KR20210067797A - Ai learning system for finding dangerous road obstacle - Google Patents

Ai learning system for finding dangerous road obstacle Download PDF

Info

Publication number
KR20210067797A
KR20210067797A KR1020190157772A KR20190157772A KR20210067797A KR 20210067797 A KR20210067797 A KR 20210067797A KR 1020190157772 A KR1020190157772 A KR 1020190157772A KR 20190157772 A KR20190157772 A KR 20190157772A KR 20210067797 A KR20210067797 A KR 20210067797A
Authority
KR
South Korea
Prior art keywords
image
unit
network
transmission target
processing unit
Prior art date
Application number
KR1020190157772A
Other languages
Korean (ko)
Inventor
나상민
Original Assignee
주식회사 디투리소스
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 주식회사 디투리소스 filed Critical 주식회사 디투리소스
Priority to KR1020190157772A priority Critical patent/KR20210067797A/en
Publication of KR20210067797A publication Critical patent/KR20210067797A/en

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0967—Systems involving transmission of highway information, e.g. weather, speed limits
    • G08G1/096766—Systems involving transmission of highway information, e.g. weather, speed limits where the system is characterised by the origin of the information transmission
    • G08G1/096783—Systems involving transmission of highway information, e.g. weather, speed limits where the system is characterised by the origin of the information transmission where the origin of the information is a roadside individual element
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0967—Systems involving transmission of highway information, e.g. weather, speed limits
    • G08G1/096733—Systems involving transmission of highway information, e.g. weather, speed limits where a selection of the information might take place
    • G08G1/09675—Systems involving transmission of highway information, e.g. weather, speed limits where a selection of the information might take place where a selection from the received information takes place in the vehicle
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0968—Systems involving transmission of navigation instructions to the vehicle
    • G08G1/096855—Systems involving transmission of navigation instructions to the vehicle where the output is provided in a suitable form to the driver
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0968—Systems involving transmission of navigation instructions to the vehicle
    • G08G1/096877—Systems involving transmission of navigation instructions to the vehicle where the input to the navigation device is provided by a suitable I/O arrangement
    • G08G1/096883—Systems involving transmission of navigation instructions to the vehicle where the input to the navigation device is provided by a suitable I/O arrangement where input information is obtained using a mobile device, e.g. a mobile phone, a PDA
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/09—Arrangements for giving variable traffic instructions
    • G08G1/0962—Arrangements for giving variable traffic instructions having an indicator mounted inside the vehicle, e.g. giving voice messages
    • G08G1/0968—Systems involving transmission of navigation instructions to the vehicle
    • G08G1/096877—Systems involving transmission of navigation instructions to the vehicle where the input to the navigation device is provided by a suitable I/O arrangement
    • G08G1/096888—Systems involving transmission of navigation instructions to the vehicle where the input to the navigation device is provided by a suitable I/O arrangement where input information is obtained using learning systems, e.g. history databases

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Remote Sensing (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Atmospheric Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Image Analysis (AREA)
  • Traffic Control Systems (AREA)

Abstract

The present invention relates to a weight-based artificial intelligence learning system for detecting a dangerous object on a road surface which improves the performance of a learning model by assigning weights. The weight-based artificial intelligence learning system for detecting a dangerous object on a road surface comprises: an image acquisition unit (110) using a camera, GPS, a gyro sensor, etc. embedded in an ordinary smartphone (100) to provide a driver with a black box function for a vehicle and acquire a road surface image; an intelligent network selection unit (131) including an image preprocessing unit (120) detecting a real-time dangerous factor in the image to preprocess the real-time dangerous factor into a priority transmission target group and a general transmission target group and an image transmission unit (130) transmitting an image, wherein the image transmission unit (130) recognizes a crossroad and a temporary standby state to intelligently select a public network (free public network), a private Wi-Fi network (generally free), and a mobile network (generally a paid network) of a user; a priority transmission unit (132) transmitting a target processed by the image preprocessing unit and classified as a priority transmission target group first; a general transmission unit (133) managing and transmitting a large quantity of generation transmission target groups in a situation in which targets processed by the image preprocessing unit and classified as general transmission target groups are intelligently recognized as an office or home; and an integrated control center (500) consisting of a system displaying all the processed information to easily identify a variety of information in accordance with authorities such as a highest manager, a city manager, a manager, and a repair manager.

Description

도로 노면 위험물 탐지를 위한 가중치 기반 인공지능 학습 시스템{AI LEARNING SYSTEM FOR FINDING DANGEROUS ROAD OBSTACLE}AI LEARNING SYSTEM FOR FINDING DANGEROUS ROAD OBSTACLE

본 발명은 도로 표면의 위험물을 검출 및 관리하는 시스템의 인공지능 학습 시스템에 관한 것으로, 더욱 상세하게는 스마트폰을 차량에 거치하고 모바일 앱을 실행하여, 차량의 운전자는 차량용 블랙박스 기능으로 사용하고, 시스템 운용사는 도로의 위험물을 검출하고, 그러한 정보를 시각화하여, 지도에 매핑하고, 실시간 정보 공유을 함으로서 도로의 안전을 확보하는데 있어, 전세계의 매우 다양한 형태의 도로 노면 위험물의 형태, 속성 등에 대하여, 마킹하고 인공지능의 판단에 대하여, 인간이 관여하여, 가중치를 부여하여, 학습 모델에 성능 향상을 이뤄내는 시스템에 관한 것이다.The present invention relates to an artificial intelligence learning system of a system for detecting and managing dangerous substances on the road surface, and more particularly, by mounting a smartphone on a vehicle and running a mobile app, the driver of the vehicle uses it as a vehicle black box function, , the system operator detects dangerous substances on the road, visualizes such information, maps it on a map, and secures road safety by sharing real-time information. It relates to a system that improves the performance of the learning model by adding weights to the marking and AI judgment.

도로 표면상의 포트홀, 도색 불량, 중앙 안전봉 불량 등은 운행 중인 자동차의 안전운행에 영향을 주고, 이로 인한 인적, 재산적 피해를 야기한다. 또한, 그 관리에 있어서도 체계적, 분석적 관리가 아닌 육안 검사에 의한 관리가 이뤄지고 있는 것이 현실이다.Potholes on the surface of the road, poor painting, and defective central safety rods affect the safe operation of a vehicle in operation, causing human and property damage. In addition, the reality is that management is conducted by visual inspection rather than systematic and analytical management.

경제가 발전함에 따라, 도로망은 계속 확충되고 있다. 이로 인해, 사람에 의한 육안 검사 식 및 파일럿 차량(탐지 차량)을 통한 탐지에는 물리적 한계가 존재한다.As the economy develops, the road network continues to expand. For this reason, there is a physical limit to the detection by the human eye and the pilot vehicle (detection vehicle).

특히, 지방도로 갈수록 그 현상이 심각하며, 신고하지 않으면 장기간 위험한 상황이 방치되어 있고, 그로 인해 추후 보수 시, 추가적 많은 비용 야기와 운전자의 안전에 심각한 위협을 발생하고 있다.In particular, the phenomenon becomes more serious as the road progresses, and if not reported, a dangerous situation is left unattended for a long period of time, resulting in additional costly additional costs and a serious threat to driver safety.

본 기술은 경제성장국에게도 도움이 되겠지만, 특히, 동남아시아, 인도, 인도네시아, 아프리카 등 신흥개발도상국 및 저개발국가 등에 저렴한 비용으로 도로를 체계적으로 관리할 수 있는 방안이 될 수 있다.This technology will be helpful to economically growing countries, but in particular, it can be a way to systematically manage roads at low cost in emerging and underdeveloped countries such as Southeast Asia, India, Indonesia, and Africa.

등록특허공보 제10-1715211호 (공고일:2017.03.06.)Registered Patent Publication No. 10-1715211 (Announcement Date: 2017.03.06.) 등록특허공보 제10-1763915호 (공고일:2017.08.14.)Registered Patent Publication No. 10-1763915 (Announcement Date: 2017.08.14.) 공개특허공보 제10-2019-0053151호 (공개일: 2019.05.71.)Laid-open Patent Publication No. 10-2019-0053151 (published date: 2019.05.71.)

본 발명은 도로 표면의 위험물을 검출 및 관리하는 시스템의 인공지능 학습 시스템에 관한 것으로, 더욱 상세하게는 스마트폰을 차량에 거치하고 모바일 앱을 실행하여, 차량의 운전자는 차량용 블랙박스 기능으로 사용하고, 시스템 운용사는 도로의 위험물을 검출하고, 그러한 정보를 시각화하여, 지도에 매핑하고, 실시간 정보 공유을 함으로서 도로의 안전을 확보하는데 있어, 전세계의 매우 다양한 형태의 도로 노면 위험물의 형태, 속성 등에 대하여, 마킹하고 인공지능의 판단에 대하여, 인간이 관여하여, 가중치를 부여하여, 학습 모델에 성능 향상을 이뤄내는 시스템에 관한 것이다.The present invention relates to an artificial intelligence learning system of a system for detecting and managing dangerous substances on the road surface, and more particularly, by mounting a smartphone on a vehicle and running a mobile app, the driver of the vehicle uses it as a vehicle black box function, , the system operator detects dangerous substances on the road, visualizes such information, maps it on a map, and secures road safety by sharing real-time information. It relates to a system that improves the performance of the learning model by adding weights to the marking and AI judgment.

일반적으로 인공지능의 학습 시스템 중, 딥러닝에서는 인간이 관여하지 않고, 빅데이터(매우 많은 데이터)에 기반한 학습을 수행하는데, 빅데이터를 구축하는데에는 매우 많은 자원(자금, 전산센터, 스토리지, 데이터 확보를 위한 투입)이 필요하게 됨으로서, 빅데이터를 만들어가는 중이거나, 적게 만들어진 상태에서는 정확도에 이슈가 있을 수 있다.In general, among the learning systems of artificial intelligence, deep learning does not involve humans and performs learning based on big data (a lot of data), but there are a lot of resources (funds, computer centers, storage, data) to build big data. input for securing) is required, so there may be an issue with accuracy in the process of making big data or in the state where it is small.

상기와 같은 이슈를 해결하기 위해, 딥러닝으로 수행하고, 도로 노면 위험물 탐지 결과에 대하여, 담당자의 선택에 의한 마킹을 하고, 이를 학습/재학습 시, 가중치 반영을 함으로서, 위험물 탐지의 정확도, 성능 향상을 이루기 위한 시스템이다.In order to solve the above issues, deep learning is performed, and the result of detection of dangerous substances on the road is marked according to the selection of the person in charge, and when learning/re-learning, the weight is reflected, so that the accuracy and performance of detection of dangerous substances It is a system for improvement.

상기한 목적을 달성하기 위한 본 발명은, 일반적인 스마트폰(100)에 내장된 카메라, GPS, 자이로센서 등을 활용하여, 운전자에게 차량용 블랙박스 기능을 제공하고, 도로 노면 영상을 취득하는 영상취득부(110); 상기 영상에서 실시간 위험 요인을 탐지하여, 우선전송대상군과 일반전송대상군으로 전처리하는 영상전처리부(120) 및 영상을 전송하는 영상송신부(130)를 포함하되, 상기 영상송신부(130)는 교차로 및 잠시 대기 중 상태를 인지하여, 퍼블릭 네트워크(공개 무료망)와 사설와이파이망(일반적으로 무료)와 사용자의 모바일 네트워크(일반적으로 유료망)를 지능적으로 선택하는 지능형 네트워크 선택부(131); 영상전처리부에서 처리되어 우선전송대상군으로 분류된 대상을 우선전송하는 우선전송부(132); 영상전처리부에서 처리되어 일반전송대상군으로 분류된 대상을 집 또는 사무실로 지능적으로 인지된 상황에서 대량의 일반전송대상군을 관리하며 전송하는 일반전송부(133)를 포함한다. The present invention for achieving the above object provides a vehicle black box function to the driver by utilizing a camera, GPS, gyro sensor, etc. built in a general smart phone 100, and an image acquisition unit for acquiring a road image (110); and an image pre-processing unit 120 that detects real-time risk factors in the image and pre-processes them into a priority transmission target group and a general transmission target group, and an image transmission unit 130 for transmitting the image, wherein the image transmission unit 130 includes an intersection and An intelligent network selection unit 131 for intelligently selecting a public network (public free network), a private Wi-Fi network (generally free), and the user's mobile network (generally a paid network) by recognizing the waiting state for a while; a priority transmission unit 132 for preferentially transmitting an object that is processed by the image preprocessor and classified as a priority transmission target group; It includes a general transmission unit 133 that manages and transmits a large number of general transmission target groups in a situation where the object classified as the general transmission target group processed by the image pre-processing unit is intelligently recognized as a home or office.

상기 영상정보처리부(400)의 인공지능 데이터 분석부(410)는 처리된 영상에서 인공지능에 의한 분석을 통해, 도로의 위험요인를 프로파일링하고, 정보화한다. 상기 영상정보처리부(400)의 데이터베이스 처리부(420)는 분석된 정보를 체계적 데이터베이스화한다.The artificial intelligence data analysis unit 410 of the image information processing unit 400 profiles and informatizes road risk factors through analysis by artificial intelligence in the processed image. The database processing unit 420 of the image information processing unit 400 systematically converts the analyzed information into a database.

통합관제센터(500)는 이상의 모든 처리된 정보를 최고책임자, 도시 책임자, 관리책임자, 보수 책임자 등 권한에 따른 다양한 정보를 쉽게 식별할 수 있도록 표현하는 시스템으로 구성된다.The integrated control center 500 is composed of a system that expresses all the processed information above so that various information according to authority such as the chief officer, city manager, management officer, maintenance officer, etc. can be easily identified.

통합관제센터(500) 내에 구비된 인간협업 가중치 처리부(501)에서는, 통합관제센터에서 모니터링하는 담당자가 모니터링 결과물을 확인 시, 필요에 따라 가중치(예시: 1 ~ 5 의 점수 지표)를 부여 하고, 그에 대한 각 위험 요인 탐지 이미지에 대하여, 담당자가 점수 지표를 마킹 및 처리하게 된다. In the human collaboration weight processing unit 501 provided in the integrated control center 500, when the person in charge of monitoring in the integrated control center checks the monitoring result, a weight (eg, score index of 1 to 5) is given as necessary, For each risk factor detection image therefor, the person in charge marks and processes the score index.

이상의 인간협업 가중치 처리부(501)에서 지표화된 것을, AI 학습 모델 관리 DB 처리부(502)에서는 학습 또는 재학습을 수행하며, 담당자가 부여한 지표의 가중치를 추가로 부여하여, 정확한 판단에 반영하고, 이를 통한 이후의 위험물 탐지 시, 그와 연관된 분석 및 판단에 있어, 어느정도의 정확도에 영향을 주었는지, DB 화하여, 학습 모델의 성능을 향상시키게 된다. What has been indexed by the human collaboration weight processing unit 501 above, the AI learning model management DB processing unit 502 performs learning or re-learning, and additionally gives the weight of the index given by the person in charge, reflects it in accurate judgment, and Upon subsequent detection of dangerous substances, the performance of the learning model is improved by converting it into a DB, which affects the degree of accuracy in the analysis and judgment related thereto.

본 발명은 도로 표면의 위험물을 검출 및 관리하는 시스템의 인공지능 학습 시스템에 관한 것으로, 더욱 상세하게는 스마트폰을 차량에 거치하고 모바일 앱을 실행하여, 차량의 운전자는 차량용 블랙박스 기능으로 사용하고, 시스템 운용사는 도로의 위험물을 검출하고, 그러한 정보를 시각화하여, 지도에 매핑하고, 실시간 정보 공유을 함으로서 도로의 안전을 확보하는데 있어, 전세계의 매우 다양한 형태의 도로 노면 위험물의 형태, 속성 등에 대하여, 마킹하고 인공지능의 판단에 대하여, 인간이 관여하여, 가중치를 부여하여, 학습 모델에 성능 향상을 이뤄내는 시스템에 관한 것이다.The present invention relates to an artificial intelligence learning system of a system for detecting and managing dangerous substances on the road surface, and more particularly, by mounting a smartphone on a vehicle and running a mobile app, the driver of the vehicle uses it as a vehicle black box function, , the system operator detects dangerous substances on the road, visualizes such information, maps it on a map, and secures road safety by sharing real-time information. It relates to a system that improves the performance of the learning model by adding weights to the marking and AI judgment.

도 1은 본 발명의 전체적인 구성을 도시하는 블록도이다.
도 2 및 도 3은 모니터링 담당자가 모니터 결과물을 확인하고, 점수 지표를 마킹 및 처리하는 화면이다.
도 4는 위험물 검출 결과를 맵에 표시한 화면이다. 1 is a block diagram showing the overall configuration of the present invention.
2 and 3 are screens in which the monitoring person checks the monitor result, and marks and processes the score index.
4 is a screen on which a detection result of a dangerous substance is displayed on a map.

도 1을 참조하여, 본 발명의 전체적인 구성을 살펴보면, 일반적인 스마트폰(100)에 내장된 카메라, GPS, 자이로센서 등을 활용하여, 운전자에게 차량용 블랙박스 기능을 제공하고, 도로 노면 영상을 취득하는 영상취득부(110); 상기 영상에서 실시간 위험 요인을 탐지하여, 우선전송대상군과 일반전송대상군으로 전처리하는 영상전처리부(120) 및 영상을 전송하는 영상송신부(130)를 포함하되, 상기 영상송신부(130)는 교차로 및 잠시 대기 중 상태를 인지하여, 퍼블릭 네트워크(공개 무료망)와 사설와이파이망(일반적으로 무료)와 사용자의 모바일 네트워크(일반적으로 유료망)를 지능적으로 선택하는 지능형 네트워크 선택부(131); 영상전처리부에서 처리되어 우선전송대상군으로 분류된 대상을 우선전송하는 우선전송부(132); 영상전처리부에서 처리되어 일반전송대상군으로 분류된 대상을 집 또는 사무실로 지능적으로 인지된 상황에서 대량의 일반전송대상군을 관리하며 전송하는 일반전송부(133)를 포함한다. Referring to FIG. 1, looking at the overall configuration of the present invention, by utilizing a camera, GPS, gyro sensor, etc. built in a general smart phone 100, a vehicle black box function is provided to the driver, and a road image is obtained. image acquisition unit 110; and an image pre-processing unit 120 that detects real-time risk factors in the image and pre-processes them into a priority transmission target group and a general transmission target group, and an image transmission unit 130 for transmitting the image, wherein the image transmission unit 130 includes an intersection and An intelligent network selection unit 131 for intelligently selecting a public network (public free network), a private Wi-Fi network (generally free), and the user's mobile network (generally a paid network) by recognizing the waiting state for a while; a priority transmission unit 132 for preferentially transmitting an object that is processed by the image preprocessor and classified as a priority transmission target group; It includes a general transmission unit 133 that manages and transmits a large number of general transmission target groups in a situation where the object classified as the general transmission target group processed by the image pre-processing unit is intelligently recognized as a home or office.

상기 영상정보처리부(400)의 인공지능 데이터 분석부(410)는 처리된 영상에서 인공지능에 의한 분석을 통해, 도로의 위험요인를 프로파일링하고, 정보화한다. 상기 영상정보처리부(400)의 데이터베이스 처리부(420)는 분석된 정보를 체계적 데이터베이스화한다.The artificial intelligence data analysis unit 410 of the image information processing unit 400 profiles and informatizes road risk factors through analysis by artificial intelligence in the processed image. The database processing unit 420 of the image information processing unit 400 systematically converts the analyzed information into a database.

통합관제센터(500)는 이상의 모든 처리된 정보를 최고책임자, 도시 책임자, 관리책임자, 보수 책임자 등 권한에 따른 다양한 정보를 쉽게 식별할 수 있도록 표현하는 시스템으로 구성된다.The integrated control center 500 is composed of a system that expresses all the processed information above so that various information according to authority such as the chief officer, city manager, management officer, maintenance officer, etc. can be easily identified.

통합관제센터(500) 내에 구비된 인간협업 가중치 처리부(501)에서는, 통합관제센터에서 모니터링하는 담당자가 모니터링 결과물을 확인 시, 도 2 및 도 3에 도시된 것과 같은 화면에서 필요에 따라 가중치(예시: 1 ~ 5 의 점수 지표)를 부여 하고, 그에 대한 각 위험 요인 탐지 이미지에 대하여, 담당자가 점수 지표를 마킹 및 처리하게 된다. In the human collaboration weight processing unit 501 provided in the integrated control center 500, when the person in charge of monitoring in the integrated control center checks the monitoring result, the weights (examples) : Score index of 1 to 5), and for each risk factor detection image, the person in charge marks and processes the score index.

이상의 인간협업 가중치 처리부(501)에서 지표화된 것을, AI 학습 모델 관리 DB 처리부(502)에서는 학습 또는 재학습을 수행하며, 담당자가 부여한 지표의 가중치를 추가로 부여하여, 정확한 판단에 반영하고, 이를 통한 이후의 위험물 탐지 시, 그와 연관된 분석 및 판단에 있어, 어느정도의 정확도에 영향을 주었는지, DB 화하여, 학습 모델의 성능을 향상시키게 된다. What has been indexed by the human collaboration weight processing unit 501 above, the AI learning model management DB processing unit 502 performs learning or re-learning, and additionally gives the weight of the index given by the person in charge, reflects it in accurate judgment, and Upon subsequent detection of dangerous substances, the performance of the learning model is improved by converting it into a DB, which affects the degree of accuracy in the analysis and judgment related thereto.

Claims (4)

일반적인 스마트폰(100)에 내장된 카메라, GPS, 자이로센서 등을 활용하여, 운전자에게 차량용 블랙박스 기능을 제공하고, 도로 노면 영상을 취득하는 영상취득부(110);
상기 영상에서 실시간 위험 요인을 탐지하여, 우선전송대상군과 일반전송대상군으로 전처리하는 영상전처리부(120) 및 영상을 전송하는 영상송신부(130)를 포함하되, 상기 영상송신부(130)는 교차로 및 잠시 대기 중 상태를 인지하여, 퍼블릭 네트워크(공개 무료망)와 사설와이파이망(일반적으로 무료)와 사용자의 모바일 네트워크(일반적으로 유료망)를 지능적으로 선택하는 지능형 네트워크 선택부(131);
영상전처리부에서 처리되어 우선전송대상군으로 분류된 대상을 우선전송하는 우선전송부(132); 영상전처리부에서 처리되어 일반전송대상군으로 분류된 대상을 집 또는 사무실로 지능적으로 인지된 상황에서 대량의 일반전송대상군을 관리하며 전송하는 일반전송부(133); 및
이상의 모든 처리된 정보를 최고책임자, 도시 책임자, 관리책임자, 보수 책임자 등 권한에 따른 다양한 정보를 쉽게 식별할 수 있도록 표현하는 시스템으로 구성되는 통합관제센터(500)를 포함하는 것을 특징으로 하는, 도로 노면 위험물 탐지를 위한 가중치 기반 인공지능 학습 시스템.
an image acquisition unit 110 that provides a vehicle black box function to the driver and acquires a road surface image by utilizing a camera, GPS, gyro sensor, etc. built in a general smart phone 100;
and an image preprocessing unit 120 that detects real-time risk factors in the image and pre-processes them into a priority transmission target group and a general transmission target group, and an image transmission unit 130 for transmitting the image, wherein the image transmission unit 130 includes an intersection and an intelligent network selection unit 131 for intelligently selecting a public network (public free network), a private Wi-Fi network (generally free), and a user's mobile network (generally a paid network) by recognizing the waiting state for a while;
a priority transmission unit 132 for preferentially transmitting an object that is processed by the image preprocessor and classified as a priority transmission target group; a general transmission unit 133 that manages and transmits a large number of general transmission target groups in a situation in which objects classified as general transmission target groups processed by the image pre-processing unit are intelligently recognized as home or office; and
Road characterized in that it includes an integrated control center 500 consisting of a system that expresses all the processed information above so that various information according to authority such as chief officer, city manager, management officer, maintenance officer, etc. can be easily identified. Weight-based artificial intelligence learning system for road surface hazardous material detection.
 청구항 1에 있어서, 상기 영상정보처리부(400)의 인공지능 데이터 분석부(410)는 처리된 영상에서 인공지능에 의한 분석을 통해, 도로의 위험요인를 프로파일링하고, 정보하며,
상기 영상정보처리부(400)의 데이터베이스 처리부(420)는 분석된 정보를 체계적 데이터베이스화하는 것을 특징으로 하는, 도로 노면 위험물 탐지를 위한 가중치 기반 인공지능 학습 시스템.
The method according to claim 1, wherein the artificial intelligence data analysis unit (410) of the image information processing unit (400) through the analysis by artificial intelligence in the processed image, profiling and information on risk factors of the road,
The database processing unit 420 of the image information processing unit 400 systematically converts the analyzed information into a database, a weight-based artificial intelligence learning system for detecting dangerous substances on the road surface.
 청구항 1 또는 청구항 2에 있어서, 상기 통합관제센터(500) 내에 구비된 인간협업 가중치 처리부(501)에서는, 통합관제센터에서 모니터링하는 담당자가 모니터링 결과물을 확인 시, 필요에 따라 가중치(예시: 1 ~ 5 의 점수 지표)를 부여 하고, 그에 대한 각 위험 요인 탐지 이미지에 대하여, 담당자가 점수 지표를 마킹 및 처리하게 되는 것을 특징으로 하는, 도로 노면 위험물 탐지를 위한 가중치 기반 인공지능 학습 시스템.
The method according to claim 1 or 2, wherein in the human collaboration weight processing unit (501) provided in the integrated control center (500), when the person in charge of monitoring in the integrated control center checks the monitoring result, weights (eg: 1 ~ 5), and for each risk factor detection image, the person in charge marks and processes the score index, a weight-based artificial intelligence learning system for detecting dangerous substances on the road surface.
 청구항 3에 있어서, 상기 이상의 인간협업 가중치 처리부(501)에서 지표화된 것을, AI 학습 모델 관리 DB 처리부(502)에서는 학습 또는 재학습을 수행하며, 담당자가 부여한 지표의 가중치를 추가로 부여하여, 정확한 판단에 반영하고, 이를 통한 이후의 위험물 탐지 시, 그와 연관된 분석 및 판단에 있어, 어느정도의 정확도에 영향을 주었는지, DB 화하여, 학습 모델의 성능을 향상시키게 된다. The method according to claim 3, wherein the above-mentioned human collaboration weight processing unit 501 indexes, the AI learning model management DB processing unit 502 performs learning or re-learning, and by additionally giving the weight of the index given by the person in charge, accurate The performance of the learning model is improved by reflecting it in the judgment, and by converting it into a DB, how much accuracy has been affected in the analysis and judgment related to the detection of dangerous substances.
KR1020190157772A 2019-11-29 2019-11-29 Ai learning system for finding dangerous road obstacle KR20210067797A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
KR1020190157772A KR20210067797A (en) 2019-11-29 2019-11-29 Ai learning system for finding dangerous road obstacle

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
KR1020190157772A KR20210067797A (en) 2019-11-29 2019-11-29 Ai learning system for finding dangerous road obstacle

Publications (1)

Publication Number Publication Date
KR20210067797A true KR20210067797A (en) 2021-06-08

Family

ID=76399533

Family Applications (1)

Application Number Title Priority Date Filing Date
KR1020190157772A KR20210067797A (en) 2019-11-29 2019-11-29 Ai learning system for finding dangerous road obstacle

Country Status (1)

Country Link
KR (1) KR20210067797A (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101715211B1 (en) 2016-11-03 2017-03-13 한국건설기술연구원 Apparatus and method for detecting status of surface of road by using image and laser
KR101763915B1 (en) 2015-03-12 2017-08-14 주식회사 에코트루먼트 System For Collecting And Analyzing Big Data By Monitoring Car's And Road's Conditions
KR20190053151A (en) 2019-04-29 2019-05-17 한기호 System and method for road irregularity and damage detection using road surface flatness vibration sensing terminal

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101763915B1 (en) 2015-03-12 2017-08-14 주식회사 에코트루먼트 System For Collecting And Analyzing Big Data By Monitoring Car's And Road's Conditions
KR101715211B1 (en) 2016-11-03 2017-03-13 한국건설기술연구원 Apparatus and method for detecting status of surface of road by using image and laser
KR20190053151A (en) 2019-04-29 2019-05-17 한기호 System and method for road irregularity and damage detection using road surface flatness vibration sensing terminal

Similar Documents

Publication Publication Date Title
CN102724482B (en) Based on the intelligent vision sensing network moving target relay tracking system of GPS and GIS
CN106296814B (en) Highway maintenance detection and virtual interactive interface method and system
CN111339905B (en) CIM well lid state visual detection system based on deep learning and multiple visual angles
CN113935441B (en) Method and device for selecting automobile automatic driving function test road and storage medium
CN114818056A (en) Traffic data integration method, device, equipment and medium based on BIM technology
WO2024016877A1 (en) Roadside sensing simulation system for vehicle-road collaboration
CN113935442B (en) Method and device for classifying automatic driving function test roads of automobile and storage medium
Kung et al. Application of deep learning and unmanned aerial vehicle on building maintenance
CN110987815A (en) Air-space-ground integrated water environment monitoring and early warning system
CN105262983A (en) Road monitoring system and method based on internet of lamps
CN116091477A (en) Deep learning multi-mode image fusion-based building exterior wall defect detection method
CN113936465A (en) Traffic incident detection method and device
CN114372311B (en) BIM-based design detection method and device, electronic equipment and storage medium
CN114495421B (en) Intelligent open type road construction operation monitoring and early warning method and system
Syed et al. Pothole detection under diverse conditions using object detection models
CN117252316A (en) Emergency evacuation path planning method, device, equipment and storage medium
Kandoi et al. Pothole detection using accelerometer and computer vision with automated complaint redressal
KR20210067797A (en) Ai learning system for finding dangerous road obstacle
US11829959B1 (en) System and methods for fully autonomous potholes detection and road repair determination
Chang et al. Safety risk assessment of electric power operation site based on variable precision rough set
CN111553199A (en) Motor vehicle traffic violation automatic detection technology based on computer vision
KR20210067790A (en) System for Finding Dangerous Road Obstacle and Rewarding
CN111661054B (en) Vehicle control method, device, electronic device and storage medium
JP6752773B2 (en) How to determine the position and / or direction of the sensor
Laureshyn et al. Exploration of a method to validate surrogate safety measures with a focus on vulnerable road users