle-resolved scattering intensity curve of the water sample to be tested, and discretize it into a scattering intensity sequence.

[0007] Based on the scattering intensity sequence, a multidimensional scattering feature vector is extracted.

[0008] The multidimensional scattering feature vector includes overall energy and angle-weighted statistical features reflecting the overall scattering energy distribution intensity, amplitude statistical morphological features describing the sharpness or flatness of the angular distribution, angular region energy distribution and proportion features characterizing the forward and backward energy distribution relationship, local angular distribution change rate features quantifying the rate of change of scattering intensity within a specific angular range, and global frequency domain morphological features.

[0009] The water concentration level is analyzed and determined based on the multidimensional scattering feature vector.

[0010] Based on the analysis and judgment results, the concentration level perception result is output.

[0011] Furthermore, the overall energy and angle-weighted statistical characteristics include the weighted integral area, scattering angle, and scattering angle broadening.

[0012] The weighted integral area is obtained by passing through 0 to... Within the scattering angle range, the scattering intensity sequence is obtained by performing a sinusoidal weighted integral operation, which is used to reflect the overall scattering energy distribution intensity.

[0013] The scattering angle is obtained by dividing the sinusoidal weighted integral of the product of the scattering intensity sequence and the corresponding scattering angle by the sinusoidal weighted integral of the scattering intensity sequence, and is used to describe the overall bias direction of the scattered energy.

[0014] The broadening of the scattering angle is obtained by quotienting the sinusoidal weighted integral of the product of the scattering intensity sequence and the square of the difference between the scattering angle and the mean scattering angle, and is used to describe the extent of the spread of the scattering angle distribution.

[0015] Furthermore, the amplitude statistical morphological characteristics include the average amplitude reflecting the overall scattering intensity level, the amplitude dispersion describing the sharpness or flatness of the angular distribution, the amplitude skewness characterizing the asymmetry of forward and backward scattering, and the amplitude kurtosis distinguishing between strong forward spikes and wide-angle scattering.

[0016] Furthermore, the corner region energy distribution and proportion characteristics include the scattered energy within a specific angle range, the ratio of forward to mid-angle energy, the ratio of forward to backward energy, and the backward scattering concentration.

[0017] The scattering energy within the specific angle range is obtained by performing a sinusoidal weighted integral operation on the scattering intensity sequence within a preset range from the starting scattering angle to the ending scattering angle.

[0018] The local angular distribution change rate characteristic is the rate of change of scattering intensity within a specific angular range, which is obtained by subtracting the scattering intensity value at the end of the preset angular range from the scattering intensity value at the beginning, and then quotienting it with the angular span of the preset angular range.

[0019] Furthermore, the global frequency domain morphological features are global frequency domain morphological coefficients obtained by performing a discrete cosine transform on the discrete intensity curve sequence. ; ; in, These are the indexes of the discrete cosine transform coefficients. The number of discrete sampling points in the angular dimension. This is the index of the discrete sampling points. In the first The scattering intensity value at each sampling point Pi is the mathematical constant of a circle.

[0020] Furthermore, the concentration levels are multiple discrete categories divided into low to high concentrations, including clear, coast, turbid, and very turbid.

[0021] Furthermore, the water concentration level analysis and judgment based on the multidimensional scattering feature vector specifically involves: using the overall energy and angle-weighted statistical features as the global scattering intensity distribution benchmark, extracting the backscattering concentration, local angle change rate features, and global frequency domain morphological coefficients from the angular region energy distribution and weight features, and supplementing the structural information of the scattering intensity distribution through the amplitude statistical morphological features.

[0022] When the backscattering concentration is higher than the first threshold and shows a negative change with the local angle change rate feature, and the distribution of the global frequency domain morphological coefficients conforms to the preset high concentration morphological pattern, the water concentration level is determined to be turbid; otherwise, it is determined to be clear.

[0023] This specification provides a water concentration sensing system based on scattering characteristics, specifically including: The data acquisition module is used to acquire the angle-resolved scattering intensity curve of the water sample to be tested and discretize it into a scattering intensity sequence.

[0024] The multidimensional scattering feature vector extraction module extracts multidimensional scattering feature vectors based on the scattering intensity sequence.

[0025] The judgment output module is used to analyze and judge the water concentration level based on the multidimensional scattering feature vector, and output the concentration level perception result based on the analysis and judgment results.

[0026] This specification provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described above.

[0027] This specification also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described above.

[0028] The beneficial effects of this invention are as follows: Compared with the prior art, this invention first constructs a multi-dimensional feature vector, and then analyzes and judges the structural changes, backscattering enhancement and angular distribution morphology differences of each dimension feature, and finally obtains the water concentration perception result. It effectively achieves high discriminative power and high reliability of water concentration level perception by extracting multi-dimensional features and mining deep structures of scattering signals under the condition of only qualitative classification or incomplete labeling of scattering data. Attached Figure Description

[0029] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0030] Figure 1 This is a schematic diagram of the process of the present invention; Figure 2 This is a comparison chart of the concentration category recognition accuracy of different machine learning models in this embodiment of the invention; Figure 3 This is a confusion matrix diagram according to an embodiment of the present invention; Figure 4 This is a feature importance distribution diagram in an embodiment of the present invention; Figure 5 This is a radar chart illustrating the importance of key features in the embodiments of the present invention. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this specification clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments in this specification without creative effort are within the scope of protection of this application.

[0032] This invention provides a water concentration sensing method based on scattering features, comprising four steps: data acquisition and discretization, construction of a multi-dimensional scattering feature vector, analysis and judgment of water concentration levels, and output of concentration level sensing results based on the analysis and judgment results. This method constructs a multi-dimensional feature vector and analyzes and judges the structural changes in scattering angle distribution, backscattering enhancement, and differences in angle distribution morphology for each dimension of the feature vector, ultimately obtaining the water concentration sensing result. This effectively and comprehensively mines the structural information of scattering signals even with only qualitatively graded or poorly labeled scattering data, thereby achieving highly discriminative and reliable water concentration level sensing.

[0033] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0034] A water concentration sensing method based on scattering characteristics, such as Figure 1 As shown, it specifically includes: Step 1: Obtain the angle-resolved scattering intensity curve of the water sample to be tested, and discretize it into a scattering intensity sequence.

[0035] Step 2: Extract multidimensional scattering feature vectors based on the scattering intensity sequence.

[0036] The multidimensional scattering feature vector includes the overall energy and angle-weighted statistical features that reflect the overall scattering energy distribution intensity, the amplitude statistical morphological features that describe the sharpness or flatness of the angular distribution, the angular region energy distribution and proportion features that characterize the forward and backward energy distribution relationship, the local angular distribution change rate features that quantify the rate of change of scattering intensity within a specific angular range, and the global frequency domain morphological features.

[0037] Optionally, the overall energy and angle-weighted statistical characteristics include the weighted integral area reflecting the overall intensity of the scattered energy distribution. The scattering angle describes the overall direction of the scattered energy. And scattering angular broadening, which describes the extent of the spread of the scattering angular distribution. ; ; ; ; in, It is a scattering intensity sequence. The scattering angle is... This is the differential symbol.

[0038] Optionally, the amplitude statistical morphological characteristics include the average amplitude reflecting the overall scattering intensity level, the amplitude dispersion describing the sharpness or flatness of the angular distribution, the amplitude skewness characterizing the asymmetry of forward and backward scattering, and the amplitude kurtosis distinguishing between strong forward spikes and wide-angle scattering.

[0039] Optionally, the corner region energy distribution and proportion characteristics include the scattered energy within a specific angular range. The ratio of forward to mid-angle energy, the ratio of forward to backward energy, and the concentration of backward scattering; ; in, This represents the scattering angle range.

[0040] The local angular distribution rate of change is the rate of change of scattering intensity within a specific angular range. ; .

[0041] Optionally, the global frequency domain morphological features are global frequency domain morphological coefficients obtained by performing a discrete cosine transform on the discrete intensity curve sequence. ; ; in, These are the indexes of the discrete cosine transform coefficients. The number of discrete sampling points that are not yet read is called again. This is the index of the discrete sampling points.

[0042] Step 3: Analyze and determine the water concentration level based on the multidimensional scattering feature vector.

[0043] Optionally, the concentration levels are multiple discrete categories divided into low to high concentrations, including clear, coast, turbid, and very turbid.

[0044] Optionally, the water concentration level analysis and judgment based on the multidimensional scattering feature vector specifically involves: using the overall energy and angle-weighted statistical features as the global scattering intensity distribution benchmark, extracting the backscattering concentration, local angle change rate features, and global frequency domain morphological coefficients from the angular region energy distribution and weight features, and supplementing the structural information of the scattering intensity distribution through the amplitude statistical morphological features.

[0045] When the backscattering concentration is higher than the first threshold and shows a negative change with the local angle change rate feature, and the distribution of the global frequency domain morphology coefficients conforms to the preset high concentration morphology pattern, the water concentration level is determined to be turbid; otherwise, it is determined to be optional, clear.

[0046] Step 4: Based on the analysis and judgment results, output the concentration level perception results.

[0047] Example This embodiment provides a water concentration sensing method based on scattering characteristics, including the following steps: S1: Data Preparation Based on the different concentration conditions set in the experiment, the scattering data were divided into four categories—clear, coast, turbid, and very turbid—in ascending order of concentration. Subsequently, within each category, training and validation data were allocated according to a predetermined ratio, with approximately 80% of the samples used for model training and the remainder for model validation. The training and validation sets constructed in this way maintained consistent category distribution, providing fundamental data support for subsequent classification modeling.

[0048] S2: Scattering Feature Construction The scattering features are divided into five categories, corresponding to the overall statistical characteristics of the scattering distribution, the amplitude distribution pattern, the corner energy distribution, the local angle variation, and the global morphological structure, thereby constructing a hierarchical and structured set of scattering features.

[0049] The first category is the overall energy and angle-weighted statistical characteristics. The weighted integral area of ​​the scattering curve This reflects the overall intensity of scattered energy distribution.

[0050] ; in, This represents the scattering curve formed by connecting the scattering intensity values ​​at various angles extracted in the experiment. The scattering angle is... This is the differential symbol.

[0051] Scattering angular centroid This describes whether the scattered energy is biased towards the forward or backward direction.

[0052] Expansion of scattering angle distribution .

[0053] The second category is amplitude statistical morphological characteristics. average amplitude of scattering curve This reflects the overall scattering intensity level.

[0054] Scattering curve amplitude dispersion This describes whether the angular distribution is sharp or flat.

[0055] Skewness of amplitude distribution This characterizes the asymmetry between forward and backward scattering.

[0056] Amplitude distribution kurtosis Distinguish between strong forward spikes and wide-angle scattering.

[0057] The third category is the energy distribution and specific gravity characteristics of the corner region. Angle range Internal scattering energy .

[0058] Forward and mid-angle energy ratio .

[0059] Forward to backward energy ratio .

[0060] Backscattering concentration .

[0061] The fourth category is the characteristic of the rate of change of local angular distribution. Angle range Rate of change of scattering intensity within .

[0062] The fifth category is global frequency domain morphological characteristics. The coefficients obtained after performing a discrete cosine transform on a discrete signal .

[0063] ; in, These are the indexes of the discrete cosine transform coefficients. The number of discrete sampling points that are not yet read is called again. This is the index of the discrete sampling points.

[0064] The above five types of features are combined to construct a feature matrix, where the first... The scattering feature vector corresponding to each sample is represented as: .

[0065] S3: Examine the effectiveness of scattering characteristics in distinguishing different water body categories. S3.1: A multilayer perceptron model is employed, which transforms the input features layer by layer through multilayer nonlinear mapping to approximate the complex nonlinear relationship between scattering features and concentration categories. For the input scattering feature vector... In its first The implicit representation of a layer can be written as:

[0066] ; in and These represent the weight matrix and bias vector of this layer, respectively. It is a non-linear activation function, and After multiple layers of mapping, the output layer estimates the probability of a sample belonging to each concentration category and uses the category with the highest probability as the prediction result. In the application scenario of this paper, the multilayer perceptron model can be regarded as a tool for performing a continuous mapping of the scattering feature space. Its role is to comprehensively utilize the coupling relationship between different types of scattering features to form a discrimination boundary for concentration categories. Compared with rule-based tree models, MLP does not rely on explicit feature thresholding but constructs a smooth decision boundary in the feature space through weight learning, thus providing another modeling perspective for the overall discriminative ability of scattering features in concentration category identification.

[0067] S3.2: A random forest model is used, which partitions the input feature space by constructing multiple decision trees. Each decision tree can be viewed as a set of recursive partitioning rules implemented on the feature space. For a given scattering feature vector... , No. Each decision tree outputs a category prediction result. The overall output of the random forest model is obtained by integrating the results of each decision, and can be expressed as:

[0068] ; in This indicates the number of decision trees. In the scattering feature modeling problem, the partitioning process of a random forest is equivalent to finding a series of threshold conditions in a multidimensional scattering feature space to distinguish the scattering distribution characteristics under different concentration categories. For example, corner energy features or local angle change rate features can form an effective partitioning basis at the decision node, thereby achieving the discrimination of different water body categories. Since random forests randomly sample both samples and features during training, the model has a weak dependence on feature scale and distribution form, making it suitable for processing input vectors composed of multiple heterogeneous scattering features.

[0069] S3.3: Input the scattering feature set into the multilayer perceptron model and the random forest model respectively, and train and evaluate them on the same training and validation sets to ensure the comparability of experimental results.

[0070] refer to Figure 2 Both the Multilayer Perceptron (MLP) and Random Forest (RF) models were able to effectively distinguish water states with different concentration levels. The overall accuracy of the MLP model was 0.9725, while that of the RF model was 0.9561. The fact that both models, with different modeling mechanisms, achieved high and similar recognition accuracies indicates that the constructed scattering features have good discriminative ability at the overall level.

[0071] refer to Figure 3 This is a confusion matrix corresponding to the two models, used to analyze the specific identification situation between different concentration categories. Figure 3 (a) and Figure 3 (b) As can be seen, both the "clear" and "very turbid" samples exhibit high recognition accuracy in both models, with the vast majority of samples being correctly classified and no misclassifications. This indicates that under low and high concentration conditions, the scattering angle distribution shows significant differences in overall intensity level and distribution pattern, and the corresponding scattering features can provide clear discrimination criteria for the model. In contrast, there is a certain degree of confusion between medium concentration categories such as "coast" and "turbid," with some samples being misclassified during the classification process. However, overall, the recognition accuracy remains high. The classification results obtained under the two different modeling frameworks verify the feasibility of water concentration category identification based on scattering features from both the overall accuracy and category distribution perspectives. Furthermore, the inter-category differences presented in the confusion matrix provide an important basis for subsequent analysis of the discrimination sensitivity of different concentration ranges from the feature level.

[0072] S3.4: Feature Importance Analysis refer to Figure 4 Although the two models have different modeling mechanisms, the key features they emphasize show strong consistency overall. Features related to backscattering, such as the energy ratio in the backscattering corner region, hold high weight in both models. Local slope features related to the backscattering region are also given high importance by the model. Under different concentration conditions, the energy distribution and variation characteristics of the backscattering region significantly contribute to the differentiation of water concentration levels. Features related to the morphology of the scattering angle distribution also show high importance in both models. Among them, the low-frequency coefficients of the multiple discrete cosine transform (DCT) are simultaneously selected as key features, reflecting the important role of the overall morphology of the scattering angle distribution in concentration differentiation. In addition to the energy variation in the local corner region, the global structural information of the scattering distribution also includes concentration-related discriminative features.

[0073] refer to Figure 5 For the multilayer perceptron model, the distribution of important features is dominated by a few key features, especially the backscattering energy ratio and the mid-angle slope, which play a leading role in model discrimination. In contrast, the random forest model exhibits a relatively balanced distribution of importance among multiple features, emphasizing not only backscattering features but also giving high weight to the mid-angle slope and some global morphological features. This difference reflects the different focuses of different models when using scattering features for discrimination, but the types of features they commonly focus on are physically consistent. Multi-class features constructed based on angle-resolved scattering information have a clear emphasis in concentration category identification, with backscattering and scattering distribution morphology features showing higher sensitivity.

[0074] In the process of distinguishing concentration levels based on angle-resolved scattering information, features related to the scattering angle distribution structure play a dominant role in model discrimination. Feature importance analysis shows that the energy distribution characteristics and related proportional characteristics in the backscattering angle region have high weights in different models. At the same time, the local slope characteristics in the middle angle region and the backscattering angle region make a significant contribution to the concentration level distinction.

[0075] This specification provides a water concentration sensing system based on scattering characteristics, specifically including: The data acquisition module is used to acquire the angle-resolved scattering intensity curve of the water sample to be tested and discretize it into a scattering intensity sequence.

[0076] The multidimensional scattering feature vector extraction module extracts multidimensional scattering feature vectors based on the scattering intensity sequence.

[0077] The judgment output module is used to analyze and judge the water concentration level based on the multidimensional scattering feature vector, and output the concentration level perception result based on the analysis and judgment results.

[0078] For specific limitations regarding a water concentration sensing system based on scattering characteristics, please refer to the limitations of a water concentration sensing method based on scattering characteristics mentioned above, which will not be repeated here. Each module in the aforementioned water concentration sensing system based on scattering characteristics can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0079] This specification also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0080] This specification provides a computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described above.

[0081] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0082] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. A water concentration sensing method based on scattering characteristics, characterized in that, include: Obtain the angle-resolved scattering intensity curve of the water sample to be tested, and discretize it into a scattering intensity sequence; Based on the scattering intensity sequence, a multidimensional scattering feature vector is extracted; The multidimensional scattering feature vector includes the overall energy and angle-weighted statistical features reflecting the overall scattering energy distribution intensity, the amplitude statistical morphological features describing the sharpness or flatness of the angular distribution, the angular region energy distribution and proportion features characterizing the forward and backward energy distribution relationship, the local angular distribution change rate features quantifying the rate of change of scattering intensity within a specific angular range, and the global frequency domain morphological features. Water concentration level analysis and judgment are performed based on the multidimensional scattering feature vector; Based on the analysis and judgment results, the concentration level perception result is output.

2. The water concentration sensing method based on scattering characteristics as described in claim 1, characterized in that: The overall energy and angle-weighted statistical characteristics include the weighted integral area, scattering angle, and scattering angle broadening; The weighted integral area is obtained by passing through 0 to... Within the scattering angle range, the scattering intensity sequence is obtained by performing a sinusoidal weighted integral operation, which is used to reflect the overall scattering energy distribution intensity. The scattering angle is obtained by taking the quotient of the sinusoidal weighted integral of the product of the scattering intensity sequence and the corresponding scattering angle, and the sinusoidal weighted integral of the scattering intensity sequence, and is used to describe the overall deflection direction of the scattered energy. The broadening of the scattering angle is obtained by quotienting the sinusoidal weighted integral of the product of the scattering intensity sequence and the square of the difference between the scattering angle and the mean scattering angle, and is used to describe the extent of the spread of the scattering angle distribution.

3. The water concentration sensing method based on scattering characteristics as described in claim 1, characterized in that: The amplitude statistical morphological characteristics include the average amplitude reflecting the overall scattering intensity level, the amplitude dispersion describing the sharpness or flatness of the angular distribution, the amplitude skewness characterizing the asymmetry of forward and backward scattering, and the amplitude kurtosis distinguishing between strong forward spikes and wide-angle scattering.

4. The water concentration sensing method based on scattering characteristics as described in claim 1, characterized in that: The corner region energy distribution and proportion characteristics include the scattered energy within a specific angle range, the ratio of forward to mid-angle energy, the ratio of forward to backward energy, and the backward scattering concentration. The scattered energy within the specific angle range is obtained by performing a sinusoidal weighted integral operation on the scattering intensity sequence within a preset range from the starting scattering angle to the ending scattering angle. The local angular distribution change rate characteristic is the rate of change of scattering intensity within a specific angular range, which is obtained by subtracting the scattering intensity value at the end of the preset angular range from the scattering intensity value at the beginning, and then quotienting it with the angular span of the preset angular range.

5. The water concentration sensing method based on scattering characteristics as described in claim 1, characterized in that: The global frequency domain morphological features are global frequency domain morphological coefficients obtained by performing a discrete cosine transform on a discrete intensity curve sequence. ; ； in, These are the indexes of the discrete cosine transform coefficients. The number of discrete sampling points in the angular dimension. This is the index of the discrete sampling points. In the first The scattering intensity value at each sampling point Pi is the mathematical constant of a circle.

6. The water concentration sensing method based on scattering characteristics as described in claim 1, characterized in that: The concentration levels are multiple discrete categories divided into low to high concentrations, including clear, coast, turbid, and very turbid.

7. The water concentration sensing method based on scattering characteristics as described in claim 6, characterized in that, The specific steps for analyzing and determining the water concentration level based on the multidimensional scattering feature vector are as follows: Using the overall energy and angle-weighted statistical characteristics as the global scattering intensity distribution benchmark, the backscattering concentration, local angle change rate characteristics, and global frequency domain morphology coefficients are extracted from the energy distribution and weighting characteristics of the corner region, and the structural information of the scattering intensity distribution is supplemented by the amplitude statistical morphology characteristics. When the backscattering concentration is higher than the first threshold and shows a negative change with the local angle change rate feature, and the distribution of the global frequency domain morphological coefficients conforms to the preset high concentration morphological pattern, and the amplitude statistical morphological feature conforms to the steep peak characteristics of strong scattering, the water concentration level is determined to be turbid; otherwise, it is determined to be clear.

8. A water concentration sensing system based on scattering characteristics, characterized in that... include: The data acquisition module is used to acquire the angle-resolved scattering intensity curve of the water sample to be tested and discretize it into a scattering intensity sequence. A multidimensional scattering feature vector extraction module extracts multidimensional scattering feature vectors based on the scattering intensity sequence; The judgment output module is used to analyze and judge the water concentration level based on the multidimensional scattering feature vector, and output the concentration level perception result based on the analysis and judgment results.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the method described in any one of claims 1 to 7.

10. A computer device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in any one of claims 1 to 7.

Images