CN-121981288-A - Water immersion identification method and device, computer equipment and storage medium

CN121981288ACN 121981288 ACN121981288 ACN 121981288ACN-121981288-A

Abstract

The application belongs to the technical field of artificial intelligence and relates to a water logging identification method, a device, computer equipment and a storage medium, wherein the method comprises the steps of receiving real-time water logging data acquired by water logging data acquisition equipment, wherein the real-time water logging data comprise real-time sensor data and real-time image data; the method comprises the steps of carrying out state estimation processing on real-time sensor data according to a Kalman filtering algorithm to obtain a water logging state estimation value, carrying out visual verification processing on the real-time image data according to an image recognition technology to obtain visual water logging confidence level, carrying out fuzzy reasoning processing according to fuzzy rules, the water logging state estimation value and the visual water logging confidence level to obtain a fuzzy reasoning result, carrying out defuzzification processing on the fuzzy reasoning result to obtain accurate alarm confidence level, and confirming the water logging recognition result according to the accurate alarm confidence level. According to the application, through fusion analysis and intelligent reasoning of the multi-source information, accurate identification and reliable early warning of the water immersion state are realized.

Inventors

GAO YONG
ZHAO CHANGJIANG
WANG RUI
YAN SHIYING
Mou Jingying

Assignees

青岛微智慧信息有限公司

Dates

Publication Date: 20260505
Application Date: 20260403

Claims (10)

1. A water immersion identification method, comprising the steps of: receiving real-time water logging data acquired by water logging data acquisition equipment, wherein the real-time water logging data comprise real-time sensor data and real-time image data; Performing state estimation processing on the real-time sensor data according to a Kalman filtering algorithm to obtain a water immersion state estimation value; Performing visual verification processing on the real-time image data according to an image recognition technology to obtain visual water confidence; carrying out fuzzy reasoning processing according to the fuzzy rule, the water immersion state estimation value and the visual water immersion confidence level to obtain a fuzzy reasoning result; performing defuzzification processing on the fuzzy reasoning result to obtain accurate alarm confidence; and confirming the water immersion identification result according to the accurate alarm confidence level.
2. The water logging identification method according to claim 1, wherein the step of performing state estimation processing on the real-time sensor data according to a kalman filter algorithm to obtain a water logging state estimation value specifically comprises the following steps: obtaining an estimated value X k-1 of the water immersion state at the previous moment; Predicting a current water logging state estimated value X k|k-1 according to the previous water logging state estimated value X k-1 , wherein the current water logging state estimated value X k|k-1 is expressed as: X k|k-1 =F k X k-1 ; wherein F k represents a state transition matrix; and performing measurement update processing on the water logging state estimated value X k|k-1 at the current moment according to the real-time sensor data Z k to obtain the water logging state estimated value X k , wherein the water logging state estimated value X k is expressed as: X k =X k|k-1 +K k （Z k -H k X k|k-1 ); Wherein X k|k-1 represents the water immersion state estimated value at the previous time, K k represents the kalman gain, Z k represents the real-time sensor data, and H k represents the observation matrix.
3. The water logging identification method according to claim 2, wherein before the step of obtaining the water logging state estimation value by performing measurement update processing on the water logging state estimation value at the current time according to the real-time sensor data, the method further comprises the steps of: Acquiring an error covariance P k-1 updated at the previous moment; Calculating the predicted error covariance P k|k-1 of the current moment according to the updated error covariance P k-1 of the previous moment, wherein the predicted error covariance P k|k-1 of the current moment is expressed as: P k|k-1 =F k P k-1 F k T +Q k ; Wherein F k represents a state transition matrix, F k T represents a transpose matrix, and Q k represents a process noise covariance; Performing covariance update processing on the error covariance P k|k-1 predicted at the current moment to obtain the kalman gain K k, , wherein the kalman gain K k is expressed as: K k =P k|k-1 H k T (H k P k|k-1 H k T +R k ） -1 ; where R k represents the observed noise covariance.
4. The method for water immersion identification according to claim 1, wherein the step of performing visual verification processing on the real-time image data according to an image identification technology to obtain visual water immersion confidence level comprises the following steps: performing image preprocessing on the real-time image data to obtain standard image data; carrying out water immersion image characteristic recognition processing on the standard image data to obtain direct characteristic data and indirect characteristic data; Performing feature fusion processing on the direct feature data and the indirect feature data to obtain fusion feature data; and predicting the probability of the real water immersion condition in the real-time image data according to the fusion characteristic data to obtain the visual water immersion confidence.
5. The water logging identification method according to claim 1, wherein the step of performing fuzzy inference processing according to a fuzzy rule, the water logging state estimation value and the visual water logging confidence level to obtain a fuzzy inference result specifically comprises the following steps: calculating membership degrees of all condition rules in the fuzzy rule according to the water immersion state estimated value and the visual water immersion confidence degree; calculating the activation intensity of each condition rule according to the membership degree of the previous condition rule of each condition rule; Generating rule output according to the activation intensity; And carrying out aggregation processing on the rule output to obtain the fuzzy reasoning result.
6. The water logging identification method according to claim 1, wherein the step of performing defuzzification processing on the fuzzy inference result to obtain an accurate alarm confidence level comprises the following steps: Calculating the centroid of the fuzzy inference result according to a barycenter method to obtain the accurate alarm confidence coefficient, wherein the accurate alarm confidence coefficient Y final is expressed as: ; Where x i represents the discrete points on the output domain, μ (x i ) represents the total membership at point x i , and n represents the total number of discrete points.
7. The water logging identification method according to claim 1, wherein the step of confirming the water logging identification result according to the accurate alarm confidence level comprises the following steps: If the accurate alarm confidence coefficient is larger than a first confidence coefficient threshold value, determining that the accurate alarm confidence coefficient is immersed in water, and triggering a primary alarm; if the accurate alarm confidence coefficient is smaller than or equal to the first confidence coefficient threshold value and larger than or equal to the second confidence coefficient threshold value, determining that the accurate alarm confidence coefficient is immersed in water, and triggering a secondary alarm; and if the accurate alarm confidence coefficient is smaller than the second confidence coefficient threshold value, confirming that the accurate alarm confidence coefficient is not immersed in water, not triggering an alarm, and recording the real-time immersed water data and a log corresponding to the accurate alarm confidence coefficient.
8. A water immersion identification apparatus, comprising: The real-time data acquisition module is used for receiving real-time water logging data acquired by the water logging data acquisition equipment, wherein the real-time water logging data comprise real-time sensor data and real-time image data; the state estimation module is used for carrying out state estimation processing on the real-time sensor data according to a Kalman filtering algorithm to obtain a water immersion state estimation value; the visual verification module is used for performing visual verification processing on the real-time image data according to an image recognition technology to obtain visual water confidence; the fuzzy reasoning module is used for carrying out fuzzy reasoning processing according to the fuzzy rule, the water immersion state estimated value and the visual water immersion confidence level to obtain a fuzzy reasoning result; The defuzzification module is used for defuzzifying the fuzzy reasoning result to obtain accurate alarm confidence; And the result confirmation module is used for confirming the water immersion identification result according to the accurate alarm confidence coefficient.
9. A computer device comprising a memory and a processor, wherein the memory has stored therein computer readable instructions which when executed by the processor implement the steps of the water logging identification method of any one of claims 1 to 7.
10. A computer readable storage medium having stored thereon computer readable instructions which when executed by a processor implement the steps of the water logging identification method according to any of claims 1 to 7.

Description

Water immersion identification method and device, computer equipment and storage medium Technical Field The present application relates to the field of artificial intelligence technologies, and in particular, to a water immersion identification method, a device, a computer device, and a storage medium. Background The secondary water supply pump house is an important facility of the urban water supply system, is usually positioned in an underground space or a building bottom layer and plays a key role in providing stable water pressure for resident users. Due to its special location and environmental conditions, the pump house is exposed to multiple water logging risks such as pipe rupture, equipment leakage, groundwater infiltration, extreme weather leading to rain water backflow. These risks are particularly pronounced in the case of ageing and untimely maintenance of the pump house equipment. Once water immersion accidents occur, expensive water pump units and electrical control equipment can be damaged, hundreds of thousands or even millions of economic losses are caused, large-scale water supply interruption can be caused, normal life of residents is seriously influenced, even secondary safety accidents such as electric shock and the like are caused due to electric leakage of electrical equipment, and life safety of personnel is threatened. Currently, most pump houses employ conventional water immersion sensors for safety monitoring, such as point electrode sensors or cable sensors. Such sensors operate on a simple conduction principle, triggering an alarm when the liquid contacts the detection site. However, the traditional monitoring mode has lower cost and simple installation, but has obvious defects in actual operation. The sensor is easy to generate false alarm due to instantaneous interference such as environmental moisture condensation, water droplet splashing due to water pipe condensation, water droplet splashing caused by equipment operation vibration, dust accumulation forming a conductive path or spider-web and the like. Especially in the plum rainy season or in the environment with larger temperature difference, the false alarm phenomenon is more frequent. Frequent false alarms not only increase the workload of operation and maintenance personnel, but also cause 'wolf' effect more seriously, so that the staff can paralyze and ignore alarm information, and delay false disposal occurs in real water immersion, thereby causing serious consequences. Therefore, the traditional detection mode has the problem of lower detection accuracy. Disclosure of Invention The embodiment of the application aims to provide a water immersion identification method, a water immersion identification device, computer equipment and a storage medium, so as to solve the problem of lower detection accuracy in a traditional detection mode. In order to solve the technical problems, the embodiment of the application provides a water immersion identification method, which adopts the following technical scheme: receiving real-time water logging data acquired by water logging data acquisition equipment, wherein the real-time water logging data comprise real-time sensor data and real-time image data; Performing state estimation processing on the real-time sensor data according to a Kalman filtering algorithm to obtain a water immersion state estimation value; Performing visual verification processing on the real-time image data according to an image recognition technology to obtain visual water confidence; carrying out fuzzy reasoning processing according to the fuzzy rule, the water immersion state estimation value and the visual water immersion confidence level to obtain a fuzzy reasoning result; performing defuzzification processing on the fuzzy reasoning result to obtain accurate alarm confidence; and confirming the water immersion identification result according to the accurate alarm confidence level. In order to solve the technical problems, the embodiment of the application also provides a water immersion identification device, which adopts the following technical scheme: The real-time data acquisition module is used for receiving real-time water logging data acquired by the water logging data acquisition equipment, wherein the real-time water logging data comprise real-time sensor data and real-time image data; the state estimation module is used for carrying out state estimation processing on the real-time sensor data according to a Kalman filtering algorithm to obtain a water immersion state estimation value; the visual verification module is used for performing visual verification processing on the real-time image data according to an image recognition technology to obtain visual water confidence; the fuzzy reasoning module is used for carrying out fuzzy reasoning processing according to the fuzzy rule, the water immersion state estimated value and the visual water immersion confidence level to obtain a fuzzy reason