CN111597992A - Scene object abnormity identification method based on video monitoring - Google Patents
Scene object abnormity identification method based on video monitoring Download PDFInfo
- Publication number
- CN111597992A CN111597992A CN202010413188.8A CN202010413188A CN111597992A CN 111597992 A CN111597992 A CN 111597992A CN 202010413188 A CN202010413188 A CN 202010413188A CN 111597992 A CN111597992 A CN 111597992A
- Authority
- CN
- China
- Prior art keywords
- image
- scene
- video
- feature
- sample
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/48—Matching video sequences
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
A scene object abnormity identification method based on video monitoring is characterized in that each frame of an acquired video is divided into an image, the type and the number of targets appearing in the image are determined, image forming samples are marked, and the images forming sample set of the video is formed as a video, wherein the video is an image sequence project; and performing target extraction on the sample set through a fast-rcnn algorithm, and judging whether the target object in the image has speed abnormality, form abnormality and position abnormality through comparison. According to the method, the problem of inaccurate target extraction caused by shielding can be solved by adopting the fast-rcnn algorithm, and the accuracy of target acquisition is improved so as to improve the accuracy of abnormal behavior detection; the abnormity detection and identification speed is accelerated, and the calculation efficiency is improved.
Description
Technical Field
The invention relates to the field of behavior identification, in particular to a scene object abnormity identification method based on video monitoring.
Background
The charging is carried out in buses, subways and elevators; some exclusive channels, such as motor vehicle lanes, do not allow bicycles to run and pedestrians to trample lawns, and if these are defined as abnormal behaviors, unnecessary loss can be reduced as long as early warning is carried out in advance, so a method is needed for judging whether abnormal behaviors exist.
The panoramic monitoring video screen can capture certain scenes, the foreground of a picture needs to be extracted by the existing detection method, then training sample characteristics are extracted, and a normal sample dictionary library is established. And finally, during testing, calculating the similarity between the test sample and the dictionary library to judge whether the test sample is abnormal. Therefore, the abnormity judgment is divided into two parts, namely foreground extraction and similarity calculation, so that the steps are complicated and the calculation amount is large; in the past, most algorithms divide image frames into blocks in a grid form or extract foreground targets by a split combination-based method, but the effect on shielding or dense multiple targets is poor.
Disclosure of Invention
The invention aims to provide a scene object abnormity identification method based on video monitoring, so that the problems in the prior art are solved.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a scene object abnormity identification method based on video monitoring comprises the following steps:
s1, dividing each frame of the video into continuous images, determining the category and the number of each pixel of the images, and marking the images to form a sample set;
s2, carrying out image segmentation and foreground extraction on the sample set through an rcnn network, training a scene model to obtain the class weight of the scene model, and comprising the following steps:
s201, performing convolution and pooling on the images in the sample set to obtain a feature map in the images;
s202, fusing the feature graph with the shallow feature through upsampling to form feature region division in the image;
s203, the feature map is subjected to convolution kernel of n x 1, and the object type existing in each pixel in the image is judged, wherein n represents the number of the object type in the image;
s3, calculating optical flow of a target object in the images according to the image sequence of the video and the scene model in the step S2 to obtain the speed characteristic of the target object;
s4, acquiring a gray scale image and an edge image of the image, thereby obtaining morphological characteristics of the target object;
s5, obtaining the position characteristics of the target parent through the action track of the target object in the image in the step S2;
s6, forming a real-time sample from the video acquired in real time through the step S1, importing the real-time sample into the scene model trained through the step S2, carrying out image segmentation and type judgment, obtaining the speed feature, the morphological feature and the position feature of the target object in the video acquired in real time through the steps S3-S5, and judging whether the target object in the video is abnormal or not in real time.
Preferably, the samples in the sample set in step S1 are divided into positive samples and negative samples; the positive sample is a sample without a pedestrian, and the negative sample is a sample with a pedestrian.
Preferably, the training of the scene model in step S2 includes two parts, namely, training and testing, that is, the sample set in step S1 is divided into a training set and a testing set, the scene model is trained through the training set, the categories in the image are divided, and then the scene model after the training set is tested through the testing set, so as to obtain a more accurate category detection position.
Preferably, the detection target of the speed feature abnormality is: in the same scene, the speed characteristics of the target object acquired in real time have a larger difference from the conventional speed characteristics.
Preferably, the detection target of the morphological feature abnormality is: and when the target is the same, the form of the target object acquired in real time is different from the conventional form.
Preferably, the detection target of the position feature abnormality is: and in the same scene, acquiring the target object which never appears in the corresponding scene in real time.
The invention has the beneficial effects that: the invention discloses a scene object abnormity identification method based on video monitoring, which adopts a fast-rcnn algorithm to solve the problem of inaccurate target extraction caused by shielding, improves the accuracy of target acquisition and improves the accuracy of abnormal behavior detection; the foreground extraction is carried out in advance through the rcnn network, so that the extracted foreground can be directly applied as comparison when the judgment of morphological abnormality and position abnormality is carried out, and the abnormality detection and identification speed is increased; in addition, the similarity between the normal sample and the morphological abnormality is not required to be calculated when the morphological abnormality is detected, so that the calculation efficiency is improved; the rcnn network is combined with the optical flow, so that objects with abnormal speed can be detected in time, prompt and response can be carried out in time, the occurrence of accidents is further reduced, and unnecessary loss is reduced.
Drawings
FIG. 1 is a block diagram of identification of scene object anomalies.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the invention, are intended for purposes of illustration only and are not intended to limit the scope of the invention.
A scene object abnormity identification method based on video monitoring is characterized in that each frame of an acquired video is divided into an image, the type and the number of targets appearing in the image are determined, image forming samples are marked, and the images forming sample set of the video is formed as a video, wherein the video is an image sequence project; and performing target extraction on the sample set through a fast-rcnn algorithm, and judging whether the target object in the image has speed abnormality, form abnormality and position abnormality through comparison. The above identification block diagram of the object anomaly in the scene is shown in fig. 1, and the identification method of the object anomaly includes the following steps:
s1, dividing the video into images with sequences, confirming the types and the number of objects in each pixel of the images, marking, recording the number of the object types in each image as n, and taking the marked images as samples to form a sample set, wherein the images have positive samples of pedestrians and have no negative samples of pedestrians; dividing the sample set into a training set and a testing set, wherein the training set and the testing set respectively comprise the positive samples and the negative samples;
s2, carrying out image segmentation and foreground extraction on the sample set by utilizing an rcnn network; the method comprises the following steps:
s201, performing convolution and pooling on the images in the sample set to obtain a feature map in the images;
s202, fusing the feature graph with the shallow feature through upsampling to form feature region division in the image;
s203, subjecting the feature map to convolution kernel of n x 1, and judging the object type in each pixel in the image;
by adopting the image segmentation and the foreground extraction, firstly, the training set is utilized to execute the steps S201-S203 to train a scene model, and then the test set is utilized to execute the steps S201-S203 to test the scene model, so as to obtain the more accurate detection position of the object type in the scene;
s3, calculating an optical flow of the target object based on the object types obtained by the segmentation in the step S2 to obtain the velocity of the target object, further obtain the average kinetic energy of the target object, and obtain the velocity characteristic of the target object by combining MHPF characteristics;
s4, obtaining a gray scale map and an edge map of the target object based on the object type of the feature map obtained in the step S2 to obtain the morphological feature of the SIRP target object;
s5, according to the step S2, the position characteristics of all object types in the image can be obtained, and the position characteristics of the object types can be judged by taking the extracted foreground as the background.
S6, forming a real-time sample of the acquired video through the step S1, introducing the real-time sample into the trained scene model in the step S2, performing image segmentation and type judgment, obtaining the speed characteristic, the form characteristic and the position characteristic of the target object in the real-time acquired video through the steps S3-S5, and judging whether the target object in the video has speed abnormality, form abnormality and position abnormality in real time;
the basis for judging the speed abnormality is as follows: in the same scene, judging whether the speed of the moving object acquired in real time is greatly different from the conventional speed or not; the basis for judging the morphological abnormality is as follows: aiming at the same object type, judging whether the object has abnormal morphological characteristics; the judgment basis of the position abnormity is as follows: whether or not there are object classes that never appeared in the same scene.
Examples
In this embodiment, the method for identifying scene object abnormality based on video monitoring is applied to a lawn, 6 types of objects including flowers, grasses, trees, birds, pedestrians and fountains are detected in the lawn, and when image segmentation and foreground extraction are performed through an rcnn network, the extracted feature maps are subjected to convolution and summation of 6 × 1, and distribution positions corresponding to the types of the objects are integrated together; the average speed of various objects moving in the lawn is calculated. When the speed of a moving object in the video acquired in real time is different from the average speed greatly; collecting the appearance of a certain object and the appearance of an object of a corresponding type in the foreground, wherein the object does not appear or rarely appears; when objects which do not belong to the object types in the scene appear in the lawn, the scene objects are judged to be abnormal.
By adopting the technical scheme disclosed by the invention, the following beneficial effects are obtained:
the invention discloses a scene object abnormity identification method based on video monitoring, which adopts a fast-rcnn algorithm to solve the problem of inaccurate target extraction caused by shielding, improves the accuracy of target acquisition and improves the accuracy of abnormal behavior detection; the foreground extraction is carried out in advance through the rcnn network, so that the extracted foreground can be directly applied as comparison when the judgment of morphological abnormality and position abnormality is carried out, and the abnormality detection and identification speed is increased; in addition, the similarity between the normal sample and the morphological abnormality is not required to be calculated when the morphological abnormality is detected, so that the calculation efficiency is improved; the rcnn network is combined with the optical flow, so that objects with abnormal speed can be detected in time, prompt and response can be carried out in time, the occurrence of accidents is further reduced, and unnecessary loss is reduced.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and improvements can be made without departing from the principle of the present invention, and such modifications and improvements should also be considered within the scope of the present invention.
Claims (6)
1. A scene object abnormity identification method based on video monitoring is characterized by comprising the following steps:
s1, dividing each frame of the video into continuous images, determining the category and the number of each pixel of the images, and marking the images to form a sample set;
s2, carrying out image segmentation and foreground extraction on the sample set through an rcnn network, training a scene model to obtain the class weight of the scene model, and comprising the following steps:
s201, performing convolution and pooling on the images in the sample set to obtain a feature map in the images;
s202, fusing the feature graph with the shallow feature through upsampling to form feature region division in the image;
s203, the feature map is subjected to convolution kernel of n x 1, and the object type existing in each pixel in the image is judged, wherein n represents the number of the object type in the image;
s3, calculating optical flow of a target object in the images according to the image sequence of the video and the scene model in the step S2 to obtain the speed characteristic of the target object;
s4, acquiring a gray scale image and an edge image of the image, thereby obtaining morphological characteristics of the target object;
s5, obtaining the position characteristics of the target parent through the action track of the target object in the image in the step S2;
s6, forming a real-time sample from the video acquired in real time through the step S1, importing the real-time sample into the scene model trained through the step S2, carrying out image segmentation and type judgment, obtaining the speed feature, the morphological feature and the position feature of the target object in the video acquired in real time through the steps S3-S5, and judging whether the target object in the video is abnormal or not in real time.
2. The method for identifying scene object abnormality based on video surveillance according to claim 1, wherein the samples in the sample set in step S1 are divided into positive samples and negative samples; the positive sample is a sample without a pedestrian, and the negative sample is a sample with a pedestrian.
3. The method for identifying scene object abnormality based on video surveillance as claimed in claim 1, wherein the training of the scene model in step S2 includes two parts of training and testing, i.e. the sample set in step S1 is divided into a training set and a testing set, the scene model is first trained through the training set, the categories in the image are divided, and then the scene model after the training set is tested through the testing set, so as to obtain a more accurate category detection position.
4. The method for identifying scene object abnormality based on video surveillance according to claim 1, wherein the detection targets of the speed feature abnormality are: in the same scene, the speed characteristics of the target object acquired in real time have a larger difference from the conventional speed characteristics.
5. The method for identifying scene object abnormality based on video surveillance according to claim 1, wherein the detection targets of morphological feature abnormality are: and when the target is the same, the form of the target object acquired in real time is different from the conventional form.
6. The method for identifying scene object abnormality based on video surveillance according to claim 1, wherein the detection targets of the position feature abnormality are: and in the same scene, acquiring the target object which never appears in the corresponding scene in real time.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010413188.8A CN111597992B (en) | 2020-05-15 | 2020-05-15 | Scene object abnormity identification method based on video monitoring |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010413188.8A CN111597992B (en) | 2020-05-15 | 2020-05-15 | Scene object abnormity identification method based on video monitoring |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111597992A true CN111597992A (en) | 2020-08-28 |
CN111597992B CN111597992B (en) | 2023-04-18 |
Family
ID=72191204
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010413188.8A Active CN111597992B (en) | 2020-05-15 | 2020-05-15 | Scene object abnormity identification method based on video monitoring |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111597992B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112861715A (en) * | 2021-02-05 | 2021-05-28 | 邵阳学院 | Image recognition method and image recognition system |
CN114898281A (en) * | 2022-07-13 | 2022-08-12 | 南通高桥体育用品有限公司 | Basketball hoop abnormity detection method and system based on computer vision |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5956424A (en) * | 1996-12-23 | 1999-09-21 | Esco Electronics Corporation | Low false alarm rate detection for a video image processing based security alarm system |
CN105184818A (en) * | 2015-09-06 | 2015-12-23 | 山东华宇航天空间技术有限公司 | Video monitoring abnormal behavior detection method and detections system thereof |
CN107729799A (en) * | 2017-06-13 | 2018-02-23 | 银江股份有限公司 | Crowd's abnormal behaviour vision-based detection and analyzing and alarming system based on depth convolutional neural networks |
CN108229390A (en) * | 2018-01-02 | 2018-06-29 | 济南中维世纪科技有限公司 | Rapid pedestrian detection method based on deep learning |
CN108509889A (en) * | 2018-03-27 | 2018-09-07 | 哈尔滨工业大学深圳研究生院 | A kind of close shot anomaly detection method and device based on skin color segmentation |
CN108932479A (en) * | 2018-06-06 | 2018-12-04 | 上海理工大学 | A kind of human body anomaly detection method |
US10223610B1 (en) * | 2017-10-15 | 2019-03-05 | International Business Machines Corporation | System and method for detection and classification of findings in images |
US20200034667A1 (en) * | 2018-07-27 | 2020-01-30 | Boe Technology Group Co., Ltd. | Image processing method, image processing apparatus, and computer storage medium |
CN110826538A (en) * | 2019-12-06 | 2020-02-21 | 合肥科大智能机器人技术有限公司 | Abnormal off-duty identification system for electric power business hall |
KR102095685B1 (en) * | 2019-12-02 | 2020-04-01 | 주식회사 넥스파시스템 | vehicle detection method and device |
CN111008562A (en) * | 2019-10-31 | 2020-04-14 | 北京城建设计发展集团股份有限公司 | Human-vehicle target detection method with feature map depth fusion |
-
2020
- 2020-05-15 CN CN202010413188.8A patent/CN111597992B/en active Active
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5956424A (en) * | 1996-12-23 | 1999-09-21 | Esco Electronics Corporation | Low false alarm rate detection for a video image processing based security alarm system |
CN105184818A (en) * | 2015-09-06 | 2015-12-23 | 山东华宇航天空间技术有限公司 | Video monitoring abnormal behavior detection method and detections system thereof |
CN107729799A (en) * | 2017-06-13 | 2018-02-23 | 银江股份有限公司 | Crowd's abnormal behaviour vision-based detection and analyzing and alarming system based on depth convolutional neural networks |
US10223610B1 (en) * | 2017-10-15 | 2019-03-05 | International Business Machines Corporation | System and method for detection and classification of findings in images |
CN108229390A (en) * | 2018-01-02 | 2018-06-29 | 济南中维世纪科技有限公司 | Rapid pedestrian detection method based on deep learning |
CN108509889A (en) * | 2018-03-27 | 2018-09-07 | 哈尔滨工业大学深圳研究生院 | A kind of close shot anomaly detection method and device based on skin color segmentation |
CN108932479A (en) * | 2018-06-06 | 2018-12-04 | 上海理工大学 | A kind of human body anomaly detection method |
US20200034667A1 (en) * | 2018-07-27 | 2020-01-30 | Boe Technology Group Co., Ltd. | Image processing method, image processing apparatus, and computer storage medium |
CN111008562A (en) * | 2019-10-31 | 2020-04-14 | 北京城建设计发展集团股份有限公司 | Human-vehicle target detection method with feature map depth fusion |
KR102095685B1 (en) * | 2019-12-02 | 2020-04-01 | 주식회사 넥스파시스템 | vehicle detection method and device |
CN110826538A (en) * | 2019-12-06 | 2020-02-21 | 合肥科大智能机器人技术有限公司 | Abnormal off-duty identification system for electric power business hall |
Non-Patent Citations (4)
Title |
---|
AKSELROD-BALLIN, A等: "Deep Learning for Automatic Detection of Abnormal Findings in Breast Mammography" * |
WANG, GQ等: "Vicinage Effects for a Nitrogen Molecular Cluster in Plasmas" * |
吴鹏: "监控视频人体状态识别方法研究" * |
周自强,等: "一种改进的YOLO目标检测方法在电缆设备异常状态识别中的应用" * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112861715A (en) * | 2021-02-05 | 2021-05-28 | 邵阳学院 | Image recognition method and image recognition system |
CN114898281A (en) * | 2022-07-13 | 2022-08-12 | 南通高桥体育用品有限公司 | Basketball hoop abnormity detection method and system based on computer vision |
Also Published As
Publication number | Publication date |
---|---|
CN111597992B (en) | 2023-04-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP3171292B1 (en) | Driving lane data processing method, device, storage medium and apparatus | |
CN104392468B (en) | Based on the moving target detecting method for improving visual background extraction | |
CN112800860B (en) | High-speed object scattering detection method and system with coordination of event camera and visual camera | |
CN106845364B (en) | Rapid automatic target detection method | |
CN103093198B (en) | A kind of crowd density monitoring method and device | |
CN105404857A (en) | Infrared-based night intelligent vehicle front pedestrian detection method | |
CN110969142B (en) | Abnormal driving scene extraction method based on network-connected vehicle natural driving data | |
CN113505638B (en) | Method and device for monitoring traffic flow and computer readable storage medium | |
CN107590486B (en) | Moving object identification method and system, and bicycle flow statistical method and equipment | |
CN110598613B (en) | Expressway agglomerate fog monitoring method | |
CN111597992B (en) | Scene object abnormity identification method based on video monitoring | |
CN111626170A (en) | Image identification method for railway slope rockfall invasion limit detection | |
CN111797738A (en) | Multi-target traffic behavior fast extraction method based on video identification | |
CN117994987A (en) | Traffic parameter extraction method and related device based on target detection technology | |
CN112735164B (en) | Test data construction method and test method | |
CN109977862A (en) | A kind of recognition methods of parking stall limiter | |
Kini | Real time moving vehicle congestion detection and tracking using OpenCV | |
CN117011341A (en) | Vehicle track detection method and system based on target tracking | |
CN111339824A (en) | Road surface sprinkled object detection method based on machine vision | |
CN113221603A (en) | Method and device for detecting shielding of monitoring equipment by foreign matters | |
CN114724119B (en) | Lane line extraction method, lane line detection device, and storage medium | |
CN110866435A (en) | Far infrared pedestrian training method with self-similarity gradient oriented histogram | |
CN112802348B (en) | Traffic flow counting method based on mixed Gaussian model | |
CN114882469A (en) | Traffic sign detection method and system based on DL-SSD model | |
Kadim et al. | Real-time vehicle counting in complex scene for traffic flow estimation using multi-level convolutional neural network |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |