WO2013143465A1 - 一种视频查询方法、装置与系统 - Google Patents

一种视频查询方法、装置与系统 Download PDF

Info

Publication number
WO2013143465A1
WO2013143465A1 PCT/CN2013/073289 CN2013073289W WO2013143465A1 WO 2013143465 A1 WO2013143465 A1 WO 2013143465A1 CN 2013073289 W CN2013073289 W CN 2013073289W WO 2013143465 A1 WO2013143465 A1 WO 2013143465A1
Authority
WO
WIPO (PCT)
Prior art keywords
video
information
query
user
list
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.)
Ceased
Application number
PCT/CN2013/073289
Other languages
English (en)
French (fr)
Inventor
王峰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
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 Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to EP13768526.9A priority Critical patent/EP2735984A4/en
Publication of WO2013143465A1 publication Critical patent/WO2013143465A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/70Information retrieval; Database structures therefor; File system structures therefor of video data
    • G06F16/78Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/783Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content

Definitions

  • the present invention relates to the field of communications, and in particular, to a video query method, apparatus, and system. Background technique
  • the network video surveillance service is a value-added service based on remote image monitoring, transmission, storage, and management of a broadband network.
  • the service utilizes a broadband network to network decentralized, independent image collection points for unified monitoring, unified storage, unified management, and resource sharing across geographies and across the entire spectrum.
  • the network video surveillance system meets the needs of mass surveillance and professional monitoring, and is independent of the network.
  • video surveillance services can be carried out on mobile streaming media.
  • the intelligent analysis of video surveillance is the study of computer vision (CV, Computer Vision) and artificial intelligence (AI).
  • CV computer vision
  • AI artificial intelligence
  • the development goal is to establish a mapping relationship between images and event descriptions, so that the computer can distinguish and identify key target objects from the numerous video images.
  • This research makes the traditional monitoring system have a more "smart" way of learning and thinking by filtering out the useless or interfering information in the image, automatically analyzing and taking the key useful information in the video source with the powerful data processing capability of the computer.
  • Intelligent analysis technology can identify objects that appear in the video and identify the specific characteristics of the object that have beenhave.
  • Vehicle identification can identify the license plate color, license plate number, vehicle color, model and other characteristics.
  • Another example is the identification of a person who can recognize the color of the person wearing the hat, whether to wear a hat, and the color of the hat.
  • Behavior recognition can identify information such as the direction of movement of the target.
  • Some video query methods in the prior art combined with intelligent analysis technology, perform preliminary analysis on acquired video, such as urban security surveillance video, and then perform query in the analyzed video according to the feature information of the target video input by the user, A list of related videos. Then, the user queries the videos in the video list one by one. Since the number of videos in the video list is large and disorderly arranged, the video query efficiency is low, and the target video has a low hit rate.
  • Embodiments of the present invention provide a video query method, apparatus, and system, which improve video query efficiency and increase target video hit rate.
  • a video query method including:
  • An unrelated video that does not include the feature query information in the video to be checked is cut to obtain a backbone video; or, an index of the video to be checked is filtered according to the feature query information, to obtain an index of the video segment that meets the query condition;
  • a video query apparatus including:
  • a feature obtaining unit configured to acquire feature query information input by the user
  • a processing unit configured to: cut an unrelated video that does not include the feature query information in the to-be-viewed video to obtain a backbone video; or, filter an index of the video to be checked according to the feature query information, and obtain an index of the video segment that meets the query condition. ; ;
  • a sorting unit configured to sort the backbone video according to the matching degree of the feature query information to obtain a trunk video list, so that the user selects and plays the video in the trunk video list, or the video segment that meets the query condition
  • the matching degree the index of the video segment with high matching degree is placed in front, and the index of the video segment with low matching degree is placed behind to form an index list of the video segment, so that the user can perform video on the index list. Select play.
  • a video query system including: any one of the above video query devices.
  • the video query method, device, and system provided by the embodiment of the present invention include: acquiring feature query information input by a user; and cutting an irrelevant video that does not include the feature query information in the to-be-viewed video to obtain a backbone video; The video is sorted according to the matching degree of the feature query information to obtain a trunk video list, so that the user selects and plays the video in the trunk video list. In this way, by cutting or filtering irrelevant video in the video to be checked, avoid When the user queries the video, it wastes a lot of time on the unrelated video. At the same time, the trunk video is sorted according to the matching degree with the feature query information, which facilitates the user to query the video, improves the video query efficiency, and increases the target video hit rate.
  • FIG. 1 is a flowchart of a video query method according to an embodiment of the present invention
  • FIG. 2 is a flowchart of another video query method according to an embodiment of the present invention.
  • FIG. 3 is a schematic diagram of a user interface according to an embodiment of the present invention.
  • FIG. 4 is a schematic diagram of a method for interacting a user with a video query device according to an embodiment of the present invention
  • FIG. 5 is a schematic structural diagram of a video query device according to an embodiment of the present invention
  • FIG. 6 is a schematic structural diagram of another video query device according to an embodiment of the present invention
  • FIG. 7 is still another video query device according to an embodiment of the present invention
  • FIG. 8 is a schematic structural diagram of still another video query apparatus according to an embodiment of the present invention.
  • An embodiment of the present invention provides a video query method, as shown in FIG. 1 , including:
  • the user may be prompted to query the information range, and then the user selects the query information in the query range according to the query purpose, and the form may be “car/blue”.
  • the index of the video to be checked is filtered according to the feature query information, and an index of the video segment that meets the query condition is obtained.
  • the step S103 may include: editing a video that includes the feature query information in the to-be-checked video to obtain the trunk video; and editing a video that does not include the feature query information in the to-be-checked video to obtain the irrelevant video.
  • S103 Sort the backbone video according to the matching degree with the feature query information to obtain a trunk video list, so that the user selects and plays the video in the trunk video list.
  • the index of the video segment that meets the query condition is sorted according to the matching degree, and the index of the video segment with high matching degree is placed in front, and the index of the video segment with low matching degree is placed behind to form an index list of the video segment, so that The user selects and plays the video in the index list.
  • the matching degree may indicate the degree of information correlation between the backbone video and the feature query information, and the ordering manner may be a high to low form. For example, if there are five features in the feature query information, the trunk video having four of the features in the trunk video list has a higher matching degree than the trunk video having three of the features, and then has four of the features.
  • the backbone video is to be placed in front of the backbone video with 3 of these features.
  • Video query efficiency increases target video hit rate.
  • the method provided by the embodiment of the present invention may further include: sorting the irrelevant video according to the matching degree with the feature query information to obtain an unrelated video list, where the sorting manner may be high to low.
  • the unrelated video list may be after the backbone video list, so that the information in the user video and the feature query information is highly correlated, and the unrelated video and feature query The information in the information is less relevant, but in order to ensure the comprehensiveness of the video information provided to the user, although the relevance of the irrelevant video is low, it is retained.
  • irrelevant videos are images that have no motion subject. For example, when the clipped video is a video, the irrelevant video is generally a video on the road without pedestrians and vehicles at night. The output of the irrelevant video list is to prevent the user from missing valid information that exists in the unrelated video.
  • the method provided by the embodiment of the present invention may further include: obtaining The video to be checked at the time period entered by the user, which can narrow the scope of the video query.
  • the user may be prompted with a time range, which is convenient for the user to select.
  • the user may also input the time segment by himself.
  • the time period input by the user may be in the form of "a certain day of a certain month to a certain day of the year".
  • the video that is not in the video generation range can be cut to obtain the video to be checked; or the video index that is not in the video generation range of the obtained video is filtered to obtain the video index to be checked.
  • the video generation range may be set according to a specific situation.
  • the video generation range may include: any one or more of a location, an action, a behavior subject, and a color of the behavior subject in a specific time period.
  • the video generation range may be a fixed range set in advance, or may be a dynamic range randomly set by a querying person.
  • the form may be "someone does something at a certain time in a certain period of time" or "some time" A certain car is driving in a certain direction at a certain intersection.”
  • the intelligent analysis system has carried out preliminary analysis through the video generation range before analyzing the video according to the feature query information, further improving the efficiency of the video analysis and reducing the workload of the user.
  • the method further includes:
  • the video summary information may be generated after the video that does not belong to the video generation range is clipped to obtain the video to be checked, but The feature query information, and the video to be checked is clipped again, and the irrelevant video and the backbone video are obtained, so the video summary information outputted in the backbone video can generate information in the range included in the video included in the main video; output in the unrelated video
  • the video summary information can generate information in the range of videos included in irrelevant videos.
  • the clipped trunk video not only includes the blue car, but also the state of the car driving eastward, so it can output "June 1, 2011 / sedan / blue / east "Travel”, but if the user enters the feature query information as "car/blue", the clipped trunk video not only contains the blue car, the car is a car, and the state of the car traveling east is also displayed, so Can output "2011 June 1st / Car / Blue / Eastward driving".
  • the summary information may be that the intelligent analysis system displays the characteristics of the video and then splicing the individual feature information into sentences, such as "the blue car travels east on June 1, 2011", or Displayed in multiple words, such as "June 1, 2011 / Sedan / Blue / Eastward”.
  • the user can basically judge the content of the corresponding video according to the video summary information, determine the order of viewing the video, and improve the efficiency of the query.
  • the video snapshot image corresponding to the video summary information may be generated according to the video summary information; when the user triggers the video summary information, the video snapshot image or video segment is displayed.
  • the snapshot image may be a photo of a behavior subject in the summary information, such as a character, a vehicle, etc.
  • the video segment may be a behavioral short film of the behavior subject in the summary information, so that the user can view the snapshot picture and the video segment more intuitively. Key information to the video, reducing query time and improving query efficiency.
  • the video query method when used in the video query field, that is, when the video to be checked is a video, the video query method provided by the embodiment of the present invention is as shown in FIG. 2, and includes:
  • the video generation range may include: any one or more of a place, a behavior, a behavior subject, and a color of the behavior subject in a specific time period.
  • the video generation range can be preset or dynamically set by the queryer. For example, the embodiment of the present invention assumes that the video generation range is preset, and the specific form thereof may be “a certain vehicle travels in a certain direction at a certain intersection in a certain period of time”.
  • the video clip is obtained by the video to be checked, that is, when a part of the video belongs to "a certain time period, a certain vehicle is driving in a certain direction at a certain intersection", the part is reserved to obtain a video to be checked, thus, according to the video
  • the generation range is a preliminary editing of the acquired video, which facilitates the storage and classification of the video, and is also beneficial for subsequent queries.
  • the electronic map can be displayed on a normal display or a touch-sensitive display in the form of a GUI (Graphical User Interface), and then the user can view the area of the video query by means of a mouse or a touch. Pick it out.
  • GUI Graphic User Interface
  • the usual areas of urban security monitoring including major intersections, important sections, security bayonet, etc. Video recording equipment is installed in these areas. Therefore, if the video is searched throughout the city, the query efficiency is low. By selecting the area range, the video query range can be narrowed, the user's query workload is reduced, and query efficiency is provided.
  • a client device such as a computer
  • the prompt content can include: a query time period, a target type, such as a vehicle or a person, and a target color.
  • the vehicle model can be further prompted, such as a car, a truck, a commercial vehicle, a bus or a special vehicle, the license plate color and the license plate number, etc.; when the user selects a person, the person can be further prompted to wear the color of the clothes, whether to wear a cap, Hat color and whether to ride a bike or the like.
  • feature query information such as "car / blue”.
  • the video to be inspected in step S205 refers to a video recorded according to the video generation range, and then obtained after the user inputs the time period and the area range.
  • Step S205 is specifically processed by: editing a video of the feature-inquiring information in the to-be-checked video to obtain the trunk video; and editing a video that does not include the feature query information in the to-be-checked video to obtain the irrelevant video.
  • the backbone video is sorted according to the matching degree of the feature query information to obtain a trunk video list.
  • the irrelevant video is sorted according to the matching degree of the feature query information to obtain an unrelated video list, and then output the location.
  • the unrelated video list is after the main video list, so that the user can click and play the information in the video list one by one, instead of being disordered as in the prior art Massive queries in the video save customers' time and improve query efficiency.
  • the above analysis process can be performed by an intelligent analysis system.
  • the video generation range is “a certain car travels in a certain direction at a certain time interval”
  • corresponding video summary information is generated corresponding to the video in each video list according to each information in the video generation range. For example, "From June 1st, 2011 to June 2nd, 2011, the blue car traveled eastward on Kunming Road”.
  • the behavior subject in the to-be-viewed video is a vehicle
  • the vehicle in the trunk video may be queried in an intelligent analysis system and/or a GIS (Geographic Information System) and/or a vehicle information base.
  • GIS Geographic Information System
  • the vehicle information library described in the embodiment of the present invention is only a generic name, and it may be a domestic police integrated platform or a traffic network platform and other platforms. In this real-time example, its main function is to pass The license plate number is queried for more detailed vehicle information.
  • the characteristics of the target vehicle in the video are: The license plate number is Jing A 123456, and then in the vehicle information database, you can find more detailed information according to the license plate, and finally get the "red car / 2005 registration Santana 2000 / car advocates three / phone 147147 / Jing A 123456".
  • the video summary information is output, and the summary information may include the above vehicle information.
  • the video summary information may be displayed all or partially.
  • the user moves the mouse, stylus, etc. to the hidden summary, all the displays are displayed.
  • the analysis of the driving direction of the vehicle is as follows: acquiring the position information of the camera corresponding to the video to be inspected; determining the driving of the vehicle according to the position information of the camera and the moving direction of the vehicle in the video to be checked relative to the camera Direction; the output video summary includes the direction in which the vehicle is traveling.
  • the system queries the GIS for the location information of the camera according to the camera ID of the video recording: including the exact location of the camera, such as Kunming Road Cross, the direction of the road segment: such as north-south trend, east-west direction, etc., and then combined with the intelligent analysis system for the target vehicle Relative to the moving direction of the camera, such as: moving from left to right, calculating the actual moving direction of the target vehicle.
  • the intelligent analysis shows that the object is in the video from left to right. mobile. Then, after calculation, it is concluded that the actual moving direction of the target object is from west to east.
  • the output video summary information can then include "Traveling from west to east,”.
  • the different feature descriptions may be output.
  • the feature query information input by the user is: the license plate number is Jing A 123456/blue, but if the vehicle is red in the vehicle information database, the different feature descriptions are displayed in the summary information, and can be displayed as "King A 123456 / blue (red)", the content in parentheses can be different from other words, to emphasize the role, easy for the user to distinguish.
  • the video snapshot image corresponding to the video summary information may be generated according to the video summary information.
  • the video snapshot image is displayed.
  • a video clip may be displayed.
  • the snapshot image may be a photo of the behavior subject in the summary information, such as a character, a vehicle, etc.
  • the video segment may be a short film of the behavior of the subject in the summary information.
  • the content of the video summary information should be covered in the image and the video segment.
  • the vehicle, the license plate, and the vehicle color are displayed through a picture or a short film in a few seconds to facilitate the user's inquiry.
  • the video summary information and the video snapshot image may be simultaneously provided.
  • the user can select the video to play in the random trunk list and the irrelevant list in order, or randomly select the video in the trunk list and the irrelevant list according to the text summary displayed in the list.
  • the associated video may be searched for in the to-be-searched video according to the feature information of the video in the selected video.
  • the video list may be directly generated without processing, or the related video clip may be made into a backbone video and an irrelevant video by referring to the method in step S202, and then the main video and the irrelevant video are respectively sorted to generate an associated video list.
  • the list can contain text summary information.
  • S210 Output the associated video list, so that the user selects and plays the video in the associated video list.
  • the output of the associated video expands the video range associated with the target video that the user is querying, increasing the hit rate of the target video.
  • a navigation link can be set in the user interface, and the user can switch between the associated video list and the original video list by clicking the navigation link.
  • the user interface 30 can be as shown in FIG. 3, and the query time period 301 can be selected by the user.
  • the feature query information 302 can be used by the user to select the feature query information according to the feature information input prompt.
  • Area A can display a list of videos, assuming that the user clicks on Play Record 1, and Area B displays a list of related videos for Record 1.
  • the area C may be an electronic map area for the user to select a video range when starting the query, but when the number of video lists is large, the electronic map may be hidden, and the area C is used to display the video list.
  • the navigation link 303 can display the currently playing video number, and the user can switch the associated video list and the original video list by clicking the navigation link.
  • the interaction method between the user and the video query device provided by the present invention is as shown in FIG. 4 .
  • the video query device displays a video query interface to the user.
  • the user selects the range of the camera device corresponding to the video to be queried on the electronic map in the video query interface, and then the user selects the time period to be queried in the query time period, selects the feature query information in the feature query information, and then Click OK to make a query.
  • the video query device obtains the video to be inspected according to the obtained time period and area range selected by the user, edits and sorts the video to be checked, and generates video summary information, etc., and presents the user interface to the user, and the specific interface is as shown in FIG. Show.
  • the above method is only an example, and the corresponding steps may be added or subtracted in practical applications. Any person skilled in the art can easily think of the changed method within the technical scope disclosed by the present invention, and therefore will not be described again.
  • the video query method provided by the embodiment of the present invention cuts off irrelevant video in the video to be inspected, so that the user avoids wasting a lot of time on the unrelated video when querying the video, and sorts the main video according to the matching degree with the feature query information.
  • the user queries the video, which improves the efficiency of the video query and increases the target video hit rate.
  • the user can obtain the video information more intuitively, which is convenient for the video query.
  • a video querying apparatus 50 includes: a feature acquiring unit 501, configured to acquire feature query information input by a user.
  • the feature acquisition unit 501 can be implemented by a mobile phone terminal, a PDA (Personal Digital Assistant), a PC (personal computer), or a monitoring center device.
  • a PDA Personal Digital Assistant
  • PC personal computer
  • the first processing unit 502 is configured to cut an unrelated video that does not include the feature query information in the to-be-viewed video to obtain a backbone video.
  • the first processing unit 502 is specifically configured to:
  • the first sorting unit 503 is configured to sort the trunk video according to the matching degree with the feature query information to obtain a trunk video list, so that the user selects and plays the video in the trunk video list.
  • the index of the video segment that meets the query condition is sorted according to the matching degree, and the index of the video segment with high matching degree is placed in front, and the index of the video segment with low matching degree is placed behind to form an index list of the video segment, so that The user selects and plays the video in the index list.
  • the processing unit cuts or filters the irrelevant video in the video to be checked, so that the user does not waste a lot of time on the unrelated video when the video is queried, and the first sorting unit selects the main video according to the feature acquired by the feature acquiring unit.
  • Query information matching degree sorting which is convenient for users to query video, improve video query efficiency, and increase target video hit rate.
  • the video query device 50 further includes:
  • the time obtaining unit 504 is configured to acquire a to-be-checked video in a time period input by the user.
  • the second sorting unit 505 is configured to sort the irrelevant videos according to the matching degree with the feature query information to obtain an unrelated video list.
  • the first sorting unit 503 and the second sorting unit 505 may be combined into one, or may be completed by a device such as a computer.
  • a video output unit 506 configured to output the trunk video list and the unrelated video list, where the unrelated video list may be after the trunk video list, so that the user is in the trunk video list and the unrelated video list The video is selected for playback.
  • the second processing unit 507 is configured to clip the video that is not in the video generation range of the acquired video to obtain a video to be checked.
  • the video generation range may include any one or more of a place, a behavior, a behavior subject, and a color of a behavior subject in a specific time period.
  • the first processing unit 502 and the second processing unit 507 can be combined into one.
  • the summary generating unit 508 is configured to generate corresponding video summary information for each video in the corresponding video list according to each information in the video generation range.
  • the summary output unit 509 is configured to output the video summary information for each video in the corresponding video list.
  • the generating unit 510 is configured to generate a video snapshot image or a video segment corresponding to the video summary information.
  • the display unit 511 is configured to display the video snapshot image or video segment when the user triggers the video summary information.
  • the receiving unit 512 is configured to receive a video selected by a user.
  • the playing unit 513 is configured to play the video selected by the user.
  • the association unit 514 is configured to belong to the video generation range according to the feature information of the video in the video selected by the user.
  • the list output unit 515 is configured to output the associated video list, so that the user selects and plays the video in the associated video list.
  • the video to be checked is a video.
  • the device 50 further includes: a range obtaining unit 516, configured to acquire a range of regions selected by the user on the electronic map.
  • the video obtaining unit 517 is configured to acquire a video in the area range as a video to be checked.
  • the apparatus 50 further includes:
  • the query unit 518 is configured to query, in the intelligent analysis system and/or the geographic information system GIS and/or the vehicle information database, related information of the vehicle in the to-be-checked video.
  • the video summary output by the digest output unit 509 includes related information of the vehicle.
  • the difference output unit 519 is configured to output the different feature description when different feature descriptions exist for the same feature in the feature query information and the related information of the vehicle.
  • the feature description can be displayed in the video summary information.
  • the information acquiring unit 520 is configured to acquire the location information of the camera corresponding to the to-be-viewed video.
  • the determining unit 521 is configured to determine the traveling direction of the vehicle according to the camera position information and the moving direction of the vehicle in the to-be-checked video with respect to the camera.
  • the video summary output by the digest output unit 509 includes the vehicle traveling direction.
  • the video query device provided by the embodiment of the present invention cuts out the irrelevant video in the video to be inspected by the processing unit, and prevents the user from wasting a lot of time on the unrelated video when the video is queried, and the first sorting unit acquires the trunk video according to the feature.
  • the feature query information matching degree obtained by the unit is sorted, which is convenient for the user to query the video, improves the video query efficiency, and increases the target video hit rate.
  • An embodiment of the present invention provides a video query system, including: any of the above-mentioned video query devices.
  • the video query system can also include:
  • the camera device can also store the above video.
  • the above video may also be stored in the central server of the urban security monitoring system.
  • the present invention does not limit the specific storage location of the video.
  • a geographic information system configured to provide spatial information to the video query device, so that the video query system queries related information of the vehicle in the video, and/or
  • a vehicle information base configured to provide vehicle information to the video query device, so that the video query system queries the related information of the vehicle in the video to be checked.
  • the foregoing program may be stored in a computer readable storage medium, and the program is executed when executed.
  • the foregoing steps include the steps of the foregoing method embodiments; and the foregoing storage medium includes: a medium that can store program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk.

Landscapes

  • Engineering & Computer Science (AREA)
  • Library & Information Science (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Television Signal Processing For Recording (AREA)

Abstract

本发明实施例提供的视频查询方法、装置与系统,涉及通信领域,能够提高视频查询效率,增大目标视频命中率。该视频查询方法包括:获取用户输入的特征查询信息;将待查视频中不包括所述特征查询信息的无关视频剪切,得到主干视频;或者,根据特征查询信息对待查视频的索引进行过滤,得到符合查询条件的视频段的索引;将所述主干视频按照与所述特征查询信息匹配度排序得到主干视频列表,以便于用户对所述主干视频列表中的视频进行选择播放;或者,对符合查询条件的视频段的索引根据匹配度排序,形成视频段的索引列表,以便于用户对索引列表中的视频进行选择播放。

Description

一种视频查询方法、 装置与系统 本申请要求于 2012 年 3 月 27 日提交中国专利局、 申请号为 201210084264.0、 发明名称为"一种视频查询方法、 装置与系统"的中国专利 申请的优先权, 其全部内容通过引用结合在本申请中。
技术领域
本发明涉及通讯领域, 尤其涉及一种视频查询方法、 装置与系统。 背景技术
现有技术中, 网络视频监控业务是基于宽带网的远程图像监控、 传输、 存储, 管理的增值业务。 该业务利用宽带网络, 将分散、 独立 的图像釆集点进行联网, 实现跨地域、 全范围内的统一监控、 统一存 储、 统一管理及资源共享。 网络视频监控系统同时满足大众监控和专 业监控需求, 独立组网, 同时可以在移动流媒体上开展视频监控业务。
视频监控的智能分析, 是对计算机视觉 (CV , Computer Vision ) 与人工智能 (AI, Artificial Intelligent ) 的研究。 其核心是视频内容分 析, 发展目标在于将图像与事件描述之间建立一种映射关系, 使计算 机从纷繁的视频图像中分辩、 识别出关键目标物体。 这一研究借助计 算机强大的数据处理能力过滤掉图像中无用的或干扰信息、 自动分析 并取视频源中的关键有用信息, 从而使传统监控系统具有更为 "聪明" 的学习思考方式。
智能分析技术可以对视频中出现的物体进行识别, 识别出物体的 具体特征已经物体行为。 如: 车辆识别, 可以识别出车牌颜色、 车牌 号码、 车辆颜色、 车型等特征。 又如人员识别, 可以识别出人物的穿 着颜色、 是否戴帽, 以及帽子颜色等。 而行为识别可以识别出目标物 的移动方向等信息。
现有技术中的一些视频查询方法, 结合智能分析技术对获取的视频, 如 城市安全监控录像, 进行初步分析, 然后根据用户输入的目标视频的特 征信息, 在分析后的视频中进行查询, 得到相关视频列表。 然后用户对 视频列表中的视频进行逐一查询, 由于视频列表中的视频数量较多, 且 无序排列, 因此视频查询效率低, 同时目标视频的命中率低。 发明内容
本发明的实施例提供一种视频查询方法、 装置与系统, 提高视频 查询效率, 增大目标视频命中率。
为达到上述目的, 本发明的实施例釆用如下技术方案:
一方面, 提供一种视频查询方法, 包括:
获取用户输入的特征查询信息;
将待查视频中不包括所述特征查询信息的无关视频剪切, 得到主 干视频; 或者, 根据特征查询信息对待查视频的索引进行过滤, 得到符合 查询条件的视频段的索引;
将所述主干视频按照与所述特征查询信息匹配度排序得到主干视 频列表,以便于用户对所述主干视频列表中的视频进行选择播放;或者, 对符合查询条件的视频段的索引根据匹配度排序, 形成视频段的索引列表, 以便于用户对索引列表中的视频进行选择播放。
一方面, 提供一种视频查询装置, 包括:
特征获取单元, 用于获取用户输入的特征查询信息;
处理单元, 用于将待查视频中不包括所述特征查询信息的无关视 频剪切, 得到主干视频; 或者, 根据特征查询信息对待查视频的索引进行 过滤, 得到符合查询条件的视频段的索引; ;
排序单元, 用于将所述主干视频按照与所述特征查询信息匹配度 排序得到主干视频列表, 以便于用户对所述主干视频列表中的视频进 行选择播放, 或者, 对符合查询条件的视频段的索引根据匹配度排序, 把 匹配度高的视频段的索引放在前面, 把匹配度低的视频段的索引放在后面, 形成视频段的索引列表, 以便于用户对索引列表中的视频进行选择播放。
一方面, 提供一种视频查询系统, 包括: 上述任意一种视频查询 装置。
本发明实施例提供的视频查询方法、 装置与系统, 包括: 获取用户 输入的特征查询信息; 将待查视频中不包括所述特征查询信息的无关视 频剪切, 得到主干视频; 将所述主干视频按照与所述特征查询信息匹配 度排序得到主干视频列表, 以便于用户对所述主干视频列表中的视频进 行选择播放。 这样一来, 通过将待查视频中无关视频剪切或过滤, 避免 用户在查询视频时在无关视频上浪费大量时间, 同时将主干视频按照与 特征查询信息匹配度排序, 方便用户查询视频, 提高了视频查询效率, 增大了目标视频命中率。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案,下面将对实 施例或现有技术描述中所需要使用的附图作简单地介绍, 显而易见地, 下面 描述中的附图仅仅是本发明的一些实施例, 对于本领域普通技术人员来讲, 在不付出创造性劳动的前提下, 还可以根据这些附图获得其他的附图。
图 1为本发明实施例提供一种视频查询方法流程图;
图 2为本发明实施例提供另一种视频查询方法流程图;
图 3为本发明实施例提供的一种视用户界面示意图;
图 4 为本发明实施例提供的一种用户与视频查询装置的交互方法 示意图;
图 5为本发明实施例提供的一种视频查询装置结构示意图; 图 6为本发明实施例提供的另一种视频查询装置结构示意图; 图 7为本发明实施例提供的又一种视频查询装置结构示意图; 图 8为本发明实施例提供的再一种视频查询装置结构示意图。 具体实施方式
下面将结合本发明实施例中的附图,对本发明实施例中的技术方案进行 清楚、 完整地描述, 显然, 所描述的实施例仅仅是本发明一部分实施例, 而 不是全部的实施例。基于本发明中的实施例, 本领域普通技术人员在没有做 出创造性劳动前提下所获得的所有其他实施例 , 都属于本发明保护的范围。
本发明实施例提供一种视频查询方法, 如图 1所示, 包括:
5101、 获取用户输入的特征查询信息。
可选的, 在用户输入特征查询信息前, 可以向用户提示查询信息范围, 然后用户根据查询目的,自行在查询范围中选择查询信息,其形式可以为 "轿 车 /蓝色" 。
5102、将待查视频中不包括所述特征查询信息的无关视频剪切,得到主 干视频。 或者, 根据特征查询信息对待查视频的索引进行过滤, 得到符合查 询条件的视频段的索引。
可选的, 步骤 S103可以包括: 剪辑所述待查视频中包括所述特征查询 信息的视频得到所述主干视频;剪辑所述待查视频中不包括所述特征查询信 息的视频得到所述无关视频。
S103、将所述主干视频按照与特征查询信息匹配度排序得到主干视频列 表, 以便于用户对所述主干视频列表中的视频进行选择播放。 或者, 对符合 查询条件的视频段的索引根据匹配度排序,把匹配度高的视频段的索引放在 前面, 把匹配度低的视频段的索引放在后面, 形成视频段的索引列表, 以便 于用户对索引列表中的视频进行选择播放。
所述匹配度可以表示所述主干视频与特征查询信息的信息相关程度,所 述排序方式可以是从高到低的形式。示例的,若特征查询信息中有 5个特征, 则上述主干视频列表中具有 4个所述特征的主干视频比具有 3个所述特征的 主干视频匹配度高,那么具有 4个所述特征的主干视频要排在具有 3个所述 特征的主干视频前面。
这样一来,通过将待查视频中无关视频剪切掉,避免用户在查询视频时 在无关视频上浪费大量时间,同时将主干视频按照与特征查询信息匹配度排 序, 方便用户查询视频, 提高了视频查询效率, 增大了目标视频命中率。
进一步的, 在步骤 S103后, 本发明实施例提供的方法还可以包括: 将所述无关视频按照与所述特征查询信息匹配度排序得到无关视频列 表, 所述排序方式可以是从高到低的形式; 输出所述主干视频列表与所述无 关视频列表, 所述无关视频列表可以在所述主干视频列表之后, 以便于用户 视频与特征查询信息中的信息相关度较高,无关视频与特征查询信息中的信 息相关度较低, 但是为了保证向用户提供的视频信息的全面性, 虽然无关视 频的相关度较低,也予以保留。通常,无关视频为一些没有动作主体的影像, 例如, 在被剪辑的视频为录像时, 无关视频一般为深夜没有行人和车辆的道 路上的录像。无关视频列表的输出是为了避免用户错失存在于无关视频中的 有效信息的情况发生。
特别的, 在步骤 S101前, 本发明实施例提供的方法还可以包括: 获取 处于用户输入的时间段的待查视频,这样可以缩小视频查询的范围。可选的, 在用户输入视频查询的时间段前, 可以向用户提示时间段范围, 便于用户选 择, 当然, 用户也可以自行输入时间段。 用户输入的时间段可以是 "某年某 月某日至某年某月某日" 的形式。
同时,还可以将获取的视频中不属于视频生成范围的视频剪切得到待查 视频; 或者, 将获取的视频中不属于视频生成范围的视频索引进行过滤得到 待查视频索引。
所述视频生成范围可以根据具体情况设定, 示例的, 可以包括: 特定时 间段中的地点、 行为、 行为主体、 行为主体的颜色中的任意一种或多种。 该 视频生成范围可以是预先设置的固定范围,也可以是查询人员随机设置的动 态范围, 示例的, 其形式可以为 "某时间段某人在某地做某事" , 也可以是 "某时间段某车在某路口向某方向行驶" 。
这样一来, 智能分析系统在根据特征查询信息对视频进行分析前, 已经 通过视频生成范围进行了初步的分析, 进一步提高了视频分析的效率, 减少 了用户的工作量。
进一步的, 所述方法还包括:
根据所述视频生成范围中的各信息,对应视频列表中的每个视频生成相 应的视频摘要信息; 对应视频列表中的每个视频输出所述视频摘要信息。 通 常视频生成范围中的视频特征要比特征查询信息中的视频特征全面,因此上 述视频摘要信息可以在将获取的视频中不属于视频生成范围的视频剪切得 到待查视频后生成, 但由于根据特征查询信息, 又对待查视频进行了再次剪 辑, 得到了无关视频和主干视频, 因此在主干视频中输出的视频摘要信息可 以为主干视频中包含的视频生成范围中的信息;在无关视频中输出的视频摘 要信息可以为无关视频中包含的视频生成范围中的信息。 示例的,假设视频 生成范围为: "2011年 6月 1 日 /轿车 /蓝色 /向东行驶" , 用户输入的时间段 为: "2011年 6月 1 日" , 而用户输入的特征查询信息为 "轿车 /蓝色" , 剪辑后的的主干视频不仅包含了蓝色的轿车,轿车向东行驶的状态也显示出 来, 因此可以输出 "2011年 6月 1 日 /轿车 /蓝色 /向东行驶" , 但如果用户输 入的特征查询信息为 "车 /蓝色" , 剪辑后的主干视频不仅包含了蓝色的车, 该车为轿车, 且轿车向东行驶的状态也显示出来, 因此仍可以输出 "2011 年 6月 1 日 /轿车 /蓝色 /向东行驶" 。
特别的, 上述摘要信息可以是智能分析系统通过对视频的特征分析后, 将各个独立的特征信息拼接成句子显示出来, 如 "2011年 6月 1 日蓝色轿 车向东行驶" , 也可以是以多个词语的形式显示出来, 如 "2011 年 6 月 1 日 /轿车 /蓝色 /向东行驶" 。
这样一来,用户可以根据视频摘要信息对相应视频的内容进行基本的判 断, 决定查看视频的先后顺序, 提高查询的效率。
进一步的,还可以根据所述视频摘要信息, 生成所述视频摘要信息对应 的视频快照图像; 当用户触发所述视频摘要信息时, 显示所述视频快照图像 或视频片段。 上述快照图片可以是摘要信息中行为主体的照片, 如人物、 车 辆等, 上述视频片段可以是摘要信息中行为主体的行为短片, 这样一来, 用 户通过快照图片和视频片段可以更为直观地看到视频中的关键信息,从而减 少查询时间, 提高查询效率。
示例的, 当所述视频查询方法在录像查询领域使用时, 即所述待查视频 为录像时, 本发明实施例提供的视频查询方法如图 2所示, 包括:
S201、 将获取的视频中不属于视频生成范围的视频剪切得到待查视频。 所述视频生成范围可以包括: 特定时间段中的地点、 行为、 行为主体、 行为主体的颜色中的任意一种或多种。 该视频生成范围可以是预先设置的, 也可以是查询人员动态设置的。 示例的, 本发明实施例假设该视频生成范围 为预设的, 其具体形式可以为 "某时间段某车在某路口向某方向行驶" 。
根据上述视频生成范围, 将视频剪辑得到待查视频, 即当视频中有一部 分属于 "某时间段某车在某路口向某方向行驶" 时, 将该部分保留得到待查 视频, 这样, 根据视频生成范围对获取的视频进行初步剪辑, 便于录像的存 储与归类, 也有益于后续的查询。
S202、 获取处于用户输入的时间段的待查视频。
S203、 获取用户在电子地图上选择的区域范围的待查视频。
在实际应用中, 电子地图可以以 GUI ( Graphical User Interface , 图形用 户界面)的形式在普通显示器或触敏显示器等设备上显示出来, 然后用户通 过鼠标或点触等方式将视频查询的区域范围框选出来。
通常的城市安全监控领域, 包括各大路口、 重要路段、 治安卡口等等, 在这些领域都设置有录像设备, 因此, 如果查询整个城市范围内的录像, 查 询效率低, 通过对区域范围的选择可以缩小视频查询范围, 减少用户的查询 工作量, 提供查询效率。
5204、 获取用户输入的特征查询信息。
可选的, 通过客户端设备, 如电脑等, 可以向用户提供特征查询选项以 便于用户选择, 提示的内容可包括: 查询时间段, 目标种类, 如车辆或人员 等, 目标颜色。 当用户选择车辆, 可以进一步提示车辆型号, 如轿车、货车、 商务车、 大巴车或特种车等, 车牌颜色和车牌号码等; 当用户选择人员, 可 以进一步提示人员穿着衣服颜色、是否戴帽、帽子颜色和是否骑车等。 当然, 用户也可以直接进行特征查询信息输入, 如 "轿车 /蓝色" 。
然后获取用户输入的特征查询信息。
5205、将待查视频中不包括所述特征查询信息的无关视频剪切,得到主 干视频。
示例的, 步骤 S205中的待查视频指的是依据视频生成范围进行初步剪 辑, 然后再在用户输入的时间段和区域范围筛选后得到的录像视频。
步骤 S205具体处理步骤为: 剪辑所述待查视频中包括所述特征查询信 息的视频得到所述主干视频;剪辑所述待查视频中不包括所述特征查询信息 的视频得到所述无关视频。
5206、 分析获取主干视频列表和无关视频列表。
将所述主干视频按照与所述特征查询信息匹配度从高到低排序得到主 干视频列表, 进一步的, 将所述无关视频按照与所述特征查询信息匹配度排 序得到无关视频列表, 然后输出所述主干视频列表与所述无关视频列表, 所 述无关视频列表在所述主干视频列表之后,这样用户可以按顺序逐个点击播 放视频列表中的信息,而不是像现有技术中,在无序的视频中进行海量查询, 节约了客户的时间, 提高了查询效率。 特别的, 上述分析过程可以由智能分 析系统完成。
示例的, 当上述视频生成范围为 "某时间段某车在某路口向某方向行 驶" , 根据所述视频生成范围中的各信息, 对应每个视频列表中的视频生成 相应的视频摘要信息, 如 "2011年 6月 1 日至 2011年 6月 2日蓝色轿车在 昆明路向东行史" 。 进一步的, 当所述待查视频中的行为主体为车辆时, 可以在智能分析系 统和 /或 GIS ( Geographic Information System, 地理信息系统)和 /或车辆信 息库中查询所述主干视频中的车辆的相关信息。 需要说明的是, 本发明实施 例中所述的车辆信息库只是一个通用叫法,它可以是国内的警务综合平台或 交管网平台等其他平台,在本实时例中其主要功能就是可以通过车牌号码查 询到更加详细完善的车辆信息。
这样一来, 可以为用户提供更为全面的信息, 方便用户进行视频查询。 如釆用智能分析得出视频中目标车辆的特征为: 车牌号为京 A 123456, 然 后再在车辆信息库中查询,可以根据车牌查询到更详细的信息,最终得到 "红 色轿车 /2005年注册桑塔纳 2000/车主张三 /电话 147147/京 A 123456" 。 然 后, 输出所述视频摘要信息, 该摘要信息可以包括上述车辆信息。 上述视频 摘要信息在输出时, 可以全部显示, 也可以部分隐藏, 在用户将鼠标、 手写 笔等移动到该隐藏摘要出时, 再进行全部显示。
其中, 车辆行驶方向的分析如下: 获取所述待查视频对应摄像装置位置 信息;根据所述摄像装置位置信息与所述待查视频中车辆相对于所述摄像装 置移动方向, 确定所述车辆行驶方向; 输出的视频摘要包括所述车辆行驶方 向。示例的, 系统根据视频录像所属摄像机标识去 GIS中查询该摄像机所在 位置信息: 包括该摄像机确切位置,如昆明路十字,路段走向: 如南北走向、 东西走向等, 然后结合智能分析系统对目标车辆相对摄像机移动方向, 如: 从左向右移动, 计算得出目标车辆实际移动方向, 示例的, 当摄像机安装在 道路南侧, 道路为东西走向, 智能分析得出物体在视频中由左向右移动。 然 后经过计算, 得出目标物体实际移动方向由西向东。 然后输出的视频摘要信 息可以包括 "由西向东行驶,,。
特别的,当所述特征查询信息与所述车辆的相关信息中对相同的特征存 在不同的特征描述时, 可以输出所述不同的特征描述。 示例的, 如用户输入 的特征查询信息为: 车牌号为京 A 123456/蓝色, 但在车辆信息库中查询得 到该车为红色, 则在摘要信息中显示上述不同的特征描述, 可以显示为 "京 A 123456/蓝色 (红色) " , 括号中内容可以带有与其他字样不同的颜色, 起到强调作用, 便于用户进行区分。
进一步的,还可以根据所述视频摘要信息, 生成所述视频摘要信息对应 的视频快照图像; 当用户触发所述视频摘要信息时, 显示所述视频快照图像 或视频片段。 上述快照图片可以是摘要信息中行为主体的照片, 如人物、 车 辆等, 上述视频片段可以是摘要信息中行为主体的行为短片, 当然, 该图片 和视频片段中应该尽量涵盖视频摘要信息中的内容, 如将车辆、 车牌、 车辆 颜色通过一张图片或几秒中的短片显示出来, 以便于用户查询。
上述主干视频列表和无关视频列表中的视频在输出时,可以同时带有上 述视频摘要信息和视频快照图像。
5207、 接收用户选定的视频。
用户可以按顺序在随机主干列表和无关列表中选择视频进行播放 ,也可 以根据列表中显示的文字摘要在主干列表和无关列表中随机选择视频进行 播放。
5208、 播放所述用户选定的视频。
5209、根据所述用户选定的视频中视频的特征信息,在所述待查视频中 获取所述用户选定的视频的关联视频索引列表。所述视频的特征信息属于所 述视频生成范围。
在用户选定视频播放时,可以根据该选定的视频中视频的特征信息在待 查视频中查找关联视频。 对于上述关联视频, 可以不进行处理, 直接生成视 频列表, 也可以参考步骤 S202中的方法, 将关联视频剪辑成为主干视频和 无关视频, 然后再分别对主干视频和无关视频进行排序生成关联视频列表。 同时, 该列表中可以带有文字摘要信息。
S210、输出所述关联视频列表, 以便用户对所述关联视频列表中的视频 进行选择播放。
需要说明的是,本发明实施例提供的视频查询方法步骤的先后顺序可以 进行适当调整, 步骤的数量也可以根据情况进行相应增减, 任何熟悉本技术 领域的技术人员在本发明揭露的技术范围内, 可轻易想到变化的方法, 因此 不再赘述。
特别的,上述关联视频的输出扩大了与用户要查询的目标视频相关的视 频范围, 增加了目标视频的命中率。 进一步的, 可以在用户界面设置导航链 接, 用户可以通过点击导航链接进行关联视频列表与原视频列表的切换。
示例的, 用户界面 30可以如图 3所示, 查询时间段 301 可以供用户选 择查询的开始时间和结束时间,特征查询信息 302可以供用户根据特征信息 输入提示选择特征查询信息。 区域 A 可以显示视频列表, 假设用户点击播 放录像 1 , 区域 B显示录像 1的相关视频列表。 特别的, 区域 C可以是电子 地图区域, 供用户在开始查询时框选视频范围, 但当视频列表数量较多时, 电子地图可以隐藏, 区域 C用来显示视频列表。 同时, 导航链接 303可以显 示当前播放的视频编号,用户可以通过点击导航链接进行关联视频列表与原 视频列表的切换。
可选的, 用户与本发明提供的视频查询装置的交互方法如图 4所示。 首 先, 视频查询装置向用户显示视频查询界面。 然后, 用户在视频查询界面中 的电子地图上框选需要查询的录像对应的摄像装置区域范围,接着用户分别 在查询时间段选择要查询的时间段, 在特征查询信息处选择特征查询信息 , 然后点击确定, 进行查询。 视频查询装置根据获取的用户选择的时间段、 区 域范围在录像中获取待查视频, 对待查视频进行剪辑、 排序, 并生成视频摘 要信息等, 在用户界面向用户展示, 具体界面如图 3所示。 特别的, 上述方 法只是举例说明, 实际应用中可以增加或减少相应步骤, 任何熟悉本技术领 域的技术人员在本发明揭露的技术范围内, 可轻易想到变化的方法, 因此不 再赘述。
本发明实施例提供的视频查询方法, 通过将待查视频中无关视频剪切 掉, 避免用户在查询视频时在无关视频上浪费大量时间, 同时将主干视频按 照与特征查询信息匹配度排序, 方便用户查询视频, 提高了视频查询效率, 增大了目标视频命中率。 同时通过视频摘要信息和视频快照图片的输出, 使 用户更直观地获取视频信息, 便于视频的查询。 本发明实施例提供的一种视频查询装置 50, 如图 5所示, 包括: 特征获取单元 501 , 用于获取用户输入的特征查询信息。
特征获取单元 501可以由手机终端、 PDA (Personal Digital Assistant, 掌 上电脑)、 PC(personal computer, 个人计算机)或监控中心设备来实现。
第一处理单元 502, 用于将待查视频中不包括所述特征查询信息的无关 视频剪切, 得到主干视频。 所述第一处理单元 502具体用于:
剪辑所述待查视频中包括所述特征查询信息的视频得到所述主干视频; 剪辑所述待查视频中不包括所述特征查询信息的视频得到所述无关视频。 第一排序单元 503 , 用于将所述主干视频按照与所述特征查询信息匹配 度排序得到主干视频列表,以便于用户对所述主干视频列表中的视频进行选 择播放。 或者, 对符合查询条件的视频段的索引根据匹配度排序, 把匹配度 高的视频段的索引放在前面, 把匹配度低的视频段的索引放在后面, 形成视 频段的索引列表, 以便于用户对索引列表中的视频进行选择播放。
这样一来,通过处理单元将待查视频中无关视频剪切或过滤掉,避免用 户在查询视频时在无关视频上浪费大量时间,同时第一排序单元将主干视频 按照与特征获取单元获取的特征查询信息匹配度排序, 方便用户查询视频, 提高了视频查询效率, 增大了目标视频命中率。
进一步的, 如图 6所示, 所述视频查询装置 50还包括:
时间获取单元 504 , 用于获取处于用户输入的时间段的待查视频。
第二排序单元 505 , 用于将所述无关视频按照与所述特征查询信息匹配 度排序得到无关视频列表。
第一排序单元 503和第二排序单元 505可以合并为一体,也可以由电脑 等设备完成。
视频输出单元 506 , 用于输出所述主干视频列表与所述无关视频列表, 所述无关视频列表可以在所述主干视频列表之后,以便于用户对所述主干视 频列表与所述无关视频列表中的视频进行选择播放。
第二处理单元 507 , 用于将获取的视频中不属于视频生成范围的视频剪 切得到待查视频。所述视频生成范围可以包括:特定时间段中的地点、行为、 行为主体、行为主体的颜色中的任意一种或多种。 第一处理单元 502和第二 处理单元 507可以合并为一体。
摘要生成单元 508 , 用于根据所述视频生成范围中的各信息, 对应视频 列表中的每个视频生成相应的视频摘要信息。
摘要输出单元 509 , 用于对应视频列表中的每个视频输出所述视频摘要 信息。
生成单元 510 , 用于生成所述视频摘要信息对应的视频快照图像或视频 片段。
显示单元 511 , 用于在用户触发所述视频摘要信息时, 显示所述视频快 照图像或视频片段。
接收单元 512 , 用于接收用户选定的视频。 播放单元 513 , 用于播放所述用户选定的视频。
关联单元 514, 用于根据所述用户选定的视频中视频的特征信息, 在所 息属于所述视频生成范围。
列表输出单元 515 , 用于输出所述关联视频列表, 以便用户对所述关联 视频列表中的视频进行选择播放。
进一步的, 所述待查视频为录像, 如图 7所示, 所述装置 50还包括: 范围获取单元 516, 用于获取用户在电子地图上选择的区域范围。
视频获取单元 517 , 用于获取所述区域范围内的录像作为待查视频。 当所述待查视频中的行为主体为车辆时, 如图 8所示, 所述装置 50还 包括:
查询单元 518, 用于在智能分析系统和 /或地理信息系统 GIS和 /或车辆 信息库中查询所述待查视频中的车辆的相关信息。
此时,所述摘要输出单元 509输出的所述视频摘要包括所述车辆的相关 信息。
差异输出单元 519 , 用于在所述特征查询信息与所述车辆的相关信息中 对相同的特征存在不同的特征描述时, 输出所述不同的特征描述。 特别的, 所述特征描述可以在视频摘要信息中显示出来。
信息获取单元 520, 用于获取所述待查视频对应摄像装置位置信息。 确定单元 521 , 用于根据所述摄像装置位置信息与所述待查视频中车辆 相对于所述摄像装置移动方向, 确定所述车辆行驶方向。
此时,所述摘要输出单元 509输出的所述视频摘要包括所述车辆行驶方 向。
本发明实施例提供的视频查询装置,通过处理单元将待查视频中无关视 频剪切掉, 避免用户在查询视频时在无关视频上浪费大量时间, 同时第一排 序单元将主干视频按照与特征获取单元获取的特征查询信息匹配度排序,方 便用户查询视频, 提高了视频查询效率, 增大了目标视频命中率。 本发明实施例提供一种视频查询系统, 包括: 上述任意所述的视频查询 装置。
进一步的, 当上述系统应用于城市安全监控系统中时, 所述视频查询系 统还可以包括:
摄像装置, 用于录像的釆集。 同时, 该摄像装置也可以存储上述录像。 特别的, 在实际应用中, 上述录像也可以存储于城市安全监控系统的中心服 务器上, 这里只是举例说明, 本发明对录像的具体存储位置不做限制。
智能分析系统, 用于对待查视频进行分析。
地理信息系统, 用于向所述视频查询装置提供空间信息, 以便于所述视 频查询系统对待查视频中的车辆的相关信息进行查询, 和 /或
车辆信息库, 用于向所述视频查询装置提供车辆信息, 以便于所述视频 查询系统对待查视频中的车辆的相关信息进行查询。
本领域普通技术人员可以理解:实现上述方法实施例的全部或部分步骤 可以通过程序指令相关的硬件来完成,前述的程序可以存储于一计算机可读 取存储介质中, 该程序在执行时, 执行包括上述方法实施例的步骤; 而前述 的存储介质包括: ROM、 RAM, 磁碟或者光盘等各种可以存储程序代码的 介质。
以上所述, 仅为本发明的具体实施方式, 但本发明的保护范围并不 局限于此, 任何熟悉本技术领域的技术人员在本发明揭露的技术范围内, 可轻易想到变化或替换, 都应涵盖在本发明的保护范围之内。 因此, 本 发明的保护范围应以所述权利要求的保护范围为准。

Claims

权利要求
1、 一种视频查询方法, 其特征在于, 包括:
获取用户输入的特征查询信息;
将待查视频中不包括所述特征查询信息的无关视频剪切, 得到主 干视频; 或者, 根据特征查询信息对待查视频的索引进行过滤, 得到符合 查询条件的视频段的索引;
将所述主干视频按照与所述特征查询信息匹配度排序得到主干视 频列表,以便于用户对所述主干视频列表中的视频进行选择播放;或者, 对符合查询条件的视频段的索引根据匹配度排序, 形成视频段的索引列表, 以便于用户对索引列表中的视频进行选择播放。
2、 根据权利要求 1所述的方法, 其特征在于, 在获取用户输入的 特征查询信息前, 所述方法还包括:
获取处于用户输入的时间段的待查视频。
3、 根据权利要求 1 所述的方法, 其特征在于, 将所述对符合查询 条件的视频段的索引根据匹配度排序的步骤具体为:把匹配度高的视频段的 索引放在前面, 把匹配度低的视频段的索引放在后面。
4、 根据权利要求 1所述的方法, 其特征在于, 在获取用户输入的 特征查询信息前, 所述方法还包括:
将获取的视频中不属于视频生成范围的视频剪切得到待查视频;或 者,将获取的视频中不属于视频生成范围的视频索引进行过滤得到待查视频 索引, 所述视频生成范围包括: 特定时间段中的地点、行为、行为主体、 行为主体的颜色中的任意一种或多种。
5、 根据权利要求 4所述的方法, 其特征在于, 所述方法还包括: 根据所述视频生成范围中的各信息, 对应视频列表中的每个视频 生成相应的视频摘要信息;
对应视频列表中的每个视频输出所述视频摘要信息;
生成所述视频摘要信息对应的视频快照图像或视频片段;
当用户触发所述视频摘要信息时, 显示所述视频快照图像或视频 片段。
6、 根据权利要求 5所述的方法, 其特征在于, 所述方法还包括: 接收用户选定的视频;
播放所述用户选定的视频;
根据所述用户选定的视频中视频的特征信息, 在所述待查视频中 属于所述视频生成范围;
输出所述关联视频列表, 以便用户对所述关联视频列表中的视频 进行选择播放。
7、 根据权利要求 1至 6任意一项权利要求所述的方法, 其特征在 于, 所述待查视频为录像, 在所述获取用户输入的特征查询信息前, 所述方法还包括:
获取用户在电子地图上选择的区域范围;
获取所述区域范围内的录像作为所述待查视频。
8、 根据权利要求 7所述的方法, 其特征在于, 当所述待查视频中 的行为主体为车辆时, 所述方法还包括:
在智能分析系统和 /或地理信息系统 GIS和 /或车辆信息库中查询所 述待查视频中的车辆的相关信息;
所述视频摘要包括所述车辆的相关信息。
9、 根据权利要求 8所述的方法, 其特征在于, 当所述特征查询信 息与所述车辆的相关信息中对相同的特征存在不同的特征描述时, 输 出所述不同的特征描述。
10、 根据权利要求 9所述的方法, 其特征在于, 所述方法还包括: 获取所述待查视频对应摄像装置位置信息;
根据所述摄像装置位置信息与所述待查视频中车辆相对于所述摄 像装置移动方向, 确定所述车辆行驶方向;
所述视频摘要包括所述车辆行驶方向。
11、 一种视频查询装置, 其特征在于, 包括:
特征获取单元, 用于获取用户输入的特征查询信息;
处理单元, 用于将待查视频中不包括所述特征查询信息的无关视 频剪切, 得到主干视频; 或者, 根据特征查询信息对待查视频的索引进行 过滤, 得到符合查询条件的视频段的索引;
排序单元, 用于将所述主干视频按照与所述特征查询信息匹配度 排序得到主干视频列表, 以便于用户对所述主干视频列表中的视频进 行选择播放, 或者, 对符合查询条件的视频段的索引根据匹配度排序, 把 匹配度高的视频段的索引放在前面, 把匹配度低的视频段的索引放在后面, 形成视频段的索引列表, 以便于用户对索引列表中的视频进行选择播放。
12、 根据权利要求 11所述的装置, 其特征在于, 所述排序单元还 用于将所述无关视频按照与所述特征查询信息匹配度排序得到无关视 频列表;
视频输出单元, 用于输出所述主干视频列表与所述无关视频列表, 择播放。
13、 根据权利要求 11所述的装置, 其特征在于, 所述处理单元还 用于将获取的视频中不属于视频生成范围的视频剪切得到待查视频, 所述视频生成范围包括: 特定时间段中的地点、 行为、 行为主体、 行 为主体的颜色中的任意一种或多种。
14、 根据权利要求 13所述的装置, 其特征在于, 所述装置还包括: 摘要生成单元, 用于根据所述视频生成范围中的各信息, 对应视 频列表中的每个视频生成相应的视频摘要信息;
摘要输出单元, 用于对应视频列表中的每个视频输出所述视频摘 要信息;
生成单元, 用于生成所述视频摘要信息对应的视频快照图像或视 频片段;
显示单元, 用于在用户触发所述视频摘要信息时, 显示所述视频 快照图像或视频片段。
15、 根据权利要求 11 至 14任意一项权利要求所述的装置, 其特 征在于, 所述待查视频为录像, 所述装置还包括:
范围获取单元, 用于获取用户在电子地图上选择的区域范围; 视频获取单元, 用于获取所述区域范围内的录像作为待查视频。
16、 根据权利要求 15所述的装置, 其特征在于, 当所述待查视频 中的行为主体为车辆时, 所述装置还包括:
查询单元, 用于在智能分析系统和 /或地理信息系统 GIS和 /或车辆 信息库中查询所述待查视频中的车辆的相关信息; 差异输出单元, 用于在所述特征查询信息与所述车辆的相关信息 中对相同的特征存在不同的特征描述时, 输出所述不同的特征描述; 信息获取单元, 用于获取所述待查视频对应摄像装置位置信息; 确定单元, 用于根据所述摄像装置位置信息与所述待查视频中车 辆相对于所述摄像装置移动方向, 确定所述车辆行驶方向;
所述摘要输出单元输出的所述视频摘要包括所述车辆的相关信 息、 所述车辆行驶方向。
17、 一种视频查询系统, 其特征在于, 包括: 权利要求 11 至 16任 意一项权利要求所述的视频查询装置。
PCT/CN2013/073289 2012-03-27 2013-03-27 一种视频查询方法、装置与系统 Ceased WO2013143465A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP13768526.9A EP2735984A4 (en) 2012-03-27 2013-03-27 METHOD, DEVICE AND SYSTEM FOR VIDEO REQUEST

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201210084264.0 2012-03-27
CN2012100842640A CN103365848A (zh) 2012-03-27 2012-03-27 一种视频查询方法、装置与系统

Publications (1)

Publication Number Publication Date
WO2013143465A1 true WO2013143465A1 (zh) 2013-10-03

Family

ID=49258227

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2013/073289 Ceased WO2013143465A1 (zh) 2012-03-27 2013-03-27 一种视频查询方法、装置与系统

Country Status (3)

Country Link
EP (1) EP2735984A4 (zh)
CN (1) CN103365848A (zh)
WO (1) WO2013143465A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105845020A (zh) * 2016-05-20 2016-08-10 深圳市喜悦智慧数据有限公司 一种实景地图制作方法及装置
EP3091735A4 (en) * 2014-01-03 2016-11-16 Hangzhou Hikvision Digital Tec METHOD AND DEVICE FOR EXTRACTION OF MONITOR RECORDING VIDEOS
CN111866428A (zh) * 2019-04-29 2020-10-30 杭州海康威视数字技术股份有限公司 一种历史录像数据处理方法及装置
CN114265952A (zh) * 2022-03-02 2022-04-01 浙江宇视科技有限公司 一种目标检索方法和装置

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104935872A (zh) * 2014-03-19 2015-09-23 华为技术有限公司 视频质检系统及方法
US9418153B2 (en) * 2014-07-29 2016-08-16 Honeywell International Inc. Video search and playback interface for vehicle monitor
CN104519319A (zh) * 2014-12-11 2015-04-15 北京中星微电子有限公司 用于电子地图的监控视频显示方法和装置
CN109871815B (zh) * 2019-02-26 2021-07-20 新华三技术有限公司 一种查询监控信息的方法及装置
CN111460218A (zh) * 2020-03-31 2020-07-28 联想(北京)有限公司 一种信息处理方法及装置
CN111641805B (zh) * 2020-04-30 2021-10-01 武汉旷视金智科技有限公司 获取视频的方法、装置、终端设备和服务器
CN114827654B (zh) * 2021-01-19 2024-05-03 腾讯科技(北京)有限公司 一种视频搜索方法、装置、设备及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1461142A (zh) * 2003-06-30 2003-12-10 北京大学计算机科学技术研究所 一种基于内容的视频片段检索方法
CN101778260A (zh) * 2009-12-29 2010-07-14 公安部第三研究所 一种基于结构化描述的视频监控管理方法及系统
CN101840422A (zh) * 2010-04-09 2010-09-22 江苏东大金智建筑智能化系统工程有限公司 基于目标特征和报警行为的智能视频检索系统和方法
CN102033937A (zh) * 2010-12-20 2011-04-27 百度在线网络技术(北京)有限公司 视频搜索结果的展示方法及其系统

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070244902A1 (en) * 2006-04-17 2007-10-18 Microsoft Corporation Internet search-based television
US7382244B1 (en) * 2007-10-04 2008-06-03 Kd Secure Video surveillance, storage, and alerting system having network management, hierarchical data storage, video tip processing, and vehicle plate analysis
CN101630361A (zh) * 2008-12-30 2010-01-20 北京邮电大学 一种基于车牌、车身颜色和车标识别的套牌车辆识别设备及方法
US20110099488A1 (en) * 2009-10-26 2011-04-28 Verizon Patent And Licensing Inc. Method and apparatus for presenting video assets
US10424342B2 (en) * 2010-07-28 2019-09-24 International Business Machines Corporation Facilitating people search in video surveillance

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1461142A (zh) * 2003-06-30 2003-12-10 北京大学计算机科学技术研究所 一种基于内容的视频片段检索方法
CN101778260A (zh) * 2009-12-29 2010-07-14 公安部第三研究所 一种基于结构化描述的视频监控管理方法及系统
CN101840422A (zh) * 2010-04-09 2010-09-22 江苏东大金智建筑智能化系统工程有限公司 基于目标特征和报警行为的智能视频检索系统和方法
CN102033937A (zh) * 2010-12-20 2011-04-27 百度在线网络技术(北京)有限公司 视频搜索结果的展示方法及其系统

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3091735A4 (en) * 2014-01-03 2016-11-16 Hangzhou Hikvision Digital Tec METHOD AND DEVICE FOR EXTRACTION OF MONITOR RECORDING VIDEOS
CN105845020A (zh) * 2016-05-20 2016-08-10 深圳市喜悦智慧数据有限公司 一种实景地图制作方法及装置
CN111866428A (zh) * 2019-04-29 2020-10-30 杭州海康威视数字技术股份有限公司 一种历史录像数据处理方法及装置
CN114265952A (zh) * 2022-03-02 2022-04-01 浙江宇视科技有限公司 一种目标检索方法和装置

Also Published As

Publication number Publication date
EP2735984A4 (en) 2016-04-20
EP2735984A1 (en) 2014-05-28
CN103365848A (zh) 2013-10-23

Similar Documents

Publication Publication Date Title
WO2013143465A1 (zh) 一种视频查询方法、装置与系统
CN103984710B (zh) 基于海量数据的视频交互查询方法及系统
Zheng et al. Diagnosing New York city's noises with ubiquitous data
EP3253042B1 (en) Intelligent processing method and system for video data
CN105701123B (zh) 人车关系的识别方法及装置
Singh et al. Social pixels: genesis and evaluation
CN107766377B (zh) 一种监控数据查询方法及装置
US11856331B1 (en) Extracting and transmitting video analysis metadata for a remote database
Morishita et al. SakuraSensor: Quasi-realtime cherry-lined roads detection through participatory video sensing by cars
CN102880692A (zh) 一种面向检索的监控视频语义描述和检测建模方法
CN109033440A (zh) 一种视频侦查多维轨迹分析方法
US20100042615A1 (en) Systems and methods for aggregating content on a user-content driven website
CN106384512B (zh) 车辆违章信息查询方法以及系统
CN111767432B (zh) 共现对象的查找方法和装置
CN116662626B (zh) 汽车数据处理方法、装置、电子设备及存储介质
CN105138525A (zh) 交通视频处理装置和方法及检索装置和方法
Tempelmeier et al. Data4urbanmobility: Towards holistic data analytics for mobility applications in urban regions
Jung et al. Real-time car tracking system based on surveillance videos
CN103631871B (zh) 一维地图的实现方法
Medvedev et al. Reporting road problems in smart cities using OpenIoT framework
CN108257406A (zh) 终端、服务器和驾驶数据分享方法
CN114265952B (zh) 一种目标检索方法和装置
Aoki et al. Busbeat: Early event detection with real-time bus GPS trajectories
CN106682168A (zh) 一种跨域城市数据可视查询系统的构建方法
Chiang et al. Quick browsing and retrieval for surveillance videos

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 13768526

Country of ref document: EP

Kind code of ref document: A1

REEP Request for entry into the european phase

Ref document number: 2013768526

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 2013768526

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE