CN113295635A - Water pollution alarm method based on dynamic update data set - Google Patents
Water pollution alarm method based on dynamic update data set Download PDFInfo
- Publication number
- CN113295635A CN113295635A CN202110585040.7A CN202110585040A CN113295635A CN 113295635 A CN113295635 A CN 113295635A CN 202110585040 A CN202110585040 A CN 202110585040A CN 113295635 A CN113295635 A CN 113295635A
- Authority
- CN
- China
- Prior art keywords
- data
- abnormal
- pollution
- water quality
- data set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000003911 water pollution Methods 0.000 title claims abstract description 36
- 238000000034 method Methods 0.000 title claims abstract description 18
- 230000002159 abnormal effect Effects 0.000 claims abstract description 90
- 238000001514 detection method Methods 0.000 claims abstract description 26
- 238000003066 decision tree Methods 0.000 claims abstract description 24
- 238000012549 training Methods 0.000 claims abstract description 18
- 238000005457 optimization Methods 0.000 claims abstract description 9
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 53
- 230000003595 spectral effect Effects 0.000 claims description 23
- 238000012360 testing method Methods 0.000 claims description 22
- 238000012937 correction Methods 0.000 claims description 17
- 238000010276 construction Methods 0.000 claims description 4
- 238000004458 analytical method Methods 0.000 abstract description 4
- 230000000694 effects Effects 0.000 abstract description 4
- 230000005856 abnormality Effects 0.000 abstract 1
- 238000001228 spectrum Methods 0.000 description 5
- 239000003440 toxic substance Substances 0.000 description 3
- IJGRMHOSHXDMSA-UHFFFAOYSA-N Atomic nitrogen Chemical compound N#N IJGRMHOSHXDMSA-UHFFFAOYSA-N 0.000 description 2
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 239000001301 oxygen Substances 0.000 description 2
- 229910052760 oxygen Inorganic materials 0.000 description 2
- 231100000167 toxic agent Toxicity 0.000 description 2
- MMDJDBSEMBIJBB-UHFFFAOYSA-N [O-][N+]([O-])=O.[O-][N+]([O-])=O.[O-][N+]([O-])=O.[NH6+3] Chemical compound [O-][N+]([O-])=O.[O-][N+]([O-])=O.[O-][N+]([O-])=O.[NH6+3] MMDJDBSEMBIJBB-UHFFFAOYSA-N 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 229910052757 nitrogen Inorganic materials 0.000 description 1
- 231100000614 poison Toxicity 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 239000010865 sewage Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Abstract
The invention belongs to the technical field of water pollution analysis, and particularly relates to a water pollution alarm method based on a dynamic update data set, which comprises the steps of acquiring a training data set, constructing an abnormality detection model, actually measuring and operating real-time data and updating an abnormal data set A2 by abnormal data on the day. By replacing the detected abnormal data of the current day with the historical data in the abnormal data set A2, the database is continuously updated, the error of the abnormal threshold is reduced, and the problem of rare samples in the existing deep algorithm is solved. The model is constructed by calling the decision tree, so that further optimization of parameters can be realized on the basis, which is an effect that a PMF model cannot realize, the accuracy of a judgment result is improved, and the robustness is better.
Description
Technical Field
The invention belongs to the technical field of water pollution analysis, and relates to a water pollution alarm method, in particular to a water pollution alarm method based on a dynamic update data set.
Background
Sudden water pollution accidents can cause serious pollution and damage to water systems and ecological environments. Because the occurrence time and place of the sudden water pollution accident have great uncertainty, and meanwhile, the damage mode and the pollution degree are difficult to determine in a short time, the serious interference is easily caused to the normal social life and the production order. Therefore, it is necessary to find an effective and reliable pollution source tracing method, quickly and accurately find the occurrence time and location of the pollution source causing the sudden water pollution, make a correct decision, and take a feasible measure for treatment.
The characteristics of the sudden water pollution event mainly comprise the following two points:
(1) complexity. The water environment pollution has complex pollution source components, and may contain a single toxic substance or a novel toxic substance generated by the mutual reaction of a plurality of toxic substances; in addition, the interference factors are more in the open environment, and the water pollution event is easy to miss the alarm.
(2) Uncertainty. The occurrence place of the sudden water pollution event is not fixed, the occurrence time and the occurrence mode have uncertain characteristics, the occurrence rule of the water pollution accident is difficult to find out in time, and the concentration and the influence range of the pollution type are difficult to determine in a short time.
In general, in water quality monitoring, water quality parameters such as Chemical Oxygen Demand (COD), Biological Oxygen Demand (BOD), nitrate nitrogen (NO3-N), turbidity (FTU) and the like are used for reflecting whether water quality is abnormal or not, but the parameters reflect comprehensive information of organic matters in a water body, and a single or specific pollution source is difficult to judge. Currently, many researchers use a positive definite matrix factorization model (PMF) as an analysis tool, but in the PMF model, the number of factors needs to be determined manually, and the selection of different numbers of factors can cause uncertainty of load distribution of each parameter in an abnormal analysis process, and can generate a large error for a determination result of an abnormal event, so that the robustness is poor.
Nowadays, with the progress of artificial intelligence technology, many pollution detection methods are completed by utilizing deep learning, and the method has the advantages of high recognition speed, high accuracy and the like. However, the method using deep learning means that a large number of positive samples and negative samples are required, which is difficult to realize for water pollution with complex pollution source components and sudden water pollution, and especially difficult to obtain negative samples.
Disclosure of Invention
In order to solve the problems, the invention designs a water quality pollution alarm method based on a dynamic update data set, and solves the problems of rare pollution samples and poor robustness of the existing algorithm in reality.
The technical scheme adopted by the invention is that,
a water quality pollution alarm method based on a dynamic update data set comprises,
step 1: acquiring spectral data when the water quality is pollution-free as a characteristic data set A1;
step 2: acquiring spectral data when water quality is polluted to serve as an abnormal data set A2;
and step 3: a1 and A2 carry out data selection according to the proportion of a model training parameter M, wherein M is A1/A2;
and 4, step 4: obtaining a training data set A3 according to the selected A1 and A2, wherein A3 is A1U A2;
and 5: the training of the anomaly detection model is started, specifically,
step 501: randomly selecting m sample points from a training data set A3 to form n subsets omega i, wherein i belongs to 1,2 and 3 … … n, and constructing a decision tree on the n subsets;
step 502: randomly selecting one characteristic in omega i, and randomly selecting one threshold value for binary splitting; the threshold value is between the maximum value and the minimum value of the specified characteristics in the current node data;
step 503: repeating the step 502 until the decision tree reaches a set height d or only one point in each leaf node;
step 504: repeating the step 502 to the step 503 until the construction of the n decision trees is completed;
step 505: calculating the average depth of the n decision trees, outputting the average depth as an abnormal threshold value, and completing the construction of an abnormal detection model;
step 6: acquiring spectral data of water quality to be detected, and inputting the spectral data into an anomaly detection model to obtain a normalized anomaly score;
and 7: comparing the normalized abnormal score with an abnormal threshold, and judging that the water quality to be detected is pollution-free when the normalized abnormal score is smaller than the abnormal threshold; when the normalized abnormal fraction is greater than or equal to the abnormal threshold value, judging that the water quality to be detected is polluted, acquiring the spectral data of the water quality to be detected, and storing the spectral data as abnormal data on the day;
and 8: repeating the step 6 to the step 7 until the preset time;
and step 9: the abnormal data set a2 is updated according to the current day abnormal data during a set period of time.
Further, the step 7 also comprises the steps of judging the water quality pollution to be detected when the normalized abnormal fraction is larger than or equal to the abnormal threshold value, and starting an alarm.
Further, the invention also includes parameter optimization of the anomaly detection model, the parameter optimization includes,
step A: acquiring an A4 set of normal data judged by an abnormal detection model as abnormal data in a water quality pollution-free state, and defining the proportion of misjudging the normal data into the abnormal data by the abnormal detection model in the water quality pollution-free state as a false alarm rate X, wherein X is A4/A1;
and B: acquiring a set A5 of judging abnormal data into normal data by an abnormal detection model in a water pollution state, and defining the proportion of misjudging the abnormal data into the normal data by the abnormal detection model in the water pollution state as a false alarm rate Y, wherein Y is A5/A2;
and C: calling an isolated forest model, and counting false alarm rate and false alarm rate under different n _ estimators and max _ samples values, wherein n _ estimators are the number of subtrees, namely the number n of decision trees, and max _ samples are the number of training samples for constructing each subtree, namely the number m of sample points;
step D: and selecting the values of n _ estimators and max _ samples corresponding to the state with the lowest false alarm rate and the state with the lowest leakage alarm rate as the parameters of n and m after optimization respectively.
Further, said step 9 comprises, after said step,
during the period of entering the automatic calibration of the system, R pieces of data are extracted from abnormal data of the current day, wherein R is equal to 1,2.. 10 …, and the same amount of historical data in the abnormal data set A2 is replaced.
Further, the spectral data of the water quality to be detected is provided with a time tag, and the time tag comprises the acquisition time of the spectral data of the water quality to be detected.
Further, the present invention also includes anomaly threshold correction, said anomaly threshold correction comprising,
step a: acquiring a plurality of groups of test data, wherein the plurality of groups of data comprise test data of water pollution and test data of non-pollution water quality, and respectively calculating the average depth of a decision tree of each group of data;
step b: setting a correction threshold according to the average depth corresponding to the test data of water pollution and the test data of no water pollution, wherein the test data of water pollution is obtained when the average depth is lower than the correction threshold, and the test data of water pollution is obtained when the average depth is lower than the correction threshold;
step c: and correcting the abnormal threshold according to the correction threshold.
The working principle and the beneficial effects of the invention are as follows:
1. by replacing the detected abnormal data of the current day with the historical data in the abnormal data set A2, the database is continuously updated, the error of the abnormal threshold is reduced, and the problem of rare samples in the existing deep algorithm is solved.
2. The model is constructed by calling the decision tree, so that further optimization of parameters can be realized on the basis, which is an effect that a PMF model cannot realize, the accuracy of a judgment result is improved, and the robustness is better.
The present invention will be described in detail with reference to the accompanying drawings.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
The technical solutions of the present invention are further described in detail below with reference to specific examples and drawings, but the scope and implementation of the present invention are not limited thereto.
In a specific embodiment, as shown in figure 1,
the invention relates to a water quality pollution alarm method based on a dynamic update data set, which comprises the following steps,
1. generation of data sets
(1) Acquiring spectral data of non-polluted water quality from historical data to serve as a characteristic data set A1;
as shown in table 1:
| wave band | 190nm | 192.2nm | …… | 579.4nm | …… | 704.8nm | …… | 748.8nm | 751.0nm |
| Value of voltage | 125 | 122 | …… | 8288 | …… | 3116 | …… | 1902 | 1825 |
| Value of voltage | 125 | 122 | …… | 8296 | …… | 3121 | …… | 1898 | 1828 |
| Value of voltage | 125 | 122 | …… | 8301 | …… | 3135 | …… | 1907 | 1838 |
| Value of voltage | 126 | 122 | …… | 8315 | …… | 3089 | …… | 1896 | 1823 |
| Value of voltage | 125 | 122 | …… | 8294 | …… | 3106 | …… | 1892 | 1827 |
| Value of voltage | 126 | 122 | …… | 8296 | …… | 3086 | …… | 1882 | 1819 |
| …… | …… | …… | …… | …… | …… | …… | …… | …… | …… |
Each row in Table 1 is a set of feature data sets, and in this embodiment, because the data band of the micro spectrometer covers 190-750nm, there are 256 sub-bands in this band to provide the spectral data.
(2) Acquiring spectral data of water pollution from historical data to serve as an abnormal data set A2;
as shown in table 2: to scan the measured value of COD exceeding sewage for multiple times by using a micro spectrometer
| Wave band | 190nm | 192.2nm | …… | 682.8nm | …… | 698.2nm | …… | 748.8nm | 751.0nm |
| Value of voltage | 127 | 127 | …… | 1109 | …… | 544 | …… | 199 | 173 |
| Value of voltage | 127 | 126 | …… | 1090 | …… | 544 | …… | 196 | 173 |
| Value of voltage | 126 | 126 | …… | 1090 | …… | 541 | …… | 192 | 173 |
| Value of voltage | 126 | 126 | …… | 1085 | …… | 531 | …… | 189 | 168 |
| Value of voltage | 126 | 126 | …… | 1094 | …… | 547 | …… | 194 | 173 |
| Value of voltage | 126 | 126 | …… | 1083 | …… | 536 | …… | 191 | 171 |
| …… | …… | …… | …… | …… | …… | …… | …… | …… | …… |
Each row in Table 2 is a set of abnormal data sets, and in this embodiment, because the data band of the micro spectrometer covers 190-750nm, there are 256 sub-bands in this band to provide the spectral data.
(3) A1 and A2 carry out data selection according to the proportion of a model training parameter M, wherein M is A1/A2; the model training parameter M is used as an adjustable parameter in a program, and a user can modify the model training parameter M according to the actual situation on site, so that the proportion of the quantity of the data selected from A1 and A2 is determined, and a more appropriate training data set is better constructed according to the actual situation.
(4) And obtaining a training data set A3 according to the selected A1 and A2, wherein A3 is A1U A2.
2. Constructing an anomaly detection model
(1) Randomly selecting m sample points from the training data set A3 to form n subsets omegai(i ∈ 1,2,3 … … n), building a decision tree on the n subsets;
(2) selecting omegaiTaking a subset of the unstructured decision tree as the characteristic of decision, taking data in the subset as current node data of the decision tree, and randomly selecting a threshold value for binary splitting; the threshold value is between the maximum value and the minimum value of the specified characteristics in the current node data;
(3) repeating the step (2) to construct a decision tree until the decision tree reaches a set height d or only one point in each leaf is reached;
(4) after n decision trees are built, defining an abnormal threshold according to the average depth of the n decision trees;
when the data set is generated, the input abnormal data set A2 is historical water quality pollution spectrum data, and therefore, the result output by the abnormal detection model is defined as an abnormal threshold value;
3. optimizing model parameters
(1) Acquiring an A4 set of normal data judged by an abnormal detection model as abnormal data in a water quality pollution-free state, and defining the proportion of misjudging the normal data into the abnormal data by the abnormal detection model in the water quality pollution-free state as a false alarm rate X, wherein X is A4/A1;
(2) acquiring a set A5 of judging abnormal data into normal data by an abnormal detection model in a water pollution state, and defining the proportion of misjudging the abnormal data into the normal data by the abnormal detection model in the water pollution state as a false alarm rate Y, wherein Y is A5/A2;
(3) in order to reduce false alarm rate and false alarm rate as much as possible and improve modeling effect, the default parameter values of the isolated forest need to be adjusted. Wherein n _ estimators are the number of subtrees, the isolated forest is composed of the subtrees, the final judgment result is determined by all the subtrees together, and the n _ estimators are the value of the number n of the decision trees; and max _ samples is the number of training samples for constructing each subtree, namely the value of the number m of sample points in the decision tree.
(4) And counting false alarm rates and false alarm rate under the states of different values of n _ estimators and max _ samples, and selecting the values of the n _ estimators and max _ samples corresponding to the states of the lowest false alarm rates and the lowest false alarm rates as parameters of n and m after optimization. Through multiple modeling tests, when the max _ samples parameter of the model is set to 300 and the n _ estimators parameter is set to 150, the average false alarm rate is 1.8%, the average false alarm rate is 0.9%, and the model effect is ideal.
4. Actual measurement operation
Acquiring spectral data of the water quality to be detected, inputting the spectral data into an anomaly detection model, outputting a normalized anomaly score by the model, comparing the normalized anomaly score with an anomaly threshold, determining that the spectral data of the water quality to be detected is not abnormal data when the anomaly score is smaller than the anomaly threshold, defining that the water quality to be detected is pollution-free, and repeating the step 6 to detect the next group of data; when the abnormal fraction is larger than the abnormal threshold value, the spectrum data of the water quality to be detected is abnormal data, the water quality pollution to be detected is defined, and an alarm is started;
according to multiple modeling, an abnormal threshold value is 0.55, the inputted spectrum data of the water quality to be detected is assumed to be [ 127.127........ 1109.......... 173], and a normalized abnormal score output by the abnormal detection model with optimized parameters is 0.68, and because 0.68 is larger than 0.55, the water quality to be detected corresponding to the data is judged to be polluted, and alarm processing is carried out.
5. Dynamically updating a sample library
When the system automatic calibration time interval (generally 00:00-01:00) is entered, R pieces of data (R is set by people for 1,2.. 10 …) are extracted from the abnormal data of the current day, the same amount of historical data in the abnormal data set A2 is replaced, and then the optimized model parameters are updated correspondingly.
The time label is included in the spectrum data of the water quality to be detected, the time label includes the acquisition time of the spectrum data of the water quality to be detected, namely, each abnormal data is also provided with the time label, the replaced abnormal data are sequentially replaced according to the sequence of the time labels, and the time of the time label is longer and earlier, so that the accuracy and the timeliness of the abnormal threshold value are ensured in the updating iteration.
6. Anomaly threshold correction
Acquiring average depths of a plurality of groups of test data decision trees, wherein the plurality of groups of test data comprise test data of water pollution and test data of non-pollution water, and the test data are used as different A2 inputs in the generation of a data set; and manually setting a correction threshold value through a plurality of groups of average depths, so that the average depth is lower than the correction threshold value and is test data without water pollution, the average depth is lower than the correction threshold value and is test data with water pollution, and correcting the abnormal threshold value according to the correction threshold value.
Claims (6)
1. A water quality pollution alarm method based on a dynamic update data set is characterized by comprising the following steps,
step 1: acquiring spectral data when the water quality is pollution-free as a characteristic data set A1;
step 2: acquiring spectral data when water quality is polluted to serve as an abnormal data set A2;
and step 3: a1 and A2 carry out data selection according to the proportion of a model training parameter M, wherein M is A1/A2;
and 4, step 4: obtaining a training data set A3 according to the selected A1 and A2, wherein A3 is A1U A2;
and 5: the training of the anomaly detection model is started, specifically,
step 501: randomly selecting m sample points from a training data set A3 to form n subsets omega i, wherein i belongs to 1,2 and 3 … … n, and constructing a decision tree on the n subsets;
step 502: randomly selecting one characteristic in omega i, and randomly selecting one threshold value for binary splitting; the threshold value is between the maximum value and the minimum value of the specified characteristics in the current node data;
step 503: repeating the step 502 until the decision tree reaches a set height d or only one point in each leaf node;
step 504: repeating the step 502 to the step 503 until the construction of the n decision trees is completed;
step 505: calculating the average depth of the n decision trees, outputting the average depth as an abnormal threshold value, and completing the construction of an abnormal detection model;
step 6: acquiring spectral data of the water quality to be detected, inputting the spectral data into an anomaly detection model to obtain a normalized anomaly score
And 7: comparing the normalized abnormal score with an abnormal threshold, and judging that the water quality to be detected is pollution-free when the normalized abnormal score is smaller than the abnormal threshold; when the normalized abnormal fraction is greater than or equal to the abnormal threshold value, judging that the water quality to be detected is polluted, acquiring the spectral data of the water quality to be detected, and storing the spectral data as abnormal data on the day;
and 8: repeating the step 6 to the step 7 until the preset time;
and step 9: the abnormal data set a2 is updated according to the current day abnormal data during a set period of time.
2. The water quality pollution alarm method based on the dynamic update data set as claimed in claim 1, wherein the step 7 further comprises judging the water quality pollution to be detected and starting an alarm when the normalized abnormal score is greater than or equal to the abnormal threshold.
3. The water quality pollution alarm method based on the dynamic update data set as claimed in claim 1, further comprising parameter optimization of the anomaly detection model, wherein the parameter optimization comprises,
step A: acquiring an A4 set of normal data judged by an abnormal detection model as abnormal data in a water quality pollution-free state, and defining the proportion of misjudging the normal data into the abnormal data by the abnormal detection model in the water quality pollution-free state as a false alarm rate X, wherein X is A4/A1;
and B: acquiring a set A5 of judging abnormal data into normal data by an abnormal detection model in a water pollution state, and defining the proportion of misjudging the abnormal data into the normal data by the abnormal detection model in the water pollution state as a false alarm rate Y, wherein Y is A5/A2;
and C: calling an isolated forest model, and counting false alarm rate and false alarm rate under different n _ estimators and max _ samples values, wherein n _ estimators are the number of subtrees, namely the number n of decision trees, and max _ samples are the number of training samples for constructing each subtree, namely the number m of sample points;
step D: and selecting the values of n _ estimators and max _ samples corresponding to the state with the lowest false alarm rate and the state with the lowest leakage alarm rate as the parameters of n and m after optimization respectively.
4. The water quality pollution alarm method based on the dynamic update data set according to the claim 1, wherein the step 9 comprises extracting R pieces of data from abnormal data in the day in an automatic calibration period of the entering system, wherein R is equal to 1,2.. 10 …, and replaces the same amount of historical data in the abnormal data set A2.
5. The water quality pollution alarm method based on the dynamic update data set as claimed in claim 1, wherein the spectral data of the water quality to be measured is provided with a time tag, and the time tag comprises the acquisition time of the spectral data of the water quality to be measured.
6. The water quality pollution alarm method based on the dynamic update data set as claimed in claim 1, further comprising abnormal threshold correction, wherein the abnormal threshold correction comprises,
step a: acquiring a plurality of groups of test data, wherein the plurality of groups of data comprise test data of water pollution and test data of non-pollution water quality, and respectively calculating the average depth of a decision tree of each group of data;
step b: setting a correction threshold according to the average depth corresponding to the test data of water pollution and the test data of no water pollution, wherein the test data of water pollution is obtained when the average depth is lower than the correction threshold, and the test data of water pollution is obtained when the average depth is lower than the correction threshold;
step c: and correcting the abnormal threshold according to the correction threshold.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110585040.7A CN113295635A (en) | 2021-05-27 | 2021-05-27 | Water pollution alarm method based on dynamic update data set |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110585040.7A CN113295635A (en) | 2021-05-27 | 2021-05-27 | Water pollution alarm method based on dynamic update data set |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113295635A true CN113295635A (en) | 2021-08-24 |
Family
ID=77325617
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110585040.7A Pending CN113295635A (en) | 2021-05-27 | 2021-05-27 | Water pollution alarm method based on dynamic update data set |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113295635A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116206427A (en) * | 2023-05-06 | 2023-06-02 | 安徽智寰科技有限公司 | Hierarchical alarm method based on universal index self-adaptive threshold |
Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101478401A (en) * | 2009-01-21 | 2009-07-08 | 东北大学 | Authentication method and system based on key stroke characteristic recognition |
| US20120213534A1 (en) * | 2011-02-18 | 2012-08-23 | Konica Minolta Business Technologies, Inc. | Image formation device |
| CN103307447A (en) * | 2013-06-03 | 2013-09-18 | 清华大学 | Technical failure information monitoring and early warning system for urban gas pipe network |
| US20160182544A1 (en) * | 2015-02-28 | 2016-06-23 | Brighterion, Inc. | Method of protecting a network computer system from the malicious acts of hackers and its own system administrators |
| CN106295684A (en) * | 2016-08-02 | 2017-01-04 | 清华大学 | A kind of the most continuous based on micro-Doppler feature/discontinuous gesture recognition methods |
| CN107645707A (en) * | 2017-08-28 | 2018-01-30 | 上海斐讯数据通信技术有限公司 | A kind of update method and system of the positioning file based on intelligent terminal |
| CN108153146A (en) * | 2017-12-12 | 2018-06-12 | 东北大学 | A kind of polynary molten steel quality MFA control system and method for blast furnace |
| CN109086808A (en) * | 2018-07-16 | 2018-12-25 | 江苏智通交通科技有限公司 | Traffic high-risk personnel recognition methods based on random forests algorithm |
| CN109288536A (en) * | 2018-09-30 | 2019-02-01 | 数坤（北京）网络科技有限公司 | Obtain the method, apparatus and system of Coronary Calcification territorial classification |
| US20190156198A1 (en) * | 2017-11-22 | 2019-05-23 | Clinc, Inc. | System and method for implementing an artificially intelligent virtual assistant using machine learning |
| US20190166024A1 (en) * | 2017-11-24 | 2019-05-30 | Institute For Information Industry | Network anomaly analysis apparatus, method, and non-transitory computer readable storage medium thereof |
| CN109828869A (en) * | 2018-12-05 | 2019-05-31 | 中兴通讯股份有限公司 | Predict the method, apparatus and storage medium of hard disk failure time of origin |
| CN109884033A (en) * | 2019-04-08 | 2019-06-14 | 西安石油大学 | The method of random forests algorithm combination laser induced breakdown spectroscopy detection metallic element |
| CN110268419A (en) * | 2019-05-08 | 2019-09-20 | 深圳市汇顶科技股份有限公司 | A kind of face identification method, face identification device and computer readable storage medium |
| CN110399935A (en) * | 2019-08-02 | 2019-11-01 | 哈工大机器人（合肥）国际创新研究院 | The real-time method for monitoring abnormality of robot and system based on isolated forest machine learning |
| CN110647909A (en) * | 2019-08-08 | 2020-01-03 | 广东工业大学 | Remote sensing image classification method based on three-dimensional dense convolution neural network |
| CN110717600A (en) * | 2019-09-30 | 2020-01-21 | 京东城市(北京)数字科技有限公司 | Sample pool construction method and device, and algorithm training method and device |
| CN111353352A (en) * | 2018-12-24 | 2020-06-30 | 杭州海康威视数字技术股份有限公司 | Abnormal behavior detection method and device |
| CN111444931A (en) * | 2019-01-17 | 2020-07-24 | 北京京东尚科信息技术有限公司 | Method and device for detecting abnormal access data |
| CN111783953A (en) * | 2020-06-30 | 2020-10-16 | 重庆大学 | 24-point power load value 7-day prediction method based on optimized LSTM network |
| CN111985567A (en) * | 2020-08-21 | 2020-11-24 | 河北先河环保科技股份有限公司 | Automatic pollution source type identification method based on machine learning |
| CN112148908A (en) * | 2020-10-23 | 2020-12-29 | 北京百度网讯科技有限公司 | Image database updating method and device, electronic equipment and medium |
-
2021
- 2021-05-27 CN CN202110585040.7A patent/CN113295635A/en active Pending
Patent Citations (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101478401A (en) * | 2009-01-21 | 2009-07-08 | 东北大学 | Authentication method and system based on key stroke characteristic recognition |
| US20120213534A1 (en) * | 2011-02-18 | 2012-08-23 | Konica Minolta Business Technologies, Inc. | Image formation device |
| CN103307447A (en) * | 2013-06-03 | 2013-09-18 | 清华大学 | Technical failure information monitoring and early warning system for urban gas pipe network |
| US20160182544A1 (en) * | 2015-02-28 | 2016-06-23 | Brighterion, Inc. | Method of protecting a network computer system from the malicious acts of hackers and its own system administrators |
| CN106295684A (en) * | 2016-08-02 | 2017-01-04 | 清华大学 | A kind of the most continuous based on micro-Doppler feature/discontinuous gesture recognition methods |
| CN107645707A (en) * | 2017-08-28 | 2018-01-30 | 上海斐讯数据通信技术有限公司 | A kind of update method and system of the positioning file based on intelligent terminal |
| US20190156198A1 (en) * | 2017-11-22 | 2019-05-23 | Clinc, Inc. | System and method for implementing an artificially intelligent virtual assistant using machine learning |
| US20190166024A1 (en) * | 2017-11-24 | 2019-05-30 | Institute For Information Industry | Network anomaly analysis apparatus, method, and non-transitory computer readable storage medium thereof |
| CN108153146A (en) * | 2017-12-12 | 2018-06-12 | 东北大学 | A kind of polynary molten steel quality MFA control system and method for blast furnace |
| CN109086808A (en) * | 2018-07-16 | 2018-12-25 | 江苏智通交通科技有限公司 | Traffic high-risk personnel recognition methods based on random forests algorithm |
| CN109288536A (en) * | 2018-09-30 | 2019-02-01 | 数坤（北京）网络科技有限公司 | Obtain the method, apparatus and system of Coronary Calcification territorial classification |
| CN109828869A (en) * | 2018-12-05 | 2019-05-31 | 中兴通讯股份有限公司 | Predict the method, apparatus and storage medium of hard disk failure time of origin |
| CN111353352A (en) * | 2018-12-24 | 2020-06-30 | 杭州海康威视数字技术股份有限公司 | Abnormal behavior detection method and device |
| CN111444931A (en) * | 2019-01-17 | 2020-07-24 | 北京京东尚科信息技术有限公司 | Method and device for detecting abnormal access data |
| CN109884033A (en) * | 2019-04-08 | 2019-06-14 | 西安石油大学 | The method of random forests algorithm combination laser induced breakdown spectroscopy detection metallic element |
| CN110268419A (en) * | 2019-05-08 | 2019-09-20 | 深圳市汇顶科技股份有限公司 | A kind of face identification method, face identification device and computer readable storage medium |
| CN110399935A (en) * | 2019-08-02 | 2019-11-01 | 哈工大机器人（合肥）国际创新研究院 | The real-time method for monitoring abnormality of robot and system based on isolated forest machine learning |
| CN110647909A (en) * | 2019-08-08 | 2020-01-03 | 广东工业大学 | Remote sensing image classification method based on three-dimensional dense convolution neural network |
| CN110717600A (en) * | 2019-09-30 | 2020-01-21 | 京东城市(北京)数字科技有限公司 | Sample pool construction method and device, and algorithm training method and device |
| CN111783953A (en) * | 2020-06-30 | 2020-10-16 | 重庆大学 | 24-point power load value 7-day prediction method based on optimized LSTM network |
| CN111985567A (en) * | 2020-08-21 | 2020-11-24 | 河北先河环保科技股份有限公司 | Automatic pollution source type identification method based on machine learning |
| CN112148908A (en) * | 2020-10-23 | 2020-12-29 | 北京百度网讯科技有限公司 | Image database updating method and device, electronic equipment and medium |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116206427A (en) * | 2023-05-06 | 2023-06-02 | 安徽智寰科技有限公司 | Hierarchical alarm method based on universal index self-adaptive threshold |
| CN116206427B (en) * | 2023-05-06 | 2023-06-30 | 安徽智寰科技有限公司 | Hierarchical alarm method based on universal index self-adaptive threshold |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11346831B2 (en) | Intelligent detection method for biochemical oxygen demand based on a self-organizing recurrent RBF neural network | |
| CN109408774B (en) | Method for predicting sewage effluent index based on random forest and gradient lifting tree | |
| CN116186566B (en) | Diffusion prediction method and system based on deep learning | |
| CN109558893B (en) | Rapid integrated sewage treatment fault diagnosis method based on resampling pool | |
| CN105335752A (en) | Principal component analysis multivariable decision-making tree-based connection manner identification method | |
| CN108319788B (en) | Method for identifying direct sewage discharge pollution source of rainwater pipe network | |
| CN112101789A (en) | Water pollution alarm grade identification method based on artificial intelligence | |
| CN113344288B (en) | Cascade hydropower station group water level prediction method and device and computer readable storage medium | |
| Bellandi et al. | Towards an online mitigation strategy for N2O emissions through principal components analysis and clustering techniques | |
| CN112735541A (en) | Sewage treatment water quality prediction method based on simple circulation unit neural network | |
| CN106935038B (en) | Parking detection system and detection method | |
| CN113295635A (en) | Water pollution alarm method based on dynamic update data set | |
| CN116432123A (en) | Electric energy meter fault early warning method based on CART decision tree algorithm | |
| CN104063609B (en) | Method of assisting in judging pollution source monitoring data validity by utilizing neural network | |
| CN111311026A (en) | Runoff nonlinear prediction method considering data characteristics, model and correction | |
| CN115358481A (en) | Early warning and identification method, system and device for enterprise ex-situ migration | |
| CN114757286A (en) | Multi-class fault data generation method based on conditional countermeasure generation network | |
| CN115064218B (en) | Method and device for constructing pathogenic microorganism data identification platform | |
| CN116629918A (en) | User consumption prediction method and system based on cross-border electronic commerce | |
| CN116702926A (en) | Air quality mode forecasting machine learning integrated correction method | |
| CN116720743A (en) | Carbon emission measuring and calculating method based on data clustering and machine learning | |
| CN112986497B (en) | Pollution gas tracing method based on gas sensor array fingerprint identification | |
| CN114741972A (en) | Construction method of seasonal prediction model of air pollutant concentration | |
| KR20230038038A (en) | Energy efficiency evaluation method and energy saving decision-making system of sewage treatment plant based on water treatment energy index | |
| CN111340095A (en) | Environmental monitoring data quality control method based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210824 |