CN102665064A - A traffic video monitoring system based on standard labeling and quick search - Google Patents

A traffic video monitoring system based on standard labeling and quick search Download PDF

Info

Publication number
CN102665064A
CN102665064A CN 201210056022 CN201210056022A CN102665064A CN 102665064 A CN102665064 A CN 102665064A CN 201210056022 CN201210056022 CN 201210056022 CN 201210056022 A CN201210056022 A CN 201210056022A CN 102665064 A CN102665064 A CN 102665064A
Authority
CN
Grant status
Application
Patent type
Prior art keywords
traffic
video
object
module
monitoring
Prior art date
Application number
CN 201210056022
Other languages
Chinese (zh)
Inventor
万忠
刘云鹏
奚李峰
张三元
张引
李瑾
毕春跃
王仁芳
Original Assignee
浙江大学
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

Links

Abstract

The present invention discloses a traffic video monitoring system based on standard labeling and quick search, comprising a traffic monitoring video capturing module, a video image content analyzing module, a traffic video key frame obtaining module, a monitoring expansion data encoding module and a data stream searching module. Based on an SVAC standard, the grammar and semantics of the monitor messages can be appropriately expanded to satisfy the requirements of quick search of monitoring videos. The system provides a unified standard interface to realize labeling and quick search of compressed monitoring video data, and the monitor messages can be appropriately stretched and clipped according to business requirement to meet demand of functions such as data transmission. Meanwhile, an encoding scheme targeting traffic objects and events is worked out and the SVAC standard is used to realize a highly-accurate, simple, effective, unified and fast labeling and quick search method of traffic objects and events.

Description

一种基于标准标记与快速检索的交通视频监控系统技术领域 Traffic video surveillance system technology fields marked with a standards-based fast retrieval

[0001] 本发明属于视频监控技术领域,涉及一种交通视频监控系统,具体地说是一种基于标准标记与快速检索的交通视频监控系统。 [0001] The present invention belongs to the technical field of video surveillance, traffic relates to a video surveillance system, in particular to a transport standard video surveillance system based on fast retrieval numerals.

背景技术 Background technique

[0002] SVAC :英文Technical Specification of Surveillance Video and Audio Coding缩写,对应中文为《安全防范监控数字视音频编解码技术要求》,由公安部第一研究所和全国安全防范报警系统标准化技术委员会(SAC/TC100)经过梳理分析,整理出安全防范监控视音频编解码和广电媒体视音频编解码的主要异同点,明确了安全防范监控数字视音频编解码特殊需求,经过反复讨论修正后,最终成为国家标准,并于2011年5月1日正式实施。 [0002] SVAC: English Technical Specification of Surveillance Video and Audio Coding abbreviation, the corresponding Chinese as "security surveillance digital video and audio codec technology requirements," Standardization Technical Committee by the First Research Institute of Ministry of Public Security Alarm System and the national security (SAC / TC100) carded analysis, sorting out the major similarities and differences between security monitoring video and audio codec and broadcast media video and audio codec, a clear security surveillance digital video and audio codec special needs, after repeated discussions amended, eventually becoming the country standards, and on May 1, 2011 formally implemented.

[0003] 伴随着“平安城市”、“智慧城市”的大力建设,道路与交通监控成为其主要技术手段之一。 [0003] With the "Green City", "smart city" strong construction, road and traffic monitoring has become one of its main technical means. 据有关数据显示,到2010年,很多大型城市已经安装完毕20多万个监控摄像头,这20多万个摄像头大多将遍布城市的道路、桥梁和公共交通系统。 According to statistics, by 2010, many large cities have installed more than 200,000 surveillance cameras, most of which more than 20 million cameras will be all over the city's roads, bridges and public transit systems. 由于业务的需要,大部分监控视频需要压缩编码后进行存档,采用的技术可以是国际压缩标准,比如H. 264,MPEG-4 等,也可以是公司或个人的私有算法,比如海康威视公司针对监控视频的压缩算法等,不管如何压缩,都会带来海量的压缩后的交通监控视频数据。 Due to business needs, most of the surveillance video archive after compression coding technology can be used in international compression standards, such as H. 264, MPEG-4, etc., can also be a private company or individual algorithms, such as Hikvision the company monitors for video compression algorithms, etc., regardless of compression, will bring traffic monitoring vast amounts of video data compression. 如何对感兴趣的监控内容进行标识和记录,以及在查询取证时能够使用统一标准的接口从海量数据中高效、快速地查询出所需信息是当前视频监控领域的重要问题。 How to monitor the content of interest to be identified and recorded, and the ability to use a uniform standard of evidence when a query interface to efficiently and quickly check out the required information from massive amounts of data is an important issue in the current field of video surveillance.

[0004] 为了方便后期的视频检索,往往需要在视频预处理或视频分析阶段对视频内容进行相应的标记,对于标记的方法,有基于标准和基于非标准的两种方案。 [0004] In order to facilitate the later retrieval, often require the appropriate markers or video content in a video pre-processing video analysis phase, a method for marking, and there are two schemes based on non-standard criteria. 对于非标准的标记方案,是利用元数据对发生地点、录制时间、图像基本信息等一些简单的描述性数据按照某种格式存储到数据库或文件中;对于标准的标记方案,大部分是利用基于MPEG-7的多媒体描述方案,目标就是产生一种描述多媒体内容数据的标准,可以对各种不同类型的多媒体信息进行标准化描述,并将该描述与所描述的内容相联系,以实现快速有效的检索。 For non-standard labeling scheme, using simple metadata descriptive data location occurs, the recording time, the image stored in the basic information and the like in accordance with a certain format in the database or file; for standard labeling scheme, mostly based on the use of MPEG-7 multimedia description schemes, the goal is to produce a standard multimedia content description data, can be a variety of different types of multimedia information in a standardized description, and the description in connection with what has been described, in order to achieve rapid and efficient retrieval.

[0005] 常见的视频检索方式有三种。 [0005] common video retrieval in three ways. 1)利用前面所提的已经存储的标记和索引信息直接对视频内容进行检索,2)将需要处理的视频序列先进行解压缩,恢复到空间像素域,再进行图像和视频分析,看内容是否满足检索的条件,3)基本上不进行解压缩处理,直接利用压缩域数据进行分析,通过在压缩域获取视频内容的特征数据来看是否满足检索条件。 1) using the marker and the index information has been stored previously mentioned direct retrieval of the video content, the video sequence 2) would need to be processed to be decompressed, to restore the spatial pixel domain, and then the video image analysis, to see whether the contents retrieval condition is satisfied, 3) substantially no decompression processing, compressed domain using direct analysis of the data acquired by the video content data, wherein the compressed domain satisfies the search condition. 但是现有技术中都存在不同程度的缺陷。 However, the prior art there are varying degrees of impairment.

[0006] 基于元数据标记的检索方式,利用简单的文本,主观的进行描述,精确性不高,同时对于描述视频内容的信息量非常少,又难以反映视频图像本身的多样性,往往要配合大量的人工检索进行辅助,对于要求一定精确度的海量数据的快速检索是不现实的。 [0006] Based on the way to retrieve metadata tags, using simple text, subjective description, accuracy is not high, but for the amount of information describing the video content is very small, and difficult to reflect the diversity of the video image itself, often with a lot of manual retrieval assist, quick search requires a certain accuracy for massive data is unrealistic.

[0007] 基于MPEG-7标准的多媒体描述方案下的检索方式,一方面技术方案相对复杂,往往需要自然语言处理技术、搜索引擎技术和分布式系统技术,另一方面并非针对监控视频, 更不会针对交通监控视频,没有充分考虑和利用此类视频的特征,同时复杂的标记与索引信息和通过编码标准压缩的视频数据是独立分开的,也没有针对监控视频的统一接口,所以不利于监控视频的标准化快速检索和对监控信息的传输操作。 [0007] Based on the retrieval mode of multimedia standard MPEG-7 description scheme, one aspect is relatively complex, often require natural language processing technology, distributed search engine technology and system technology, on the other hand the video monitor is not directed, but do not will be for traffic surveillance video did not fully consider and utilize features such video, as well as complex and marked with the index information by encoding standard compressed video data is separate and distinct, there is no uniform interface for video surveillance, it is not conducive to monitoring standardization fast retrieval and transfer operations to monitor video information.

[0008] 基于内容的图像检索,包括针对像素域和压缩域数据两种,它通过分析和理解多媒体信息的视觉信息,根据得到的多媒体低级特征的匹配来进行检索。 [0008] The content-based image retrieval, comprising a pixel domain for the two kinds of data and the compressed domain, it is to be appreciated and retrieve multimedia information by analyzing visual information according to a matching of the obtained low-level features of multimedia. 对于纯粹的像素域数据分析,首先需要把压缩的视频数据全部解码,然后在像素域进行视频分析,对于海量的监控数据,不管是在空间上还是时间上来讲,都是不现实的。 For pure pixel domain data analysis, first of all we need to decode compressed video data and video pixel domain analysis, monitoring data for the mass, whether in space or in time terms, is unrealistic. 对于压缩域数据分析,只存在很少量的图像解码过程,可以忽略,但是由于分析过程复杂,需要消耗大量的时间,而且和像素域分析相比,准确度偏低,同时也没有针对监控视频的统一标准接口。 For compressed domain data analysis, there is only a small amount of image decoding process can be ignored, but due to the complexity of the analysis process, consumes a lot of time and pixel-domain analysis and comparison, the accuracy is low, but there is no video for monitoring the unified standard interface.

发明内容 SUMMARY

[0009] 为了克服现有技术中存在的以上缺陷,本发明提供一种交通视频监控系统,基于SVAC标准,并对其中监控信息部分的语法和语义进行适当的扩展,以满足监控视频快速检索的需要。 [0009] In order to overcome the above drawbacks present in the prior art, the present invention provides a video surveillance system traffic, based SVAC standard, and wherein the syntax and semantics of the monitoring information portion proper scale to meet the rapid retrieval of the video monitor need. 使用SVAC标准,完全是针对视频监控网络,可以解决目前视频监控系统中视音频编解码标准不统一导致的系统难以互联互通的问题,这种互通性的解决,提供了一个统一标准的接口来实现压缩监控视频数据的标记和快速检索,并可以根据业务需要对监控信息进行适当的伸缩和裁剪,来满足传输等功能的需要,同时制定出一套针对交通对象和事件的编码方案,最终利用SVAC标准来实现一种准确率高、简单有效和统一快速的交通对象和交通事件的标记和快速检索方法,技术方案为: Use SVAC standard, entirely for video surveillance network, you can solve the video and audio codec standard video surveillance system does not lead to a unified system of interconnection of difficult issues, such interoperability solutions, provides a unified standard interface to achieve compression the surveillance video data tags and fast retrieval, and can be based on business needs for monitoring information appropriate scaling and cropping, to meet the needs of transmission and other functions, as well as develop a set of traffic object and coding scheme for the event, the final use of SVAC standard to achieve a high accuracy, simple and effective and rapid unification of the object mark traffic and traffic incidents and rapid retrieval methods, technical solutions for:

[0010] 一种基于标准标记与快速检索的交通视频监控系统,包括交通监控视频捕捉模块和视频图像内容分析模块,还包括获取交通视频关键帧模块、监控扩展数据编码模块和对码流检索模块,其中: [0010] A standard video surveillance system marker and Transportation rapid retrieval, including traffic monitoring video image capture module and video content analysis module, further comprising a video key frame based on the acquired traffic module, monitor module and the extended data encoding code stream retrieval module ,among them:

[0011] 视频图像内容分析模块的作用为从关键帧中提取出可分类的交通对象和交通事件; [0011] the role of video content analysis module is extracted from the key frames and objects traffic traffic events can be categorized;

[0012] 获取交通视频关键帧模块的作用为从运动镜头中将存在事件的图像设置为关键帧,传递交通事件和相关对象; [0012] GET action video key frame module there is provided an image in the event of movement of the lens is from a key frame, and transmitting the traffic event related objects;

[0013] 监控扩展数据编码模块的作用为监控扩展数据进行扩展,来满足对交通对象的语法表达,监控扩展数据单元通过extension_id进行区分,取extension_id值为0x6来表示监控对象扩展语法,从extension_id开始到reserve_bits的语义与其他监控扩展数据单元的相同语法元素的语义是一致的,循环语法的语义为从每个区域中找到每个对象,并且从region_object_id中获取到对象的详细特征信息,语法元素值的含义由应用本身决定, 对交通对象的region_object_id进行定义,通过扩展后的SVAC标准内容对交通对象和交通事件进行编码; Acts [0013] The extension data encoding monitoring module to monitor the extension data extended to meet the grammatical object of traffic monitoring extension data unit distinguished by EXTENSION_ID, taking a value of 0x6 EXTENSION_ID extended syntax to indicate the monitored object, beginning from EXTENSION_ID identical to the semantics of the syntax elements other monitoring unit reserve_bits extension data is consistent, the semantics of the syntax for the cycle to find each object from each area, and acquires the detailed characteristic information to region_object_id the object, the value of the syntax element the meaning is determined by the application itself, for region_object_id traffic object definition, encodes the target traffic and traffic incident by SVAC standard content after expansion;

[0014] 对码流检索模块的作用为根据已知的SVAC码流和交通对象或交通事件的特征编码,从码流的监控扩展数据单元中找到交通对象或交通事件所在的图像位置以及所在图像中区域的位置,记录下图像位置所对应的区域具体参数,从编码的视频图像数据中找到图像数据并解码显示,同时标记出相应的感兴趣区域。 [0014] The code stream according to known SVAC traffic object or feature encoding and traffic event, the extension data from the monitoring unit to find the code stream in the image position of the vehicle or the object is located and the traffic event is located on the image retrieval module role stream region position, region specific parameters corresponding to a position of the image recorded, the image data is found from the encoded video image data and decoded for display, and mark the corresponding region of interest.

[0015] 本发明也可以通过以下方式实现,所述关键帧信息直接来自视频图像内容分析模块,同时包含了检测出的交通对象或交通事件,对于事件,SVAC有针对事件的监控事件扩展语法结构:event_extension,其中的region_event_id 表不事件特征,将region_event_id针对交通视频业务进行编码标准定义。 [0015] The present invention may also be achieved by the keyframe information directly from the video content analysis module, it contains the traffic object or detected traffic event, for event monitoring SVAC has extended for a syntax structure Event : event_extension, which region_event_id table does not feature event, the region_event_id encode standard definition video for the transportation business.

[0016] 进一步优选,在找关键帧之前,先将运动镜头按照其中间的车辆复杂度进行分类。 [0016] Further preferably, before looking for key frames, first lens classified according to vehicle motion complexity between them.

[0017] 与现有技术相比,本发明的有益效果: [0017] Compared with the prior art, the beneficial effects of the invention:

[0018] (1)检索速度非常快:由于只有文件流查找和解码关键帧的操作,没有大量视频图像的解码过程和低层编码视频数据的分析操作。 [0018] (1) very fast retrieval: Since only the file stream operation to find and decode a key frame, the decoding process does not analyze a large number of operations and low-level video image encoded video data.

[0019] (2)检索准确率高:由于对象与事件的分析和获取是在视频监控捕获阶段,而用于实时监控视频的质量与存档视频质量相比是高分辨率和高清晰度的,这就大大提高了视频内容分析的准确度。 [0019] (2) retrieval accuracy rate: Since the analysis and acquisition of objects and events in video surveillance capture stage, and the quality and archived video quality for real-time monitoring of video compared to a high-resolution and high definition, This greatly improves the accuracy of video content analysis.

[0020] (3)检索标准的高度统一:完全遵循SVAC编码标准,同时也对交通对象和事件进行了统一的编码,这就可以对任何不同地区的遵循SVAC标准的压缩交通视频进行统一的检索处理。 [0020] (3) a high degree of unity search criteria: full compliance SVAC coding standards, but also on traffic objects and events were unified coding, which can be unified retrieval of any traffic video compression standard to follow SVAC different regions deal with.

附图说明 BRIEF DESCRIPTION

[0021] 图1是本发明一种基于标准标记与快速检索的交通视频监控系统示意图; [0021] FIG. 1 is a schematic view of the present invention, a standard marker and Transportation video surveillance system based fast retrieval;

[0022] 图2是运动镜头示意图; [0022] FIG. 2 is a schematic view of movement of the lens;

[0023] 图3是关键帧位置图; [0023] FIG. FIG. 3 is a key frame position;

[0024] 图4是根据交通事件或交通对象特征编码查询相关的图像与图像中区域标识流程图; [0024] FIG. 4 is a traffic event or a traffic object feature encoding the relevant image in the image region identifier flowchart;

[0025] 图5是查找图像中区域详细信息的流程图; [0025] FIG. 5 is a flowchart for more information in the image region;

[0026] 图6是检索出图像并显示流程图。 [0026] FIG. 6 is a flowchart showing an image and retrieved.

具体实施例 Specific Example

[0027] 下面结合附图和本发明实施例作进一步详细地说明。 [0027] accompanying drawings and embodiments of the invention will be further described in detail below in conjunction.

[0028] 参照图1,一种基于标准标记与快速检索的交通视频监控系统,包括交通监控视频捕捉模块和视频图像内容分析模块,还包括获取交通视频关键帧模块、监控扩展数据编码模块和对码流检索模块,其中: [0028] Referring to FIG 1, a standard video surveillance system marker and Transportation rapid retrieval, including traffic monitoring video image capture module and video content analysis module, further comprising obtaining traffic video key frame module, monitor module and the extended data encoding based on stream retrieval module, wherein:

[0029] 视频图像内容分析模块的作用为从关键帧中提取出可分类的交通对象和交通事件; [0029] the role of video content analysis module is extracted from the key frames and objects traffic traffic events can be categorized;

[0030] 获取交通视频关键帧模块的作用为从运动镜头中将存在事件的图像设置为关键帧,传递交通事件和相关对象; [0030] GET action video key frame module there is provided an image in the event of movement of the lens is from a key frame, and transmitting the traffic event related objects;

[0031] 监控扩展数据编码模块的作用为监控扩展数据进行扩展,来满足对交通对象的语法表达,监控扩展数据单元通过extension_id进行区分,取extension_id值为0x6来表示监控对象扩展语法,从extension_id开始到reserve_bits的语义与其他监控扩展数据单元的相同语法元素的语义是一致的,循环语法的语义为从每个区域中找到每个对象,并且从region_object_id中获取到对象的详细特征信息,语法元素值的含义由应用本身决定, 对交通对象的region_object_id进行定义,通过扩展后的SVAC标准内容对交通对象和交通事件进行编码; Acts [0031] The extension data encoding monitoring module to monitor the extension data extended to meet the grammatical object of traffic monitoring extension data unit distinguished by EXTENSION_ID, taking a value of 0x6 EXTENSION_ID extended syntax to indicate the monitored object, beginning from EXTENSION_ID identical to the semantics of the syntax elements other monitoring unit reserve_bits extension data is consistent, the semantics of the syntax for the cycle to find each object from each area, and acquires the detailed characteristic information to region_object_id the object, the value of the syntax element the meaning is determined by the application itself, for region_object_id traffic object definition, encodes the target traffic and traffic incident by SVAC standard content after expansion;

[0032] 对码流检索模块的作用为根据已知的SVAC码流和交通对象或交通事件的特征编码,从码流的监控扩展数据单元中找到交通对象或交通事件所在的图像位置以及所在图像中区域的位置,记录下图像位置所对应的区域具体参数,从编码的视频图像数据中找到图像数据并解码显示,同时标记出相应的感兴趣区域。 [0032] The code stream according to known SVAC traffic object or feature encoding and traffic event, the extension data from the monitoring unit to find the code stream in the image position of the vehicle or the object is located and the traffic event is located on the image retrieval module role stream region position, region specific parameters corresponding to a position of the image recorded, the image data is found from the encoded video image data and decoded for display, and mark the corresponding region of interest.

[0033] 本发明也可以通过以下方式实现,所述关键帧信息直接来自视频图像内容分析模块,同时包含了检测出的交通对象或交通事件,对于事件,SVAC有针对事件的监控事件扩展语法结构:event_extension,其中的region_event_id 表不事件特征,将region_event_id 针对交通视频业务进行编码标准定义。 [0033] The present invention may also be achieved by the keyframe information directly from the video content analysis module, it contains the traffic object or detected traffic event, for event monitoring SVAC has extended for a syntax structure Event : event_extension, which region_event_id table does not feature event, the region_event_id encode standard definition video for the transportation business.

[0034] 进一步优选,在找关键帧之前,先将运动镜头按照其中间的车辆复杂度进行分类。 [0034] Further preferably, before looking for key frames, first lens classified according to vehicle motion complexity between them.

[0035] 从图1中可以知道,本系统包含了交通监控视频捕捉、视频图像内容分析、获取交通视频关键帧、按照SVAC标准编码视频和对SVAC压缩码流进行检索五大功能。 [0035] can be known from FIG. 1, the traffic monitoring system includes a video capture, video content analysis, GET key frame video, and video coding standard in accordance SVAC SVAC compressed stream to retrieve five functions. 其中,交通监控视频捕捉和视频图像内容分析作为已知的功能模块来提供必要的接口数据,而获取交通视频关键帧、SVAC编码中的监控扩展数据编码部分和对码流检索部分的实现是本发明的核心技术环节,下面将做详细的介绍。 Wherein, the traffic monitoring video image capture and video content analysis as known in the function module to provide the necessary interface data, acquired traffic key frame video monitor extension data encoding portion SVAC coding retrieval section and effected stream are the core technical aspects of the invention, will be described in detail below.

[0036] 在监控视频中,人们最关心的两类重要信息是对象和事件,显然对于交通监控视频,就是交通对象和交通事件,人们在对监控视频进行检索的时候一般也是按照对象和事件进行查找,比如检索条件是“一辆白色的桑塔纳”或“两车相撞事件”。 [0036] In the surveillance video, the two types of important information people are most concerned about objects and events, it is clear for traffic surveillance video, the object is to transport and traffic incident, people are generally carried out in accordance with the objects and events at the time of retrieval of surveillance video Find, such as search query is "a white Santana" or "two-car collision." 所以一个重要的工作就是找到关心的对象和事件,再遵循SVAC标准进行编码。 Therefore, an important job is to find objects and events of interest, and then follow the SVAC standard coding. 我们先说事件,一般来说事件并不存在于捕获的每一个视频帧中,而且在整个监控的过程当中,数量也是有限的,它是由视频图像内容分析模块在实时监控的时候所分析产生的结果,当分析到已经定义的交通事件后,会把产生事件的视频帧、事件类型和事件相关的对象交给SVAC编码模块进行处理, 具体处理过程稍后部分会有详细介绍。 Let's say an event, the event does not exist in general, each video frame is captured, but also in the whole monitoring process, the number is limited, it is by video content analysis module in real-time monitoring when analyzed produced the result, when analyzing the traffic incident has been defined, will produce video frame event, event type objects and events related to the SVAC encoding module for processing, the specific part of the process will be described in detail later. 现在我们接着说对象,对象不同于事件,它可能存在于捕获的每一个视频帧中,如果我们对每一帧图像都进行对象的分析和获取,这会给图像内容分析模块带来巨大的工作量,而且无法保证处理的实时性,所以我们要利用监控视频和道路交通的特点,找出需要获取对象的视频帧(我们称之为关键帧)进行对象的分析和获取,与所有的视频帧相比,关键帧的数量就少了很多,不管是在编码前的对象分析阶段还是编码后的检索阶段,这都会带来了很大的速度改进。 Now we went on an object, the object is different from the event, which may be present in each video frame capture, if we were to each frame of image analysis and acquisition targets, this will give the image content analysis module tremendous work amount, and we can not ensure real-time processing, so we have to take advantage of the characteristics of video surveillance and road traffic, find a video frame (we called keyframes) need to get the object of analysis and acquisition targets, with all the video frames compared to the number of key frames of a lot less, whether it is in the object code analysis phase before or after the retrieval phase encoding, which it will bring a lot of speed improvements. 下面就详细描述获取交通视频关键帧的技术实现方案。 Here are described in detail for technical traffic video key frame of implementation.

[0037] 获取交通视频关键帧:对于交通监控视频来讲,在某个很长的时间内,背景都是固定不变的,此处我们可以认为背景就是固定的。 [0037] acquiring key frames of video traffic: for traffic surveillance video is concerned, within a very long time, the background is fixed, here we can say that the background is fixed. 而且,交通监控的另一个特点就是经常会有一段时间内没有任何交通对象出现,只有静止的背景显示。 Moreover, traffic monitoring Another feature is that there is often a period of time without any traffic objects appear, only stationary background. 此处,我们做以下定义,将连续两次静止背景视频帧之间存在运动交通对象的所有视频帧称之为一个运动镜头,如图2所 Here, we make the following definitions, the background of all consecutive video frames still exist traffic object motion between video frames of a motion referred to as lens, as shown in FIG 2

示。 Shows. [0038] 从图2可以看到,第n帧和第n+61帧是没有任何运动对象的静态背景帧,第n+1 帧到第n+60帧存在运动对象,这连续的60帧就是一个运动镜头。 [0038] As seen in Figure 2, the n-th frame and frame n + 61 is no moving object in a static background frame, the first frame to the n + 1 n + 60 there is a moving object, which is a continuous 60 a movement of the lens. 我们需要做的工作就是从运动镜头中找到最重要的一帧图像作为关键帧,在找关键帧之前,我们先将运动镜头按照其间的车辆复杂度进行分类,1)简单运动镜头:在整个镜头中,只有一辆车在开始帧位置进入,在结束帧位置淡出,此时取车头或车尾最近位置的帧为关键帧,因为车辆最核心的信息是车牌号,在最近位置的时候是车牌号最清晰的时候,同时也可以看到整个车身信息。 We need to do is to find the most important movement of the lens of an image as a key frame, before looking keyframes, we first movement of the lens are classified according to the complexity of the vehicle between them, 1) simple movement of the lens: the entire lens only a car into the frame at the start position, at the end of the fade-out frame position, this time taking the front or rear frame position of the nearest keyframe, because the core of the information about the vehicle license plate number, when the most recent position of license plate No. when the clearest, but can also see the entire body of information. 如图3所示。 As shown in Figure 3.

[0039] 如果有多辆车,要求多辆车进出镜头的时间基本保持一致,此处再分两种情况,一Ir] , Ut fà, Ir] fàÆBiËAJP^.Æ—I^MWÏbMBi*] (^3-4#,«$i^lk#nT«) ±mk&mk,vim^ 2-3 ^f!^, fMÄMöliiIMtWiifê, 1*1 «[0040]SVAC um[0041] S WM SVAC ÂMfêrMfôgtfei^ m SVAC:ë?i® SVAC «5CM 5. 2. 3. 8jtmimïï, w^®gfetr Mm, ®j$$#riiM, leurif .limité, Bf M^Mfêi^èif M Mm^mmarpfix ■.[0042]roi_extension(){ extension_id extension_length position_idc camera_idc region_num reserve_bitsfor(i=0; i<region_num; i++){ region_top_left_mbx[i][0043]region_top_left_mby[i] region_width_in_mbs_rainusl[i]region_height_in_mbs_minusl[i]}}[0044] ffljlfeît 5. 2. 4. 10. 2 ^^X^È7ÛÂ的表示交通对象的信息,所以需要对监控扩展数据进行一定的扩展,来满足对交通对象的语法表达。 [0039] If multiple vehicles, vehicles requires more time out of the lens remains substantially the same, then two cases where a Ir], Ut fà, Ir] fàÆBiËAJP ^ .Æ-I ^ MWÏbMBi *] (^ ! 3-4 #, «$ i ^ lk # nT«) ± mk & mk, vim ^ 2-3 ^ f ^, fMÄMöliiIMtWiifê, 1 * 1 «[0040] SVAC um [0041] S WM SVAC ÂMfêrMfôgtfei ^ m SVAC: ë ? i® SVAC «5CM 5. 2. 3. 8jtmimïï, w ^ ®gfetr Mm, ®j $$ # riiM, leurif .limité, Bf M ^ Mfêi ^ èif M Mm ^ mmarpfix ■. [0042] roi_extension () { extension_id extension_length position_idc camera_idc region_num reserve_bitsfor (i = 0; i <region_num; i ++) {region_top_left_mbx [i] [0043] region_top_left_mby [i] region_width_in_mbs_rainusl [i] region_height_in_mbs_minusl [i]}} [0044] ffljlfeît 5. 2. 4. 10 2 ^^ traffic information indicates the object X ^ È7ÛÂ, it is necessary to monitor the expansion of certain data expansion to meet the grammatical object of traffic. 根据标准5. 2. 4. 10. 1描述可知,监控扩展数据单元通过extensionjd进行区分,目前已经使用了0xl-0x5共五个数,我们取extensionjd值为0x6来表示监控对象扩 4. 5. 2. The standard 10.1 description that, to monitor the extension data unit distinguished by extensionjd, now using five total number 0xl-0x5, 0x6 we take the value represented extensionjd monitored object extender

展语法,其语法结构如下所示: Show syntax, the syntax structure is shown below:

[0045] [0045]

object_extension() { extension—id extension—length position_idc camera一idc region—num reserve—bits object_extension () {extension-id extension-length position_idc camera a idc region-num reserve-bits

for(i=0; i<region—num; i++) { for (i = 0; i <region-num; i ++) {

object_nura[i] object_nura [i]

for(j=0; j<object_num[i]; j++){ region一object—id[i, j] for (j = 0; j <object_num [i]; j ++) {region an object-id [i, j]

} }

} }

[0046] 对新增语法结构的语义解释如下:从extension_id开始到reserve_bits的语义与其他监控扩展数据单元的相同语法元素的语义是一致的。 [0046] The semantics of the new syntax structure is explained as follows: From the start to the semantic extension_id same syntax element with semantics of other monitoring unit reserve_bits extension data is consistent. 下面循环语法的语义为从每个区域中找到每个对象,并且从region_object_id中获取到对象的详细特征信息,这个语法元素值的含义由应用本身决定,那么在本发明中,我们对交通对象的region_object_id进行了以下定义,希望可以成为SVAC标准针对交通监控业务的一部分。 Below is a semantic loop syntax found in each region for each object, and acquires the detailed characteristic information to region_object_id the object, the meaning of this syntax element values ​​determined by the application itself, then the present invention, the object of our traffic region_object_id carried out the following definitions, hoping to become part of the SVAC standard for traffic monitoring service. 具体如表1所示。 As specifically shown in Table 1.

[0047] [0047]

Figure CN102665064AD00091

[0048] ^ 1 [0049] u(n) n #Ti^Íl^h region_ob ject_id ËltÈÎ^Ix SÄ 2+4+6+3+41 = 56bitSo[0050] [0048] ^ 1 [0049] u (n) n # Ti ^ Íl ^ h region_ob ject_id ËltÈÎ ^ Ix SÄ 2 + 4 + 6 + 3 + 41 = 56bitSo [0050]

Figure CN102665064AD00092

[0051][0052] [0051] [0052]

Figure CN102665064AD00101

[0053] 有了上面的扩展语法定义,我们在对交通监控视频采用SVAC标准压缩编码的时候,就可以准确的将关键帧的交通事件或交通对象的特征描述写入到码流中,那么在解码过程当中只要按照语法结构就可以快速的找到相应的事件或对象。 [0053] With extended syntax defined above, we adopt when SVAC standard compression coding of video traffic monitoring, can be accurately characterized by traffic or traffic event keyframe description of the object code is written to the stream, then among the decoding process as long as you can quickly find the corresponding event or object in accordance with the grammatical structure.

[0054] 在SVAC码流中检索过程:首先检索的第一步是根据已知的SVAC码流和交通对象或交通事件的特征编码(编码基本格式如上述region_event_id和region_ob ject_id的定义描述)从码流的监控扩展数据单元中找到对象或事件所在的图像位置以及所在图像中区域的位置(即存在于图像中的第几个感兴趣区域,因为图像中可能存在η个存放对象的感兴趣区域),如图4所示。 [0054] In the search process stream SVAC code: The first step is to retrieve the coded according to the characteristics known SVAC stream or traffic events and traffic object (encoding basic format as described above and region_ob ject_id region_event_id defined below) from the code extension data monitoring unit finds an object or an event stream and a position where the image region where the image (i.e., present in the image region of interest of a few, because there may be a storage target η region of interest in the image) ,As shown in Figure 4. nal_unit的语法结构参见SVAC标准5. 2. 3. I部分,语义描述参见5. 2. 4. 2部分。 Syntax structure nal_unit see SVAC standard part 5. 2. 3. I, 5. 2. semantic description see section 4.2. 监控扩展数据单元的语法结构参见SVAC标准。 Syntax structure extension data unit monitor see SVAC standard. 5. 2. 3. 8部分,语义描述参见5. 2. 4. 10部分。 5. 2. 3.8 part, semantic description see section 5. 2. 4.10. 对于本发明所扩展的object_extension的语法和语义见本文档的“基于扩展SVAC标准的监控扩展数据编码”部分。 Of this document for the present invention for the expanded object_extension syntax and semantics "extended data encoding based on the extended monitoring SVAC standard" section. 该流程会把查找到的图像位置与对应的区域位置存放到一个列表数据结构FoundPicRegionList中,利用该数据在流程“查找图像中区域详细信息”中找到可表征区域的具体参数,包括区域在图像中的左上点坐标(Left, top)以及宽高(Width, Height),如图5所示。 This process will find the position of the image region corresponding to a position of a list data structure stored in the FoundPicRegionList, using the data in the process "for more information in the image area" may be found in the parameters that characterize the specific area, including areas in the image the upper left point coordinates (Left, top), and the width and height (width, Height), as shown in FIG. 根据查找后结果对FoundPicRegionList数据进行更新,记录下图像位置所对应的区域具体参数,然后根据更新后数据从编码的视频图像数据中找到图像数据并解码显示,同时标记出相应的感兴趣区域,如图6所示。 The results FoundPicRegionList after finding the updated data, the position of the image region corresponding to the specific parameters of the record, and then to find the updated data from the encoded video image data and the image data decoded for display, while the corresponding region of interest marker, such as 6 shown in FIG.

[0055] 效率分析:整个检索过程主要以SVAC编码的文件流查找定位为主,这个查找过程是非常快的。 [0055] efficiency analysis: The entire retrieval process mainly SVAC find files encoded stream based positioning, the search process is very fast. 由于对象与事件都是来自关键帧,一般我们把关键帧设置为I帧,可以直接解码,无需再解码需要参考的其他帧。 Since the object with the event are from key frames, keyframes we generally set to I-frame can be decoded directly, no longer need to refer to the decoding of other frames.

Claims (3)

  1. 1. 一种基于标准标记与快速检索的交通视频监控系统,包括交通监控视频捕捉模块和视频图像内容分析模块,其特征在于,还包括获取交通视频关键帧模块、监控扩展数据编码模块和对码流检索模块,其中: 视频图像内容分析模块的作用为从关键帧中提取出可分类的交通对象和交通事件; 获取交通视频关键帧模块的作用为从运动镜头中将存在事件的图像设置为关键帧,传递交通事件和相关对象; 监控扩展数据编码模块的作用为监控扩展数据进行扩展,来满足对交通对象的语法表达,监控扩展数据单元通过extension_id进行区分,取extension_id值为0x6来表示监控对象扩展语法,从extension_id开始到reserve_bits的语义与其他监控扩展数据单元的相同语法元素的语义是一致的,循环语法的语义为从每个区域中找到每个对象,并且从region_object_id中获取到对象的详细 A standard video surveillance system marker and Transportation rapid retrieval, including traffic monitoring video capture module and a video image based on the content analysis module, wherein the vehicle further comprising a video key frame acquisition module, monitor module and the extended data of the encoding code flow search module, wherein: the role of video content analysis module to extract traffic events and traffic object from the key frame classification; GET action video key frame module there is provided an image in the event of movement of the lens from the critical frame, and transmitting the traffic event related objects; monitoring role extension data encoding module to monitor for extended data expansion, to meet the grammatical object of traffic monitoring extension data unit distinguished by EXTENSION_ID, taking EXTENSION_ID monitored object is represented 0x6 extended syntax, semantics of the same from the beginning to extension_id syntax element with semantics of other monitoring reserve_bits extension data unit is consistent semantic loop syntax for each object found from each region, and to obtain from the detailed object region_object_id 征信息,语法元素值的含义由应用本身决定,对交通对象的region_object_id进行定义,通过扩展后的SVAC标准内容对交通对象和交通事件进行编码; 对码流检索模块的作用为根据已知的SVAC码流和交通对象或交通事件的特征编码,从码流的监控扩展数据单元中找到交通对象或交通事件所在的图像位置以及所在图像中区域的位置,记录下图像位置所对应的区域具体参数,从编码的视频图像数据中找到图像数据并解码显示,同时标记出相应的感兴趣区域。 Extrinsic information, meaning the value of the syntax element is determined by the application itself, for region_object_id traffic object is defined, and encodes the traffic event by the traffic object SVAC extended standard contents; the role of the module is retrieved code stream according to known SVAC stream and traffic object or feature encoding traffic event, to find the position of the image position traffic object or traffic events are located and a region where the image from the monitor extension data unit stream, the region-specific parameters of the image corresponding to the position recorded, find the image data from the encoded video image data and decoded for display, and mark the corresponding region of interest.
  2. 2.根据权利要求I所述的基于标准标记与快速检索的交通视频监控系统,其特征在于,所述关键帧信息直接来自视频图像内容分析模块,同时包含了检测出的交通对象或交通事件,对于事件,SVAC有针对事件的监控事件扩展语法结构:event_extension,其中的region_event_id表示事件特征,将region_event_id针对交通视频业务进行编码标准定义。 According to claim I of the standard video surveillance system marker and Transportation based fast retrieval, wherein the key frame information directly from the video content analysis module, contains the traffic object or traffic events detected, for the event, SVAC have extended syntax structure for monitoring events events: event_extension, which represents an event of region_event_id features, will region_event_id encode standard definition video for the transportation business.
  3. 3.根据权利要求I所述的基于标准标记与快速检索的交通视频监控系统,其特征在于,在找关键帧之前,先将运动镜头按照其中间的车辆复杂度进行分类。 The numerals I according to standard video traffic monitoring system based fast retrieval, characterized in that, prior to looking for the key frame, the first movement of the lens in accordance with the classification of the vehicle between the complexity of claims.
CN 201210056022 2012-03-01 2012-03-01 A traffic video monitoring system based on standard labeling and quick search CN102665064A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201210056022 CN102665064A (en) 2012-03-01 2012-03-01 A traffic video monitoring system based on standard labeling and quick search

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201210056022 CN102665064A (en) 2012-03-01 2012-03-01 A traffic video monitoring system based on standard labeling and quick search

Publications (1)

Publication Number Publication Date
CN102665064A true true CN102665064A (en) 2012-09-12

Family

ID=46774463

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201210056022 CN102665064A (en) 2012-03-01 2012-03-01 A traffic video monitoring system based on standard labeling and quick search

Country Status (1)

Country Link
CN (1) CN102665064A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103268702A (en) * 2013-05-03 2013-08-28 福建工程学院 Method for obtaining evidence of illegal occupying of non-bus vehicles on bus lane
CN103412859A (en) * 2012-10-11 2013-11-27 华迪计算机集团有限公司 Method and device for rapidly searching for enormous videos based on media asset management system
CN104504732A (en) * 2014-12-25 2015-04-08 合肥寰景信息技术有限公司 Video content retrieval method based on key frame extraction
CN105450978A (en) * 2014-06-24 2016-03-30 杭州海康威视数字技术股份有限公司 Method and device for achieving structural description in video monitoring system

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101547351A (en) * 2008-03-24 2009-09-30 展讯通信(上海)有限公司 Method for generating and processing video data stream and equipment thereof
CN101778260A (en) * 2009-12-29 2010-07-14 公安部第三研究所 Method and system for monitoring and managing videos on basis of structured description
CN101902617A (en) * 2010-06-11 2010-12-01 公安部第三研究所 Device and method for realizing video structural description by using DSP and FPGA
CN102207966A (en) * 2011-06-01 2011-10-05 华南理工大学 Video content quick retrieving method based on object tag

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101547351A (en) * 2008-03-24 2009-09-30 展讯通信(上海)有限公司 Method for generating and processing video data stream and equipment thereof
CN101778260A (en) * 2009-12-29 2010-07-14 公安部第三研究所 Method and system for monitoring and managing videos on basis of structured description
CN101902617A (en) * 2010-06-11 2010-12-01 公安部第三研究所 Device and method for realizing video structural description by using DSP and FPGA
CN102207966A (en) * 2011-06-01 2011-10-05 华南理工大学 Video content quick retrieving method based on object tag

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103412859A (en) * 2012-10-11 2013-11-27 华迪计算机集团有限公司 Method and device for rapidly searching for enormous videos based on media asset management system
CN103412859B (en) * 2012-10-11 2016-09-14 华迪计算机集团有限公司 Massive video media asset management system, method and apparatus for rapid retrieval
CN103268702A (en) * 2013-05-03 2013-08-28 福建工程学院 Method for obtaining evidence of illegal occupying of non-bus vehicles on bus lane
CN103268702B (en) * 2013-05-03 2015-09-23 福建工程学院 One kind of non-public PDI forensic methods illegal occupation of bus lanes
CN105450978A (en) * 2014-06-24 2016-03-30 杭州海康威视数字技术股份有限公司 Method and device for achieving structural description in video monitoring system
CN104504732A (en) * 2014-12-25 2015-04-08 合肥寰景信息技术有限公司 Video content retrieval method based on key frame extraction

Similar Documents

Publication Publication Date Title
Meng et al. CVEPS-a compressed video editing and parsing system
US7035468B2 (en) Methods and apparatus for archiving, indexing and accessing audio and video data
US6909745B1 (en) Content adaptive video encoder
US6810086B1 (en) System and method of filtering noise
US20030210821A1 (en) Methods and apparatus for generating, including and using information relating to archived audio/video data
US20060059510A1 (en) System and method for embedding scene change information in a video bitstream
US20030026340A1 (en) Activity descriptor for video sequences
US6970513B1 (en) System for content adaptive video decoding
US20110122255A1 (en) Method and apparatus for detecting near duplicate videos using perceptual video signatures
US20060187358A1 (en) Video entity recognition in compressed digital video streams
US20030174893A1 (en) Digital image storage method
Gunsel et al. Temporal video segmentation using unsupervised clustering and semantic object tracking
US20130170557A1 (en) Method and System for Video Coding with Noise Filtering
US20080267290A1 (en) Coding Method Applied to Multimedia Data
US20120106806A1 (en) Face Recognition in Video Content
US20120076357A1 (en) Video processing apparatus, method and system
Wang et al. A confidence measure based moving object extraction system built for compressed domain
US7773670B1 (en) Method of content adaptive video encoding
Pua et al. Real time repeated video sequence identification
US8699581B2 (en) Image processing device, image processing method, information processing device, and information processing method
Erol et al. Linking presentation documents using image analysis
Liu et al. Key frame extraction from MPEG video stream
US20090256972A1 (en) Methods and apparatus to generate and use content-aware watermarks
US20110267544A1 (en) Near-lossless video summarization
US7277485B1 (en) Computer-readable medium for content adaptive video decoding

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
C12 Rejection of a patent application after its publication