1263890 玖、發明說明 【發明所屬之技術領域】 本發明是有關於偵測景象改變(scene change),且特別 是有關於一種在一連串複數個視訊圖框(video frame)中智 慧型(intelligently)偵測一景象改變之位置的方法以及應用 此方法的監視系統。 【先前技術】 在傳統的監視系統(surveillance or security system) 中’經常需要使用照相機(camera)持續觀察一被監視區域 (surveillance area),以擷取許多物體的一連串影像。有時會 用錄影機(video recorder)將被擷取的影像記錄到一錄影媒 介(recording medium)上,以供保存與重複放影。常見的是 以電腦對所操取的影像進行影像處自,而幻吏用榮幕顯示 所擷取的影像,以供人員或操作者(operator)觀看檢查。 常見的監視一被監視區域的目的包括確保此區域中的 目標或物體(object)的安全與不受侵擾。然而,當—目桿在 -段時間中的-連串影像已被擷取且儲存下來以後,: 此段時間中此目標已經被移動,或消失而令操 需要找出此目標在何時被移動的肖 、 能連續地顯示所儲存下來的-連串影像,=:::: 到尾觀看’以找出此目標在一連串影像 (―㈣的位置。這種做法是不方 ^ ^ 與缺乏效率的,而且以人來觀看檢 二」,間的 一奋易有遺漏景象改變 1263890 主η #者’當為了只檢查單-物體於何時發生變動的 月 > 下,卻必須觀看整個被監視區域,如此是不經濟的, 而且,h查中也易受到此區域中其他物體影像的干擾。 /此,有需要發展出一種摘測景象改變的方法,以避免 傳、充上以人來觀看檢查景象改變的位置之做法。 【發明内容】 ^因此本發明的目的就是在提供一種智慧型偵測一景象 改變的方法,用以避免傳統上以人來觀看檢查景象改變的 位置之做法。 ^本發明的另一目的是在提供一種智慧型偵測一景象改 變的裝置以及一種監視系統,皆應用上述的偵測景象改變 之方法。 、—根據本發明之上述目的,提出一種智慧型偵測在一連串 複數個視訊圖框中一景象改變之位置的方法。此方法包含 乂下步驟。首先,在這些視訊圖框之至少一視訊圖框中選 擇代表一目標之一影像;再產生資料,這些資料代表在每 一視訊圖框中此影像的資訊;然後,處理這些資料以找出 這些視訊圖框之間在此資訊中有關視覺的(visual)改變·,之 後’再辨識出這些有關視覺的改變之視訊圖框位置,以辨 識出可能的景象改變之視訊圖框位置。 依照本發明一較佳實施例,一種監視系統包含影像擷取 單元、一景象改變偵測裝置以及一第五元件。影像擷取單 元被安排以觀察一景象。此景象改變偵測裝置包含第一元 1263890 !、第二元件、第三元件以及第四元件。第一元件用以在 一連串複數個視訊圖框之至少_雜#门 在 ^ 硯訊圖框中選擇代砉—曰 才示之一影像。第二元件用以產生資 _ — 貝料,些資料代表/斤 一視訊圖框中此影像的資訊。第三 母 、、 ~"凡件用以處理這itb資粗 以找出這些視訊圖框之間在此資 一、" J仕此貝巩中有關視覺的改變。 四凡件用以辨識出這些有關視覺 币 凡見的改變之視訊圖框位置, :辨識出可能的景象改變之視訊圖框位置。當偵測到—景 象改變時’第五元件用以產生一輪出信號。 【實施方式】 在監視系統中,照相機根據所擷取的影像而產生的視訊 。虎會被分為一連串複數個視訊圖框。在視訊㈨―)領域 中,圖框(frame)即是一連串影像中的單一靜態(stiu)影像。 :發月提t、冑智慧型偵測在一連串複數個視訊圖框中一 τ、象改交之 立置的方法。帛丨_是此智慧型债測景象改變 :法的机私圖。請參照帛i圖,此偵測景象改變之方法包 -、下乂驟首先’選擇在這些視訊圖框之至少一視訊圖 杧中代表一目標之一影像,如第丨圖中步驟丨〇〇所示。再 產生資料,這些資料代表在每一視訊圖框中此影像的資 〇Tl如步驟i OS所示。然後,處理這些資料以找出這些視 图才之間在此資訊中有關視覺的(visual)改變,如步驟104 ^ °之後’再辨識出這些有關視覺的改變之視訊圖框位 \jp ’ W辨識出可能的景象改變之視訊圖框位置,如步驟1 06 所示。 1263890 在此敘述本發明的一較佳實施例。第2圖係為依昭本發 明:較佳實施例的一監視系統之示意圖。在第2圖中的監 視系統具有三個影像擷取單元(image pickup unit)2〇〇、 與204、t腦主機2〇6卩及顯示勞$ 2i2。影像榻取單元 、202與204例如為照相機或攝影機(video camera)。電 腦主機206内部具有一視訊獲得卡(vide〇邮㈣c叫或數 位視訊裝置(digital video㈣⑷期、儲存裝置 deviCe)2G9以及處理n(pr()eess()r)21()。儲存裝i谢例如 為磁碟機(magnetic disk drives)。電腦主機2〇6藉由例如一 視訊配接卡(video adapter)(未繪出)連接到顯示螢幕2丨2。處 理器210電性連接到視訊獲得卡2〇8以及儲存裝置2〇9。 影像擷取單元200、202與204分別連接到電腦主機2〇6 中的視訊獲得卡208,且被安排以觀察一被監視區域中物體 的厅、象匕們的功能是將代表物體(目標)的光學資訊 Optical information)轉換成代表此物體的影像信號。視訊獲 付卡208也可能為一種視訊配接卡,且通常具有記憶體 (memory)以儲存資料。視訊獲得卡2〇8的功能是將影像擷 取單7L 200、202與204所產生的類比視訊信號轉換為數位 的格式而且壓縮視訊資料。有的數位視訊裝置則負責獲得 〜像且、、二由;}:示準的串列或並列介面〇r parauei interface)將影像傳送至一電腦中。壓縮視訊資料例如根據 一壓縮演算法(compressi〇n alg〇rhhm),而此壓縮演算法例 々使用種減才貝壓'% (1 〇 s s y c 〇 m p r e s s i ο η)技術。壓縮後的視 訊資料然後經常會被儲存於儲存裝置2〇9。 1263890 個視=視訊獲得卡208處理後的視訊資料是_連串複數 21 回才 這二連串視訊圖框經常被顯示於顯示螢幕 以供操作者監視觀看。在此實施例中,偵測景象改 二糟由處理H 21G來進行,以辨識出儲存於儲存裝置 這二連串複數個視訊圖框中一景象改變之位置。首 =里σσ 2 1 G在這些已儲存的視訊圖框之至少一視訊圖 a 4擇代表-目標之一影像。此選擇動作可能發生於例 操作者在顯示螢幕212上觀看此視訊圖框,且下指令 搜尋此目標的景象何時發生改變,例如藉由點選⑷叫^顯 丁螢幕2 1 2上此視訊圖框中代表此目標之影像。處理器2 j 〇 接著產生資料’這些資料代表在每一視訊圖框中此影像的 貧訊。此資訊即源於壓縮後的視訊資料。然後,處理器21〇 =理這些資料以找^些視訊圖框之間纟此資訊中有關視 覺的(visual)改變。此資訊中有關視覺的改變可以依照一預 定的門檻值(threshoM)來決定,也就是當有關視覺的改變量 :旦超過此門檻值,則被處理器21〇決定為有關視覺的改 蜒。此外,有關視覺的改變例如可設定為採取至少持續一 段時間之有關視覺的改變。最後,處理器2丨〇再辨識出這 些有關視覺的改變之視訊圖框位置,以辨識出可能的景象 改變之視訊圖框位置。此外,處理器21〇可被設計為在偵 測到景象改變之視訊圖框位置後’產生一輸出信號以告知 操作者此目標在何時發生景象改變。 在第2圖中的監視系統中,這些視訊圖框之間在此資訊 中有關視覺的改變可被解釋為例如一闖入者(intruder)進入 1263890 被監視的區域而改變此目標的景象。n 口此,當偵測到一畢 象改變時,可採取各種措施以辨識出— 厅、 闖入者的存在。再 者’如弟2圖所不’監視糸統一般來#、吾θ 木况遏具有錄影機(video rec〇rder)214設備,連接到影像擷取單 早70 2〇〇、202 與 204, 用以在需要的時候將影像擷取單元20〇 、202與204的輸出 記錄㈣⑽)到錄影媒介上。錄影媒介例如為錄影帶或磁帶 (video tape)。另-種可能是直接使用攝錄影機 作為影像擷取單元200、202與204,以、鹿么+ Μ避免需要使用單獨 的錄影機。錄影機2 14可能連接到電腦主 而接受電 腦主機206巾某裝置例如處理器、21〇的控制。舉例而今, 當偵測到闖入者時,可能有必要增進所錄下的影像:S 質,以利掌握景象的變化,故處理3 21〇纟此時使錄影: 214進入動作的狀態(activated),而且使錄影機214進入動 作狀態的過程可能包括修改錄影機214之錄影程序,以提 升因此所錄下的影像之品質。另一種方式為,處理写21〇 使錄影機2 1 4只有當闖入者被偵測到的時候才錄影。 在另一實施例中,一連串視訊圖框係源自於經過編輯的 視訊(edited video),而非直接取自影像擷取單元。在又一實 施例中,一連串視訊圖框係源自於一來源錄像 recording),且景象改變之彳貞測被實施以幫助這些一連串視 讯圖框之編輯。錄像(rec〇rding)即是指已被記錄而儲存下來 的影像。 由上述本發明較佳實施例可知,應用本發明的優點 包括避免傳統上以人來觀看檢查景象改變的位置之做法, l26389〇 所以能提升效率 #者所下的指令 法〇 、降低成本。而且可以自動搜尋或回應操 進行自動搜尋景象改變,所以是進步的方 、雖然本發明已以較佳實施例揭露如上,然其並非用 =限定本發明’任何熟習此技#者,在不脫離本發明之 :、申和fe圍内’當可作各種之更動與潤飾,因此本發明 保護範圍當視後附之申請專利範圍所界定者為準。 【圖式簡單說明】 的、特徵、和優點能更明顯 並配合所附圖式,作詳細 為讓本發明之上述和其他目 易it,下文特舉一較佳實施例, 說明如下: 圖疋此智慧型偵測景象改變方法的流程圖。 示意圖。2目係為依照本發明—較佳實施例的—監視系統之 【元件代表符號簡單說明】 100 200 206 209 102 、 1〇4 與 1〇6 : 2〇2與204 :影像 電腦主機 儲存裝置 顯示螢幕 步騍 擷取單元 :視訊獲得卡 21 〇 ··處理器 214 :錄影機 2121263890 BACKGROUND OF THE INVENTION 1. Field of the Invention This invention relates to detecting scene changes, and more particularly to intelligently detecting in a series of video frames. A method of measuring the position of a scene change and a monitoring system to which the method is applied. [Prior Art] In a conventional surveillance system (reveillance or security system), it is often necessary to use a camera to continuously observe a surveillance area to capture a series of images of many objects. The captured video is sometimes recorded on a recording medium for storage and repeat playback. It is common to use a computer to image the image being manipulated, and the illusion to display the captured image with a glory for the person or operator to view the inspection. A common purpose of monitoring a monitored area is to ensure that objects or objects in this area are safe and unaffected. However, when the - series of images in the - time period have been captured and stored, this target has been moved or disappeared during this time so that the operation needs to find out when the target was moved. Xiao, can continuously display the stored - series of images, =:::: End to watch 'to find out the target in a series of images (- (four) position. This is not true ^ ^ and inefficiency And the people who come to watch the second check, there is a lack of scene change 1263890 main η #者', in order to check only the single-object when the month of change>, but must watch the entire monitored area This is uneconomical, and h investigation is also susceptible to interference from other objects in this area. / /, there is a need to develop a method of picking up scene changes to avoid transmission, charging and watching. [Invention] The object of the present invention is therefore to provide a method for intelligently detecting a scene change in order to avoid the traditional practice of checking the position of a scene change by a person. Another object of the present invention is to provide an intelligent detection device for scene change and a monitoring system, which are all applicable to the method for detecting scene change. According to the above object of the present invention, an intelligent detection method is provided. a method of changing the position of a scene in a plurality of video frames. The method includes the following steps. First, selecting one of the images in at least one of the video frames of the video frame; generating data, the data Represents information about the image in each video frame; then, processing the data to find visual changes in the information between the video frames, and then 'recognizing these visual changes The video frame position is used to identify the position of the video frame of the possible scene change. According to a preferred embodiment of the present invention, a monitoring system includes an image capturing unit, a scene change detecting device and a fifth component. The capture unit is arranged to observe a scene. The scene change detecting device comprises a first element 1263890!, a second component, a third component and a fourth component. The first component is configured to select one of the images in at least one of the plurality of video frames, and the second component is used to generate the image. _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ First, " J Shi this about the visual changes in Begong. Four pieces are used to identify the position of the video frame of the change in the visual currency, to identify the location of the video frame of the possible scene change. Detected - the fifth component is used to generate a round of signal when the scene changes. [Embodiment] In the surveillance system, the camera generates video according to the captured image. The tiger will be divided into a series of video frames. . In the field of video (9)-), a frame is a single static image of a series of images. : Sending a month to mention t, 胄 smart detection in a series of video frames in a τ, like the method of changing the standing.帛丨 _ is this smart debt measurement scene change: the law of the machine. Please refer to the 帛i diagram, the method for detecting the scene change--the next step is to first select to represent one of the targets in at least one of the video frames of the video frame, as in the step 丨 in the figure 丨〇〇 Shown. Re-generate data representing the information Tl of this image in each video frame as shown in step i OS. Then, process the data to find out the visual changes in the information between the views, such as after step 104 ^ 're-identify these visual changes to the visual frame\pp 'W Identify the location of the video frame for possible scene changes, as shown in step 106. 1263890 A preferred embodiment of the present invention is described herein. Figure 2 is a schematic representation of a monitoring system of the preferred embodiment. The monitoring system in Fig. 2 has three image pickup units 2, 204, and t brain hosts 2〇6卩 and displays $2i2. The image couching units, 202 and 204 are, for example, cameras or video cameras. The host computer 206 internally has a video acquisition card (vide mail (4) c or digital video device (digital video (4) (4), storage device deviCe) 2G9 and processing n (pr ()eess () r) 21 (). It is a magnetic disk drive. The host computer 2〇6 is connected to the display screen 2丨2 by, for example, a video adapter (not shown). The processor 210 is electrically connected to the video acquisition card. 2〇8 and storage device 2〇9. The image capturing units 200, 202 and 204 are respectively connected to the video obtaining card 208 in the host computer 2〇6, and are arranged to observe the hall of the object in the monitored area, symbolizing Their function is to convert optical information representing an object (target) into an image signal representing the object. The video-paid card 208 may also be a video adapter card and typically has a memory to store data. The function of the video acquisition card 2〇8 is to convert the analog video signals generated by the image capture units 7L 200, 202 and 204 into a digital format and compress the video data. Some digital video devices are responsible for obtaining images like ~, , and two; }: the serial or parallel interface 〇r parauei interface) to transfer images to a computer. The compressed video data is for example based on a compression algorithm (compressi〇n alg〇rhhm), and the compression algorithm uses the technique of reducing the pressure of '% (1 〇 s s y c 〇 m p r e s s i ο η). The compressed video data is then often stored in the storage device 2〇9. 1263890 The video data processed by the video capture card 208 is _ consecutive plural 21 back. These two series of video frames are often displayed on the display screen for the operator to monitor and watch. In this embodiment, the detection of the scene is performed by the processing of the H 21G to identify the location of a scene change in the plurality of video frames stored in the storage device. First = σσ 2 1 G At least one of the video frames in the stored video frames a 4 represents one of the target images. This selection action may occur when the operator views the video frame on the display screen 212, and when the instruction to search for the target changes, for example, by clicking (4), the video is displayed on the display screen. The box represents the image of this target. Processor 2 j 〇 then generate data 'These data represent the poorness of this image in each video frame. This information is derived from the compressed video material. The processor 21 then uses the data to find visual changes in the information between the video frames. The visual change in this information can be determined according to a predetermined threshold (threshoM), that is, when the amount of visual change is exceeded: the threshold is determined by the processor 21 as a visual change. Furthermore, the change in vision can be set, for example, to take a visual change that lasts at least for a period of time. Finally, the processor 2 recognizes the position of the video frame for visual changes to identify the location of the video frame for possible scene changes. In addition, processor 21A can be designed to generate an output signal after the location of the video frame in which the scene change is detected to inform the operator when the scene has changed. In the surveillance system of Figure 2, the visual changes in the information between these video frames can be interpreted as, for example, an intruder entering the 1263890 monitored area to change the view of the target. n, in this case, when a change in the image is detected, various measures can be taken to identify the presence of the hall and the intruder. In addition, 'If the brother 2 map does not monitor the general system to #, I θ wood condition suppression with video recorder (video rec〇rder) 214 equipment, connected to the image capture single early 70 2〇〇, 202 and 204, It is used to record (4) (10) the output of the image capturing units 20, 202 and 204 to the video medium when needed. Video media such as for video tapes or tapes (video tape). Another possibility is to use the camcorder directly as the image capturing units 200, 202 and 204, so that the deer + Μ avoids the need to use a separate video recorder. The video recorder 2 14 may be connected to the computer main body and receive control of a computer host 206 such as a processor, 21 。. For example, when an intruder is detected, it may be necessary to enhance the recorded image: S quality, in order to grasp the change of the scene, so process 3 21〇纟 at this time to make the video: 214 into the action state (activated) And the process of causing the video recorder 214 to enter an operational state may include modifying the video recording program of the video recorder 214 to enhance the quality of the recorded video. Another way is to process the write 21 〇 so that the recorder 2 1 4 only records when the intruder is detected. In another embodiment, a series of video frames are derived from edited video rather than directly from the image capture unit. In yet another embodiment, a series of video frames are derived from a source video recording, and scene change detection is implemented to aid in the editing of the series of video frames. Recording (rec〇rding) refers to an image that has been recorded and stored. It can be seen from the above-described preferred embodiments of the present invention that the advantages of applying the present invention include avoiding the traditional practice of checking the position of the scene change by humans, so that the efficiency of the instructions can be improved and the cost can be reduced. Moreover, the automatic search or response operation can be performed to automatically search for scene changes, so that it is a progressive party. Although the present invention has been disclosed in the preferred embodiment as above, it is not intended to limit the present invention to any of those skilled in the art. The invention is not limited to the scope of the patent application, and the scope of the invention is defined by the scope of the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS The above and other objects of the present invention will be described in detail with reference to the accompanying drawings. A flow chart of a smart detection scene change method. schematic diagram. 2 is a simple description of the component representative symbols of the monitoring system according to the present invention - a preferred embodiment 100 200 206 209 102 , 1 〇 4 and 1 〇 6 : 2 〇 2 and 204 : image computer main storage device display Screen Step Capture Unit: Video Acquisition Card 21 处理器 Processor 214: Video Recorder 212