A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
DRAIN WATER ABNORMALITY DETECTION DEVICE, LEARNED MODEL GENERATION DEVICE, DRAIN WATER ABNORMALITY DETECTION METHOD AND LEARNED MODEL GENERATION METHOD
To prevent drain water that is colored by a non-abnormality factor color being the color that may be erroneously recognized as an abnormality by being mixed to drain water from being erroneously detected as an abnormality and improve detection accuracy of an abnormality of drain water.SOLUTION: A drain water abnormality detection device for detecting an abnormality of drain water from a facility that appears as a hue comprises: an image acquisition unit which acquires a drain water image being a photographed image of the drain water; and a detection unit which detects an abnormality of drain water from the facility by using a learned model generated with learning using as teacher data association information that associates a plurality of simulated drain water images being the photographed images of a plurality of simulated drain waters which is obtained by mixing a component of an abnormality factor color showing an abnormality of drain water and a component of a non-abnormality factor color being the color that may be erroneously recognized as an abnormality by being mixed to drain water and in which mutually-different hues appear with an evaluation of a degree of an abnormality of each simulated drain water image with the drain water image as input and the evaluation of the degree of the abnormality as output, and outputs the result.SELECTED DRAWING: Figure 1
【課題】排水に混入することで異常と誤認識され得る色である非異常要因色によって着色された排水を異常と過検知することを抑制し、排水の異常の検知精度を向上させる。【解決手段】色彩として現れる、施設からの排水の異常を検知する排水異常検知装置は、排水の撮像画像である排水画像を取得する画像取得部と、排水の異常を示す異常要因色の成分と、排水に混入することで異常と誤認識され得る色である非異常要因色の成分とを混合して得られる複数の模擬排水であって、それぞれ異なる色彩が現れている複数の模擬排水の撮像画像である複数の模擬排水画像と、各前記模擬排水画像の異常の程度の評価とを対応付ける対応情報を教師データとして用いた学習により生成され、排水画像を入力とし、異常の程度の評価を出力とする学習済みモデルを用いて、前記施設からの排水の異常を検知し、その結果を出力する検知部と、を備える。【選択図】図１
DRAIN WATER ABNORMALITY DETECTION DEVICE, LEARNED MODEL GENERATION DEVICE, DRAIN WATER ABNORMALITY DETECTION METHOD AND LEARNED MODEL GENERATION METHOD
To prevent drain water that is colored by a non-abnormality factor color being the color that may be erroneously recognized as an abnormality by being mixed to drain water from being erroneously detected as an abnormality and improve detection accuracy of an abnormality of drain water.SOLUTION: A drain water abnormality detection device for detecting an abnormality of drain water from a facility that appears as a hue comprises: an image acquisition unit which acquires a drain water image being a photographed image of the drain water; and a detection unit which detects an abnormality of drain water from the facility by using a learned model generated with learning using as teacher data association information that associates a plurality of simulated drain water images being the photographed images of a plurality of simulated drain waters which is obtained by mixing a component of an abnormality factor color showing an abnormality of drain water and a component of a non-abnormality factor color being the color that may be erroneously recognized as an abnormality by being mixed to drain water and in which mutually-different hues appear with an evaluation of a degree of an abnormality of each simulated drain water image with the drain water image as input and the evaluation of the degree of the abnormality as output, and outputs the result.SELECTED DRAWING: Figure 1
【課題】排水に混入することで異常と誤認識され得る色である非異常要因色によって着色された排水を異常と過検知することを抑制し、排水の異常の検知精度を向上させる。【解決手段】色彩として現れる、施設からの排水の異常を検知する排水異常検知装置は、排水の撮像画像である排水画像を取得する画像取得部と、排水の異常を示す異常要因色の成分と、排水に混入することで異常と誤認識され得る色である非異常要因色の成分とを混合して得られる複数の模擬排水であって、それぞれ異なる色彩が現れている複数の模擬排水の撮像画像である複数の模擬排水画像と、各前記模擬排水画像の異常の程度の評価とを対応付ける対応情報を教師データとして用いた学習により生成され、排水画像を入力とし、異常の程度の評価を出力とする学習済みモデルを用いて、前記施設からの排水の異常を検知し、その結果を出力する検知部と、を備える。【選択図】図１
DRAIN WATER ABNORMALITY DETECTION DEVICE, LEARNED MODEL GENERATION DEVICE, DRAIN WATER ABNORMALITY DETECTION METHOD AND LEARNED MODEL GENERATION METHOD
排水異常検知装置、学習済みモデル生成装置、排水異常検知方法、及び学習済みモデル生成方法
OGAWA KAORI (author)
2023-03-29
Patent
Electronic Resource
Japanese
IPC:
G01N
Untersuchen oder Analysieren von Stoffen durch Bestimmen ihrer chemischen oder physikalischen Eigenschaften
,
INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
/
E03F
SEWERS
,
Abwasserkanäle
/
G06T
Bilddatenverarbeitung oder Bilddatenerzeugung allgemein
,
IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
/
H04N
PICTORIAL COMMUNICATION, e.g. TELEVISION
,
Bildübertragung, z.B. Fernsehen
ABNORMALITY DETECTION DEVICE, WATER GATE SYSTEM, AND ABNORMALITY DETECTION METHOD
| European Patent Office | 2021
ABNORMALITY DETECTION DEVICE, WATER GATE SYSTEM, AND ABNORMALITY DETECTION METHOD
| European Patent Office | 2022
ABNORMALITY DETECTION DEVICE, AND ABNORMALITY DETECTION METHOD
| European Patent Office | 2022
WATER-LEVEL ABNORMALITY DETECTION PROGRAM AND WATER-LEVEL ABNORMALITY DETECTION DEVICE
| European Patent Office | 2022
ABNORMALITY DETECTION DEVICE, SLUICE SYSTEM, AND ABNORMALITY DETECTION METHOD
| European Patent Office | 2022