KR-102961780-B1 - SYSTEM AND APPARATUS FOR ROBUST L-BAND LEAKAGE DETECTION IN WET CONDITIONS

KR102961780B1KR 102961780 B1KR102961780 B1KR 102961780B1KR-102961780-B1

Abstract

An L-band based water and sewage leak detection device robust in a wet environment according to an embodiment of the present invention comprises: an image receiving unit that receives L-band synthetic aperture radar images from a satellite; an environmental data collection unit that collects environmental data including rainfall amount, relative humidity, soil type, terrain slope, and location information of water and sewage pipes; a signal distortion prediction unit that fuses the synthetic aperture radar images and the environmental data to predict the amount of signal distortion according to environmental conditions; an image correction unit that corrects the original synthetic aperture radar images in an environment-adaptive manner based on the predicted amount of signal distortion; a leak detection unit that analyzes candidate leak areas around water and sewage pipes based on the corrected images; and a learning management unit that receives field verification results for the candidate leak areas and updates the parameters of the signal distortion prediction unit and the leak detection unit.

Inventors

이성욱
이상덕
이재혁
김혜정
이민재

Assignees

수자원기술 주식회사

Dates

Publication Date: 20260512
Application Date: 20251219

Claims (10)

In a device for detecting water and sewage pipe leaks using L-band synthetic aperture radar imaging in a humid environment, SAR image receiver for receiving L-band synthetic aperture radar images from a satellite; An environmental data collection unit that collects environmental data including rainfall, relative humidity, soil type, topographic slope, and location information of water and sewage pipelines; A signal distortion prediction unit that fuses the synthetic aperture radar image and the environmental data to predict the amount of signal distortion according to environmental conditions; An image correction unit that corrects the original synthetic aperture radar image in an environment-adaptive manner based on the above-mentioned predicted signal distortion amount; A leak detection unit that analyzes candidate leak areas around water and sewer pipes based on the above-mentioned corrected image; and A device comprising a learning management unit that receives field verification results for the above-mentioned leak candidate area and updates parameters of a signal distortion prediction unit and a leak detection unit.
In paragraph 1, The above signal distortion prediction unit An apparatus comprising a multivariate prediction model that predicts the amount of change in scattering intensity of the synthetic aperture radar image using at least one of rainfall, soil moisture, relative humidity, and terrain slope as an input value.
In paragraph 1, The above signal distortion prediction unit A device implemented as at least one of a random forest-based model or a prediction model using long-term time series data, and which quantitatively calculates the distortion rate of each environmental variable affecting an L-band synthetic aperture radar signal.
In paragraph 1, The above image correction unit is a device that automatically raises the detection threshold when rainfall is high according to the predicted amount of signal distortion, strengthens the scattering intensity correction filter in clay areas, and applies an environment-adaptive correction algorithm that separates and removes building reflection components in urban areas.
In paragraph 1, The above image correction unit is a device that automatically adjusts speckle noise intensity according to a predicted humidity change pattern and increases the weight of a speckle noise removal filter for areas where humidity has increased.
In paragraph 1, The above leak detection unit is, A device comprising an image analysis model that detects at least one of an abnormal moisture pattern around a water pipe path, a local scattering rise area not explained by environmental changes, and an abnormal pattern of temporal rate of change in a corrected synthetic aperture radar image.
In paragraph 6, The above image analysis model is, A device implemented with a synthetic neural network-based detection model that learns the spatial patterns of synthetic aperture radar images, wherein the detection model divides a suspected leakage area within an image into pixel units and outputs candidate leakage points.
In paragraph 1, The above learning management department is, A device that receives verification data including the location of a leak confirmed at the site, whether a leak exists, and whether there is a false positive, updates the environmental variable weights of a signal distortion prediction unit using the verification data, and retrains the image analysis model of a leak detection unit.
In paragraph 1, The above leak detection unit is, A device that calculates the temporal rate of change by comparing a corrected synthetic aperture radar image with an image from a previous time point, detects a continuous moisture increase pattern independent of environmental changes, and assigns priority to candidate leakage areas.
A method for detecting water and sewage leaks using L-band synthetic aperture radar imaging in a humid environment, A step of receiving a synthetic aperture radar image at an image receiving unit; A step of collecting environmental data, such as rainfall, relative humidity, soil type, terrain slope, and pipeline location, from an environmental data receiving unit; A step of predicting the amount of signal distortion for each environmental condition by fusing the synthetic aperture radar image and the environmental data in the signal distortion prediction unit; A step of correcting the synthetic aperture radar image in an environment-adaptive manner based on the amount of signal distortion predicted by the image correction unit; A step of detecting leak candidate regions based on a corrected image in a leak detection unit; and A method comprising a step of retraining the prediction step and the detection step using field verification data in a learning management unit.

Description

System and Apparatus for Robust L-Band Leakage Detection in Wet Conditions The present invention relates to a technology field for remotely detecting leaks in underground water and sewage facilities. More specifically, the present invention belongs to a new technology field that is clearly distinguished from existing SAR-based leak detection technology in that it is not merely a leak detection technology using L-band SAR images, but rather an environment-adaptive signal processing technology that quantitatively predicts and corrects non-linear and complex signal distortions occurring in wet environments that have not been resolved by conventional image-based leak detection technology. Early detection is crucial because leaks in underground water and sewage pipes cause urban ground subsidence, water resource loss due to water leakage, and hygiene and environmental problems due to sewage leakage. Recently, L-band synthetic aperture radar-based pipeline leak detection technology, which utilizes satellite-based remote sensing technology to identify leaks in extensive areas, is being widely used. Conventional SAR-based leak detection technology has not solved the problem that the scattering intensity of the SAR signal changes in a pattern similar to leaks when soil moisture increases or surface wetness rises after rainfall, and most of them have structural limitations in which false positives increase rapidly in wet environments by performing detection based only on simple speckle removal filters, static threshold adjustments, or signal characteristics based on dry conditions. More specifically, it is significantly affected by changes in rainfall, humidity, and soil moisture conditions. For instance, if surface and soil moisture rise abnormally after rainfall, SAR signal scattering increases even in the absence of leakage, leading to the problem of misclassification as an 'abnormal moisture zone'. There is a problem in that the amount of signal distortion caused by soil and topographic conditions is not reflected. Generally, clay areas, sandy soil areas, and poorly drained areas have very different basic scattering characteristics, so the degree of SAR signal distortion varies significantly even under the same rainfall conditions; however, conventional technology has not been able to quantitatively reflect this. There is a lack of correction models that respond to non-linear environmental variations. Existing correction technologies utilize simple descending filtering and speckle removal filters, making it difficult to accurately predict and correct distortions in SAR signals when multiple complex environmental parameters act simultaneously. Therefore, there is a need for a technology that quantitatively reflects rainfall, soil, and humidity conditions in a wet environment to predict in advance how much the L-band SAR signal will be distorted, and adaptively corrects SAR images based on this to enable leak detection without false positives. FIG. 1 is an overall configuration diagram of an L-band-based water and sewage leak detection device that is robust in a wet environment according to an embodiment of the present invention. Figure 2 is a device configuration diagram of the SAR image receiver shown in Figure 1. Figure 3 is a device configuration diagram of the signal distortion prediction unit illustrated in Figure 1. Figure 4 is a device configuration diagram of the image correction unit illustrated in Figure 1. Figure 5 is a device configuration diagram of the leak detection unit shown in Figure 1. FIG. 6 is a flowchart illustrating a robust L-band-based water and sewage leak detection method in a wet environment according to an embodiment of the present invention. Various embodiments of the present disclosure are described below in conjunction with the accompanying drawings. As various embodiments of the present disclosure may be subject to various modifications and may have various forms, specific embodiments are illustrated in the drawings and described in detail. However, this is not intended to limit the various embodiments of the present disclosure to specific forms, and it should be understood that they include all modifications and/or equivalents and substitutions that fall within the spirit and scope of the various embodiments of the present disclosure. Similar reference numerals have been used for similar components in connection with the description of the drawings. In various embodiments of the present disclosure, terms such as “comprising” or “having” are intended to specify the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. In various embodiments of the present disclosure, expressions such as “or” include any and all combinations of the words listed