TW202129597A - Image capturing method and electronic device - Google Patents
Image capturing method and electronic device Download PDFInfo
- Publication number
- TW202129597A TW202129597A TW109101963A TW109101963A TW202129597A TW 202129597 A TW202129597 A TW 202129597A TW 109101963 A TW109101963 A TW 109101963A TW 109101963 A TW109101963 A TW 109101963A TW 202129597 A TW202129597 A TW 202129597A
- Authority
- TW
- Taiwan
- Prior art keywords
- images
- image
- image capturing
- focal lengths
- target sub
- Prior art date
Links
Images
Abstract
Description
本發明是有關於一種影像擷取方法,且特別是有關於一種影像擷取方法與使用所述影像擷取方法的電子裝置。The present invention relates to an image capturing method, and more particularly to an image capturing method and an electronic device using the image capturing method.
傳統的影像辨識操作會需要所辨識的影像中的目標物件的影像(如,用於人臉辨識操作的人臉影像)是清晰的。為了獲得清晰的所述目標物件的所述影像,影像擷取裝置需要去掃描/對準所述目標物件,以使焦距可以準確地位於所述目標物件上,進而獲得清晰的已對焦的所述目標物件的所述影像。The traditional image recognition operation requires that the image of the target object in the recognized image (for example, the face image used for the face recognition operation) is clear. In order to obtain a clear image of the target object, the image capture device needs to scan/align the target object so that the focal length can be accurately located on the target object, thereby obtaining a clear focused image The image of the target object.
然而,在多個目標物件的總數目很大(多於兩個以上)的情境下,為了獲得清晰的每一個目標物件的影像,往往因需要針對每一個目標物件來進行對焦與拍攝而耗費了大量的時間和資源,導致了對應的影像擷取操作的效率低下,更使後續的影像辨識操作需耗費時間與運算資源來接收與合成所述多個目標物件的多個影像。However, in a situation where the total number of multiple target objects is large (more than two), in order to obtain a clear image of each target object, it is often costly due to the need to focus and shoot each target object. A large amount of time and resources result in low efficiency of the corresponding image capturing operation, and the subsequent image recognition operation requires time and computing resources to receive and synthesize multiple images of the multiple target objects.
本發明提供一種影像擷取方法與電子裝置,可根據多個焦距擷取多個影像且辨識所述多個影像中的對應所述多個焦距的多個目標區域,以獲得所述多個目標區域內的多個目標子影像,進而根據所述多個目標子影像產生對應所述時間點的單一的重建影像,其中於所述重建影像中的多個物件影像皆為對焦的。The present invention provides an image capturing method and electronic device, which can capture multiple images according to multiple focal lengths and identify multiple target regions corresponding to the multiple focal lengths in the multiple images to obtain the multiple targets A plurality of target sub-images in the area are further generated according to the plurality of target sub-images to generate a single reconstructed image corresponding to the time point, wherein the plurality of object images in the reconstructed image are all in focus.
本發明的一實施例提供適用於電子裝置的一種影像擷取方法。所述電子裝置包括多個影像擷取裝置,其中所述多個影像擷取裝置分別使用不同的多個焦距來進行拍攝。所述方法包括:經由分別對應所述多個焦距的所述多個影像擷取裝置,同時地對一監控場景進行拍攝,以擷取對應所述多個焦距的於同一個時間點的多個影像,其中所述多個影像的總數目相等於所述多個焦距的總數目;根據多個焦段,辨識所擷取所述多個影像的各自的目標區域,其中所述多個目標區域對應所述多個焦段,並且所述多個焦段分別對應所述多個焦距;保留於所述多個影像的所述多個目標區域內的多個目標子影像,其中於所述多個目標子影像中的多個物件影像皆為對焦的;直接根據所述多個目標子影像產生對應所述時間點的單一的重建影像;以及輸出所述重建影像。An embodiment of the present invention provides an image capturing method suitable for electronic devices. The electronic device includes a plurality of image capturing devices, wherein the plurality of image capturing devices use different focal lengths to shoot. The method includes: simultaneously shooting a monitoring scene through the plurality of image capturing devices corresponding to the plurality of focal lengths, so as to capture a plurality of images corresponding to the plurality of focal lengths at the same time point. Image, wherein the total number of the plurality of images is equal to the total number of the plurality of focal lengths; according to the plurality of focal lengths, the respective target regions of the captured images are identified, wherein the plurality of target regions correspond to The multiple focal lengths, and the multiple focal lengths respectively correspond to the multiple focal lengths; multiple target sub-images retained in the multiple target regions of the multiple images, wherein the multiple target sub-images Multiple object images in the image are all in focus; directly generate a single reconstructed image corresponding to the time point according to the multiple target sub-images; and output the reconstructed image.
本發明的一實施例提供適用於電子裝置的一種影像擷取方法,其中所述電子裝置包括可變焦的影像擷取裝置。所述方法包括:經由的所述影像擷取裝置,根據多個焦距依序對監控場景連續地進行多次拍攝,以擷取對應所述多個焦距的多個影像,其中所述多個影像的總數目相同於所述多個焦距的總數目,並且所述多個影像的多個時間點為連續的;根據多個焦段,辨識所擷取所述多個影像的各自的目標區域,其中所述多個目標區域對應所述多個焦段,並且所述多個焦段分別對應所述多個焦距;保留於所述多個影像的所述多個目標區域內的多個目標子影像,其中於所述多個目標子影像中的多個物件影像皆為對焦的;直接輸出該子影像;直接根據所述多個目標子影像產生單一的重建影像;以及輸出所述重建影像。An embodiment of the present invention provides an image capturing method suitable for an electronic device, wherein the electronic device includes a zoomable image capturing device. The method includes: using the image capturing device to sequentially take multiple shots of the monitoring scene according to multiple focal lengths to capture multiple images corresponding to the multiple focal lengths, wherein the multiple images The total number of is the same as the total number of the multiple focal lengths, and the multiple time points of the multiple images are continuous; according to the multiple focal lengths, the respective target regions of the captured multiple images are identified, wherein The multiple target regions correspond to the multiple focal lengths, and the multiple focal lengths respectively correspond to the multiple focal lengths; multiple target sub-images retained in the multiple target regions of the multiple images, wherein Multiple object images in the multiple target sub-images are all focused; directly output the sub-image; directly generate a single reconstructed image based on the multiple target sub-images; and output the reconstructed image.
本發明的一實施例提供一種電子裝置。所述電子裝置包括:固定焦距的多個影像擷取裝置、儲存裝置以及處理器。所述多個影像擷取裝置用以分別使用不同的多個焦距來進行拍攝。儲存裝置用以儲存資料,其中所述資料包括多個程式碼。所述處理器用以存取且執行所述多個程式碼以實現影像擷取方法。此外,所述多個影像擷取裝置用以同時地對監控場景進行拍攝,以擷取對應所述多個焦距的於同一個時間點的多個影像,其中所述多個影像的總數目相等於所述多個焦距的總數目。所述處理器用以根據多個焦段,辨識所擷取所述多個影像的各自的目標區域,其中所述多個目標區域對應所述多個焦段,並且所述多個焦段分別對應所述多個焦距。所述處理器更用以保留於所述多個影像的所述多個目標區域內的多個目標子影像,其中於所述多個目標子影像中的多個物件影像皆為對焦的。所述處理器更用以直接根據所述多個目標子影像產生對應所述時間點的單一的重建影像,其中所述處理器更用以輸出所述重建影像。An embodiment of the invention provides an electronic device. The electronic device includes: a plurality of image capturing devices with a fixed focal length, a storage device, and a processor. The multiple image capturing devices are used for shooting with different multiple focal lengths respectively. The storage device is used for storing data, wherein the data includes a plurality of codes. The processor is used for accessing and executing the multiple program codes to realize the image capturing method. In addition, the multiple image capturing devices are used to simultaneously shoot the monitoring scene to capture multiple images corresponding to the multiple focal lengths at the same time point, wherein the total number of the multiple images is the same Equal to the total number of the multiple focal lengths. The processor is used for identifying respective target regions of the captured images according to multiple focal lengths, wherein the multiple target regions correspond to the multiple focal lengths, and the multiple focal lengths correspond to the multiple focal lengths respectively. Focal lengths. The processor is further configured to retain a plurality of target sub-images in the plurality of target regions of the plurality of images, wherein the plurality of object images in the plurality of target sub-images are all in focus. The processor is further used to directly generate a single reconstructed image corresponding to the time point according to the multiple target sub-images, and the processor is further used to output the reconstructed image.
本發明的一實施例提供一種電子裝置。所述電子裝置包括:可變焦的影像擷取裝置、儲存裝置以及處理器。所述影像擷取裝置用以從多個焦距中選擇其中之一來進行拍攝。儲存裝置用以儲存資料,其中所述資料包括多個程式碼。處理器用以存取且執行所述多個程式碼以實現影像擷取方法。此外,所述影像擷取裝置用以根據所述多個焦距依序對監控場景連續地進行多次拍攝,以擷取對應所述多個焦距的多個影像,其中所述多個影像的總數目相同於所述多個焦距的總數目,並且所述多個影像的多個時間點為連續的。所述處理器更用以根據所述多個焦段,辨識所擷取所述多個影像的各自的目標區域,其中所述多個目標區域對應所述多個焦段,並且所述多個焦段分別對應所述多個焦距。所述處理器更用以保留於所述多個影像的所述多個目標區域內的多個目標子影像,其中於所述多個目標子影像中的多個物件影像皆為對焦的。所述處理器更用以直接根據所述多個目標子影像產生單一的重建影像。所述處理器更用以輸出所述重建影像。An embodiment of the invention provides an electronic device. The electronic device includes: a zoomable image capturing device, a storage device, and a processor. The image capturing device is used for selecting one of a plurality of focal lengths for shooting. The storage device is used for storing data, wherein the data includes a plurality of codes. The processor is used for accessing and executing the multiple program codes to realize the image capturing method. In addition, the image capturing device is used to sequentially take multiple shots of the monitoring scene according to the multiple focal lengths, so as to capture multiple images corresponding to the multiple focal lengths, wherein the total number of the multiple images The target is the same as the total number of the multiple focal lengths, and the multiple time points of the multiple images are continuous. The processor is further configured to identify respective target regions of the captured images according to the multiple focal lengths, wherein the multiple target regions correspond to the multiple focal lengths, and the multiple focal lengths are respectively Corresponding to the multiple focal lengths. The processor is further configured to retain a plurality of target sub-images in the plurality of target regions of the plurality of images, wherein the plurality of object images in the plurality of target sub-images are all in focus. The processor is further used to directly generate a single reconstructed image based on the multiple target sub-images. The processor is further used to output the reconstructed image.
基於上述,本發明的實施例所提供的影像擷取方法與電子裝置,可根據多個焦距對監控場景進行拍攝,以擷取多個影像,根據分別對應所述多個焦距的多個焦段來保留所述多個影像各自的目標子影像,以直接根據所述多個目標子影像來產生且輸出重建影像。如此一來,可在不分析所述多個影像及不貼合/縫合所述多個目標子影像的情況下,快速且有效率地使所產生的所述重建影像中位於多個深度的多個物件影像皆為對焦且清晰的,進而增進使用所述重建影像的影像辨識操作的精確度。Based on the above, the image capturing method and the electronic device provided by the embodiments of the present invention can shoot a monitoring scene according to multiple focal lengths to capture multiple images, according to multiple focal lengths corresponding to the multiple focal lengths. The target sub-images of each of the multiple images are retained to directly generate and output reconstructed images based on the multiple target sub-images. In this way, without analyzing the multiple images and not fitting/stitching the multiple target sub-images, multiple images located at multiple depths in the generated reconstructed image can be quickly and efficiently made. Each object image is focused and clear, thereby improving the accuracy of the image recognition operation using the reconstructed image.
為讓本發明的上述特徵和優點能更明顯易懂,下文特舉實施例,並配合所附圖式作詳細說明如下。In order to make the above-mentioned features and advantages of the present invention more comprehensible, the following specific embodiments are described in detail in conjunction with the accompanying drawings.
有關本發明之前述及其他技術內容、特點與功效,在以下配合參考圖式之一較佳實施例的詳細說明中,將可清楚的呈現。以下實施例中所提到的方向用語,例如:上、下、左、右、前或後等,僅是參考附加圖式的方向。因此,使用的方向用語是用來說明並非用來限制本發明。The foregoing and other technical content, features, and effects of the present invention will be clearly presented in the following detailed description of a preferred embodiment with reference to the drawings. The directional terms mentioned in the following embodiments, for example: up, down, left, right, front or back, etc., are only directions for referring to the attached drawings. Therefore, the directional terms used are used to illustrate but not to limit the present invention.
[第一實施例][First Embodiment]
圖1是根據本發明的第一實施例所繪示的監控場景的示意圖。請參照圖1,假設影像辨識裝置想要辨識一個場景(如圖1所繪示的監控場景SC1)中的多個特定物件的影像(如,監控場景SC1中的使用者U1~U4的臉部的影像FC1~FC4)。所述監控場景SC1例如是會議室,所述使用者U1~U4例如是與會者,但本發明並不限定監控場景SC1與特定物件的類型。例如,所述監控場景SC1也可為教室,所述使用者U1~U4例如是學生。Fig. 1 is a schematic diagram of a monitoring scene according to the first embodiment of the present invention. Please refer to Figure 1. Suppose that the image recognition device wants to recognize images of multiple specific objects in a scene (such as the surveillance scene SC1 shown in Figure 1) (such as the faces of users U1~U4 in the surveillance scene SC1). The images FC1~FC4). The monitoring scene SC1 is, for example, a conference room, and the users U1 to U4 are participants, for example, but the present invention does not limit the types of the monitoring scene SC1 and specific objects. For example, the monitoring scene SC1 may also be a classroom, and the users U1 to U4 are, for example, students.
一般來說,影像辨識裝置需要經由影像擷取裝置所提供的包含多個特定物件的影像FC1~FC4的監控場景SC1的多個影像,以對所獲得的多個影像進行影像辨識操作。所述多個特定物件(亦稱,目標物件)位於所述監控場景SC1中的不同位置,並且所述多個目標物件與用以拍攝所述監控場景SC1的影像擷取裝置之間的距離是不同的。若影像擷取裝置在不執行對焦的情況下對所述監控場景SC1進行拍攝,所擷取到影像中的多個目標物件的多個影像FC1~FC4會是模糊的,並且所述多個目標物件的多個影像FC1~FC4不能被有效地執行對應的影像辨識操作。Generally, the image recognition device needs to perform image recognition operations on the obtained multiple images through multiple images of the monitoring scene SC1 including images FC1 to FC4 of multiple specific objects provided by the image capturing device. The plurality of specific objects (also known as target objects) are located at different positions in the monitoring scene SC1, and the distance between the plurality of target objects and the image capturing device used to shoot the monitoring scene SC1 is different. If the image capture device shoots the surveillance scene SC1 without performing focusing, the multiple images FC1 to FC4 of multiple target objects captured in the image will be blurred, and the multiple targets Multiple images FC1~FC4 of the object cannot be effectively performed corresponding image recognition operations.
為了擷取清晰的多個目標物件的多個影像FC1~FC4,一般的傳統作法是,影像擷取裝置會利用距離偵測裝置(如,紅外線測距裝置或雷射測距裝置)來偵測影像擷取裝置與所述多個目標物件之間的多個距離,以根據所述多個距離來對焦至所述多個目標物件。每當所述多個目標物件的其中之一被對焦時,所述影像擷取裝置便拍攝所述監控場景SC1一次,以獲得一個具有所述被對焦的所述目標物件的一個影像。換句話說,以本場景而言,傳統方法至少需要經由執行四次的測距操作、四次的對焦操作與四次的拍攝操作,才可獲得可供影像辨識操作的四個合格影像。接著把所述四個合格影像傳送給影像辨識裝置。此外,影像辨識裝置更需要對所述四個合格影像的每一個執行一次影像辨識操作(共執行四次影像辨識操作),以獲得所有目標物件的影像的辨識結果。也就是說,針對上述的僅有四個目標物件的監控場景SC1的例子,為了輸出應用於影像辨識操作的合格影像,影像擷取裝置需要耗費大量的時間、資源(如,運算資源、儲存空間)於目標物件的影像的偵測、擷取與傳送。若監控場景SC1具有更多的目標物件,則整體耗費的時間與資源更會大量地增加。In order to capture multiple images FC1~FC4 of multiple target objects clearly, the general traditional method is that the image capture device uses a distance detection device (such as an infrared distance measurement device or a laser distance measurement device) to detect The multiple distances between the image capturing device and the multiple target objects are used to focus on the multiple target objects according to the multiple distances. Whenever one of the multiple target objects is focused, the image capturing device captures the monitoring scene SC1 once to obtain an image with the focused target object. In other words, in this scenario, the traditional method requires at least four ranging operations, four focusing operations, and four shooting operations to obtain four qualified images for image recognition operations. Then the four qualified images are sent to the image recognition device. In addition, the image recognition device further needs to perform an image recognition operation on each of the four qualified images (a total of four image recognition operations are performed) to obtain the recognition results of the images of all the target objects. In other words, for the above example of the surveillance scene SC1 with only four target objects, in order to output qualified images for image recognition operations, the image capture device needs to consume a lot of time and resources (such as computing resources, storage space). ) In the detection, capture and transmission of the image of the target object. If the surveillance scene SC1 has more target objects, the overall time and resources consumed will increase greatly.
為了解決上述傳統作法所面臨的技術問題,本發明的多個實施例所提供的影像擷取方法與電子裝置,可產生用以影像辨識的具有監控場景中所有的目標物件的影像的單一的重建影像,其中所述重建影像中的可被辨識的目標物件的影像皆是清晰或已對焦的(另外,所述重建影像為合格影像)。如此一來,除了增加獲得所述重建影像的速度之外,更可使影像辨識裝置可藉由單一的重建影像來執行影像辨識,進而增進了影像辨識操作整體的效率。以下先藉由圖2、圖3與第一實施例來說明本發明的影像擷取方法與電子裝置的具體細節。In order to solve the technical problems faced by the above-mentioned traditional practices, the image capturing method and electronic device provided by various embodiments of the present invention can generate a single reconstruction of images with all target objects in the monitoring scene for image recognition An image, wherein the images of the recognizable target object in the reconstructed image are all clear or focused (in addition, the reconstructed image is a qualified image). In this way, in addition to increasing the speed of obtaining the reconstructed image, the image recognition device can also perform image recognition through a single reconstructed image, thereby improving the overall efficiency of the image recognition operation. Hereinafter, the specific details of the image capturing method and the electronic device of the present invention will be described with reference to FIGS. 2, 3 and the first embodiment.
圖2是根據本發明的第一實施例所繪示的影像辨識系統的方塊示意圖。請參照圖2,本實施例的影像辨識系統1包括電子裝置10與人工智慧(AI)影像辨識裝置20。所述電子裝置10包括處理器110、主記憶體120、通訊電路單元130、儲存裝置140、影像擷取裝置陣列150。處理器110電性連接至主記憶體120、通訊電路單元130、儲存裝置140、影像擷取裝置陣列150。FIG. 2 is a block diagram of the image recognition system according to the first embodiment of the present invention. Please refer to FIG. 2, the
所述AI影像辨識裝置20包括處理器210、主記憶體220、通訊電路單元230、儲存裝置240。處理器210電性連接至主記憶體220、通訊電路單元230、儲存裝置240。The AI
在本實施例中,處理器110與處理器210為具備運算能力的硬體。處理器110用以存取且執行一或多個程式碼,以管理電子裝置10的整體運作;處理器210用以執行一或多個程式碼,以管理AI影像辨識裝置20的整體運作。在本實施例中,處理器110與處理器210,例如是一核心或多核心的中央處理單元(Central Processing Unit,CPU)、可程式化之微處理器(微處理器(Micro-processor))、數位訊號處理器(Digital Signal Processor,DSP)、可程式化控制器、特殊應用積體電路(Application Specific Integrated Circuits,ASIC)、可程式化邏輯裝置(Programmable Logic Device,PLD)或其他類似裝置。In this embodiment, the
主記憶體120、220用以分別接收處理器110、210的指示來暫存資料。例如,主記憶體120可暫存來自影像擷取裝置150所擷取的影像資料;主記憶體220可暫存來自電子裝置10所傳輸的資料(如,重建影像)。主記憶體120、220例如是動態隨機存取記憶體或靜態隨機存取記憶體。The
通訊電路單元130經由無線或有線的方式建立網路連線以連接通訊電路單元230,進而利用所建立的網路連線來傳輸資料。通訊電路單元130、230例如是具有一無線通訊電路單元(未繪示),並支援全球行動通信(Global System for Mobile Communication,GSM)系統、個人手持式電話系統(Personal Handy-phone System,PHS)、碼多重擷取(Code Division Multiple Access,CDMA)系統、無線相容認證(Wireless Fidelity,WiFi)系統、全球互通微波存取(Worldwide Interoperability for Microwave Access,WiMAX)系統、第三代無線通信技術(3G)、第四代無線通信技術(4G)、第五代無線通信技術(5G)、長期演進技術(Long Term Evolution, LTE)、紅外線(Infrared)傳輸、藍芽(bluetooth)通訊技術的其中之一或其組合,且不限於此。除此之外,通訊電路單元130、230也可具有一有線通訊電路單元(未繪示),並透過有線通訊的方式來傳輸或是接收資料。The
儲存裝置140、240分別用以經由處理器110、處理器210的指示來記錄一些需要長時間儲存的資料。儲存裝置140、240可以是任何型態的硬碟機(hard disk drive,HDD)或非揮發性記憶體儲存裝置(如,固態硬碟(SSD))。儲存裝置140所儲存的所述資料例如為用以管理電子裝置10的韌體或是軟體(如,程式碼143);以及多個資料庫(如,監控場景資料庫141、鏡頭規格資料庫142或其他適合的資料庫等…)。儲存裝置240所儲存的所述資料例如為用以管理AI影像辨識裝置20的韌體或是軟體;以及多個資料庫(如,目標影像樣本資料庫、AI影像辨識模型資料庫或其他適合的資料庫等…)。The
在本實施例中,處理器110用以存取且載入所述程式碼143以執行影像擷取操作,進而實現本發明所提供的影像擷取方法。此外,處理器210用以執行多個程式碼,以對所接收的重建影像執行影像辨識操作。In this embodiment, the
監控場景資料庫141用以記錄多個監控場景各自的空間資訊。所述空間資訊包括但不限於所述監控場景的高度、寬度與深度。鏡頭規格資料庫142用以記錄所述多個影像擷取裝置150(1)~150(N)各自的鏡頭規格。所述鏡頭規格例如是鏡頭的焦距、有效視野等。應注意的是,上述的多個資料庫亦可經由連接至網際網路或區域網路的網路連線所傳輸的資料被更新,或是經由輸入裝置所輸入的資料而被更新。The
所述影像擷取裝置陣列150包括多個影像擷取裝置150(1)~150(N)。N為大於1的正整數。所述多個影像擷取裝置150(1)~150(N)各自例如為具有擷取影像資料功能的攝影機或相機。所述多個影像擷取裝置150(1)~150(N)用以分別對監控場景SC1進行拍攝。各影像擷取裝置可包括鏡頭、感光元件以及光圈等。鏡頭例如是標準鏡頭、廣角鏡頭及變焦鏡頭等。感光元件例如是電荷耦合元件(Charge Coupled Device,CCD)、互補性氧化金屬半導體(Complementary Metal-Oxide Semiconductor,CMOS)元件或其他元件,鏡頭與感光元件或其組合在此皆不設限。各影像擷取裝置可連續擷取(拍攝)多個影像,即,擷取(拍攝)一動態影像(亦稱,視訊串流,Video Stream)。所述影像擷取裝置亦可擷取(拍攝)單一的靜態影像。所述動態影像或靜態影像會被暫存至主記憶體120,以讓處理器110對所述動態影像或靜態影像執行進一步的處理。The image
圖3是根據本發明的第一實施例所繪示的影像擷取方法的流程圖。請參照圖3,在步驟S301中,經由分別對應多個焦距的多個影像擷取裝置,同時地對監控場景進行拍攝,以擷取對應所述多個焦距的於同一個時間點的多個影像。FIG. 3 is a flowchart of the image capturing method according to the first embodiment of the present invention. Please refer to FIG. 3, in step S301, through multiple image capturing devices corresponding to multiple focal lengths, the monitoring scene is photographed at the same time, so as to capture multiple focal lengths corresponding to the multiple focal lengths at the same time point. image.
具體來說,在第一實施例中,所述多個影像擷取裝置150(1)~150(N)的焦距會被預先設定為不同的固定焦距。所述多個影像擷取裝置150(1)~150(N)會各自利用被預先設定的所述固定焦距一起同時地對監控場景SC1進行拍攝。換言之,於每個時間點,N個影像擷取裝置共可同時擷取/輸出監控場景SC1的N個影像。所述N個影像會被暫存至主記憶體120中。Specifically, in the first embodiment, the focal lengths of the plurality of image capturing devices 150(1)-150(N) are preset to different fixed focal lengths. The multiple image capturing devices 150(1)-150(N) will each use the preset fixed focal length to simultaneously capture the surveillance scene SC1. In other words, at each time point, N image capturing devices can simultaneously capture/output N images of the monitoring scene SC1. The N images will be temporarily stored in the
接著,在步驟S303中,根據多個焦段,處理器110辨識所擷取所述多個影像的各自的目標區域,其中所述多個目標區域對應所述多個焦段,並且所述多個焦段分別對應所述多個焦距。Then, in step S303, according to the multiple focal lengths, the
具體來說,所述多個焦段是根據所述多個影像擷取裝置150(1)~150(N)各自的所述焦距、所述多個鏡頭規格與所述多個影像擷取裝置150(1)~150(N)的總數目而決定的。更詳細來說,針對對應第一影像擷取裝置的所述多個焦段中的第一焦段,所述第一焦段的起始點至所述多個影像擷取裝置中的第一影像擷取裝置的第一焦距的第一焦距前距離以及所述第一焦段的終止點至所述第一焦距的第一焦距後距離是根據所述第一影像擷取裝置的第一鏡頭規格與所述多個影像擷取裝置的所述總數目而決定的。Specifically, the multiple focal lengths are based on the respective focal lengths of the multiple image capturing devices 150(1) to 150(N), the multiple lens specifications, and the multiple image capturing devices 150 (1) Determined by the total number of ~150(N). In more detail, for the first focal length of the plurality of focal lengths corresponding to the first image capturing device, the starting point of the first focal length to the first image capturing of the plurality of image capturing devices The first focal length before the first focal length of the device and the first focal length from the end point of the first focal length to the first focal length are based on the first lens specification of the first image capturing device and the Determined by the total number of multiple image capturing devices.
圖4A是根據本發明的第一實施例或第二實施例所繪示的對應多個焦距的多個焦段的示意圖。請參照圖4A,為了方便說明,以下的實施例會將N設定為4。舉例來說,假設每個影像擷取裝置150(1)~150(4)皆具有相同的視野深度FVL(可根據每個影像擷取裝置150(1)~150(4)而得知);影像擷取裝置150(1)~150(4)分別被設定為焦距FL1~FL4。在本實施例中,影像擷取裝置150(1)~150(4)各自的焦段FS1~FS4的總和會被設定為相等於所述視野深度FVL。此外,影像裝置150(1)的焦段FS1的起始點至焦距FL1的焦距前距離FS1_F以及影像裝置150(1)的焦段FS1的終止點至焦距FL1的焦距後距離FS1_B是根據影像擷取裝置150(1)的鏡頭規格與所述多個影像擷取裝置150的所述總數目(如,4)而決定的。相似地,影像裝置150(2)的焦段FS2的起始點至焦距FL2的焦距前距離FS2_F以及影像裝置150(2)的焦段FS2的終止點至焦距FL2的焦距後距離FS2_B是根據影像擷取裝置150(2)的鏡頭規格與所述多個影像擷取裝置150的所述總數目而決定;影像裝置150(3)的焦段FS3的起始點至焦距FL3的焦距前距離FS3_F以及影像裝置150(3)的焦段FS3的終止點至焦距FL3的焦距後距離FS3_B是根據影像擷取裝置150(3)的鏡頭規格與所述多個影像擷取裝置150的所述總數目而決定;影像裝置150(4)的焦段FS4的起始點至焦距FL4的焦距前距離FS4_F以及影像裝置150(4)的焦段FS4的終止點至焦距FL4的焦距後距離FS4_B是根據影像擷取裝置150(4)的鏡頭規格與所述多個影像擷取裝置150的所述總數目而決定。4A is a schematic diagram of multiple focal lengths corresponding to multiple focal lengths according to the first embodiment or the second embodiment of the present invention. Please refer to FIG. 4A. For the convenience of description, N is set to 4 in the following embodiment. For example, suppose that each image capturing device 150(1)~150(4) has the same depth of field FVL (which can be known according to each image capturing device 150(1)~150(4)); The image capturing devices 150(1)~150(4) are respectively set to focal lengths FL1~FL4. In this embodiment, the sum of the respective focal lengths FS1 to FS4 of the image capturing devices 150(1) to 150(4) is set to be equal to the depth of view FVL. In addition, the distance FS1_F from the starting point of the focal length FS1 of the imaging device 150(1) to the focal length FL1 and the distance after the focal length FS1_B from the end point of the focal length FS1 of the imaging device 150(1) to the focal length FL1 are based on the image capturing device The lens specification of 150(1) is determined by the total number (eg, 4) of the plurality of
圖4B是根據本發明的第一實施例所繪示的監控場景的拍攝區域與對應的多個焦段的示意圖。請參照圖4B,具體來說,如上述,所述監控場景資料庫記錄有所述監控場景SC1的空間資訊。此外,所述空間資訊包括所述監控場景SC1的高度、寬度與深度。根據所述監控場景的所述空間資訊、對應所述多個影像擷取裝置的所述多個焦段以及對應所述多個影像擷取裝置的所述多個鏡頭規格來決定所述多個影像各自的所述目標區域的位置、形狀與大小。舉例來說,針對所述多個影像中的經由影像擷取裝置150(1)所擷取的影像,處理器110根據所述監控場景SC1的所述空間資訊的所述高度、所述寬度與所述深度中的一或多者、對應所述影像擷取裝置150(1)的所述焦段FS1以及對應所述影像擷取裝置150(1)的所述鏡頭規格來決定所述影像中的目標區域TA1的位置、形狀以及多個邊長。其中,所述影像擷取裝置150(1)的所述鏡頭規格可決定對應監控場景SC1的拍攝區域SA1,並且拍攝區域SA1的深度為所述視野深度FVL。所述影像擷取裝置150(1)所擷取的影像的範圍即為所述拍攝區域SA1的範圍。4B is a schematic diagram of a shooting area and corresponding multiple focal lengths of a surveillance scene according to the first embodiment of the present invention. Please refer to FIG. 4B. Specifically, as described above, the monitoring scene database records spatial information of the monitoring scene SC1. In addition, the spatial information includes the height, width, and depth of the surveillance scene SC1. The multiple images are determined according to the spatial information of the monitoring scene, the multiple focal lengths corresponding to the multiple image capturing devices, and the multiple lens specifications corresponding to the multiple image capturing devices The position, shape and size of the respective target area. For example, for the images captured by the image capturing device 150(1) among the plurality of images, the
請再回到圖3,接著,在步驟S305中,處理器110保留於所述多個影像的所述多個目標區域內的多個目標子影像,其中於所述多個目標子影像中的多個物件影像皆為聚焦的。Please return to FIG. 3 again. Next, in step S305, the
具體來說,針對從影像擷取裝置所擷取的影像,處理器110會根據所述影像中的所辨識的目標區域來僅保留於所述目標區域內的影像資訊(如,多個像素的像素值),其中所述影像資訊可視為所述影像中的目標子影像。應注意的是,在另一實施例中,所述影像擷取裝置150(1)亦可自身辨識所述目標區域TA1。在一實施例中,處理器110會根據所述影像中的所辨識的目標區域來辨識所述目標區域之外的移除區域,以刪除在移除區域內的影像資訊。以下利用圖5A~5C、6A~6B、7A~7B、8A~8B來說明。Specifically, for the image captured from the image capturing device, the
圖5A是根據本發明的第一實施例所繪示的所擷取的對應焦段FS1的影像IMG1的示意圖。圖5B是根據本發明的第一實施例所繪示的辨識影像IMG1中的對應焦段FS1的目標區域TA1的示意圖。圖5C是根據本發明的第一實施例所繪示的保留對應焦段FS1的目標區域TA1內的目標子影像TIMG1的示意圖。FIG. 5A is a schematic diagram of the captured image IMG1 corresponding to the focal length FS1 according to the first embodiment of the present invention. FIG. 5B is a schematic diagram of the target area TA1 corresponding to the focal length FS1 in the recognition image IMG1 according to the first embodiment of the present invention. 5C is a schematic diagram of retaining the target sub-image TIMG1 in the target area TA1 corresponding to the focal length FS1 according to the first embodiment of the present invention.
請參照圖5A,舉例來說,假設影像擷取裝置150(1)已擷取影像IMG1,使用者U1位於焦段FS1內,並且使用者U1的臉部的影像AFC1為已對焦的清晰影像。在本實施例中,影像擷取裝置150(1)可根據反差檢測操作來確定當前所擷取的影像是否為已對焦(亦稱,合焦)。例如,影像擷取裝置150(1)可通過改變影像擷取裝置150(1)自身鏡頭的鏡片組的位置,在對應至不同的位置的被投射至影像擷取裝置150(1)的感光元件上的不同影像中尋找反差(對比度)最高的目標影像,並且將對應至所述目標影像的鏡片組的位置作為對焦點(亦稱,合焦點),以使用所述對焦點來擷取影像,進而可使已擷取影像IMG1中的使用者U1的臉部的影像AFC1為已對焦的清晰影像。Referring to FIG. 5A, for example, suppose that the image capturing device 150(1) has captured the image IMG1, the user U1 is located in the focal range FS1, and the image AFC1 of the face of the user U1 is a clear image that has been focused. In this embodiment, the image capturing device 150(1) can determine whether the currently captured image is in focus (also referred to as in focus) according to the contrast detection operation. For example, the image capturing device 150(1) can change the position of the lens group of the lens of the image capturing device 150(1) to correspond to different positions to be projected to the photosensitive element of the image capturing device 150(1). Find the target image with the highest contrast (contrast) among the different images on the above, and use the position of the lens group corresponding to the target image as the focus point (also known as the focus point) to use the focus point to capture the image, In turn, the image AFC1 of the face of the user U1 in the captured image IMG1 can be a clear and focused image.
請參照圖5B,處理器110可根據所述監控場景SC1的所述空間資訊的所述高度、所述寬度與所述深度中的一或多者、對應所述影像擷取裝置150(1)的所述焦段FS1以及對應所述影像擷取裝置150(1)的所述鏡頭規格來決定所述影像中的目標區域TA1的多個頂點(如,點TA1(1)~TA1(4))位於影像IMG1中的座標位置,以決定目標區域TA1的位置(如,位於影像IMG1的最下方的對應焦段FS1的位置)、形狀(如,矩形)以及多個邊長。影像IMG1中的目標區域TA1之外的區域會被視為移除區域EA1,並且位於移除區域EA1的影像不會被保留。Referring to FIG. 5B, the
請參照圖5C,在辨識影像IMG1中的目標區域TA1後,處理器110可保留目標區域TA1內的目標子影像TIMG1。影像IMG1中的移除區域EA1內的其他子影像不會被保留或被刪除。應注意的是,如上述,在所述另一實施例中,影像擷取裝置150(1)可自身辨識所述目標區域TA1。此外,影像擷取裝置150(1)可僅保留且傳送位於所述目標區域TA1內的目標子影像IMG1至主記憶體120中,進而減少主記憶體120的空間耗費。Referring to FIG. 5C, after identifying the target area TA1 in the image IMG1, the
請參照圖6A,相似地,假設於同一時間點,影像擷取裝置150(2)也擷取影像IMG2,使用者U2位於焦段FS2內,並且使用者U2的臉部的影像AFC2為已對焦的清晰影像。請參照圖6B,在辨識影像IMG2中的目標區域TA2後,處理器110可保留目標區域TA2內的目標子影像TIMG2。請參照圖7A,相似地,假設於同一時間點,影像擷取裝置150(3)也擷取影像IMG3,使用者U3位於焦段FS3內,並且使用者U3的臉部的影像AFC3為已對焦的清晰影像。請參照圖7B,在辨識影像IMG3中的目標區域TA3後,處理器110可保留目標區域TA3內的目標子影像TIMG3。請參照圖8A,相似地,假設於同一時間點,影像擷取裝置150(4)也擷取影像IMG4,使用者U4位於焦段FS4內,並且使用者U4的臉部的影像AFC4為已對焦的清晰影像。請參照圖8B,在辨識影像IMG4中的目標區域TA4後,處理器110可保留目標區域TA4內的目標子影像TIMG4。6A, similarly, assuming that at the same time point, the image capturing device 150(2) also captures the image IMG2, the user U2 is located in the focal range FS2, and the image AFC2 of the face of the user U2 is in focus Clear image. Referring to FIG. 6B, after identifying the target area TA2 in the image IMG2, the
請再回到圖3,接著,在步驟S307中,處理器110直接根據所述多個目標子影像產生對應所述時間點的單一的重建影像。具體來說,在一實施例中,處理器110根據所述多個焦段將所述多個目標子影像複製至背景影像以產生所述重建影像,其中所述多個目標子影像彼此不重合。Please return to FIG. 3 again. Next, in step S307, the
圖9是根據本發明的第一實施例所繪示的根據多個目標子影像所產生的重建影像的示意圖。請參照圖9,舉例來說,處理器110可根據所述多個焦段的順序,依序(如,由背景影像RIMG的底部開始至背景影像RIMG的頂部)將所述多個目標子影像TIMG1~TIMG4複製至背景影像RIMG(如,空白影像RIMG),以產生重建影像RIMG。值得一提的是,所述重建影像RIMG中的多個目標物件的影像AFC1~AFC4皆為已對焦的清晰影像。更詳細地說,在一實施例中,所述多個目標物件的影像AFC1~AFC4皆大於一預定像素門檻值。此外,所述預定像素門檻值是依據所述AI影像辨識裝置所執行的影像辨識操作而對應設定的。FIG. 9 is a schematic diagram of a reconstructed image generated from multiple target sub-images according to the first embodiment of the present invention. Referring to FIG. 9, for example, the
此外,在另一實施例中,處理器110可不需依序將所述多個目標子影像TIMG1~TIMG4複製至所述背景影像RIMG1,以產生所述重建影像RIMG。例如,處理器110可隨機地或利用其他規則來複製所述多個目標子影像TIMG1~TIMG4至所述背景影像RIMG1,以產生所述重建影像RIMG。In addition, in another embodiment, the
在產生所述重建影像RIMG後,接續至步驟S309,處理器110輸出所述重建影像RIMG。具體來說,處理器110可指示所述通訊電路單元130將所述重建影像RIMG經由所建立的網路連線傳送至所述AI影像辨識裝置20(如,箭頭A9所示),以使所述AI影像辨識裝置20可對所述重建影像RIMG進行影像辨識操作。After the reconstructed image RIMG is generated, step S309 is continued, and the
圖10是根據本發明的第一實施例所繪示的多個影像擷取操作的時序示意圖。請參照圖10,在時間點T1,處理器110獲得了影像擷取裝置150(1)所擷取的目標子影像TIMG1(1)、影像擷取裝置150(2)所擷取的目標子影像TIMG2(1)、影像擷取裝置150(3)所擷取的目標子影像TIMG3(1)、以及影像擷取裝置150(4)所擷取的目標子影像TIMG4(1);在時間點T2,處理器110獲得了影像擷取裝置150(1)所擷取的目標子影像TIMG1(2)、影像擷取裝置150(2)所擷取的目標子影像TIMG2(2)、影像擷取裝置150(3)所擷取的目標子影像TIMG3(2)、以及影像擷取裝置150(4)所擷取的目標子影像TIMG4(2),處理器110根據所述目標子影像IMG1(1)~IMG4(1)產生對應時間點T1的重建影像RIMG(1),並且所述重建影像RIMG(1)被傳送至所述AI影像辨識裝置20;在時間點T3,處理器110獲得了影像擷取裝置150(1)所擷取的目標子影像TIMG1(3)、影像擷取裝置150(2)所擷取的目標子影像TIMG2(3)、影像擷取裝置150(3)所擷取的目標子影像TIMG3(3)、以及影像擷取裝置150(4)所擷取的目標子影像TIMG4(3),處理器110根據所述目標子影像IMG1(2)~IMG4(2)產生對應時間點T2的重建影像RIMG(2),並且所述重建影像RIMG(2)被傳送至所述AI影像辨識裝置20;在時間點T4,處理器110獲得了影像擷取裝置150(1)所擷取的目標子影像TIMG1(4)、影像擷取裝置150(2)所擷取的目標子影像TIMG2(4)、影像擷取裝置150(3)所擷取的目標子影像TIMG3(4)、以及影像擷取裝置150(4)所擷取的目標子影像TIMG4(4),處理器110根據所述目標子影像IMG1(3)~IMG4(3)產生對應時間點T3的重建影像RIMG(3),並且所述重建影像RIMG(3)被傳送至所述AI影像辨識裝置20,以此類推。FIG. 10 is a schematic diagram of a sequence of multiple image capturing operations according to the first embodiment of the present invention. 10, at time T1, the processor 110 obtains the target sub-image TIMG1 (1) captured by the image capturing device 150(1) and the target sub-image captured by the image capturing device 150(2) TIMG2(1), the target sub-image TIMG3(1) captured by the image capturing device 150(3), and the target sub-image TIMG4(1) captured by the image capturing device 150(4); at time point T2 , The processor 110 obtains the target sub-image TIMG1(2) captured by the image capturing device 150(1), the target sub-image TIMG2(2) captured by the image capturing device 150(2), and the image capturing device The target sub-image TIMG3(2) captured by 150(3) and the target sub-image TIMG4(2) captured by the image capturing device 150(4), the processor 110 according to the target sub-image IMG1(1) ~IMG4(1) generates a reconstructed image RIMG(1) corresponding to time point T1, and the reconstructed image RIMG(1) is transmitted to the AI image recognition device 20; at time point T3, the processor 110 obtains the image capture The target sub-image TIMG1(3) captured by the capture device 150(1), the target sub-image TIMG2(3) captured by the image capture device 150(2), and the target subimage TIMG2(3) captured by the image capture device 150(3) The target sub-image TIMG3(3) and the target sub-image TIMG4(3) captured by the image capturing device 150(4), the processor 110 generates the corresponding time according to the target sub-images IMG1(2)~IMG4(2) The reconstructed image RIMG(2) at point T2, and the reconstructed image RIMG(2) is transmitted to the AI image recognition device 20; at time point T4, the processor 110 obtains the image captured by the image capturing device 150(1) The captured target sub-image TIMG1(4), the target sub-image TIMG2(4) captured by the image capture device 150(2), the target sub-image TIMG3(4) captured by the image capture device 150(3), And the target sub-image TIMG4(4) captured by the image capturing device 150(4), the processor 110 generates a reconstructed image RIMG(3) corresponding to the time point T3 according to the target sub-images IMG1(3)~IMG4(3) ), and the reconstructed image RIMG(3) is sent to the AI image recognition device 20, and so on.
也就是說,假設所述多個時間點T1~T4各自皆對應一個時間框(Time Frame),第一實施例僅需耗費兩個時間框的時間長度,便可輸出第一個重建影像給AI影像辨視裝置20。特別是,第一實施例僅需耗費一個時間框,便可輸出接續的重建影像給AI影像辨識裝置20(如圖10所示,相鄰的兩個重建影像之間的時間間隔為一個時間框)。如此一來,可證明本發明所提供的電子裝置與其所使用的影像擷取方法,相較於傳統方式,可快速且有效率地產生用以影像辨識操作的重建影像。In other words, assuming that each of the multiple time points T1 to T4 corresponds to a time frame (Time Frame), the first embodiment only needs two time frames to output the first reconstructed image to AI.
[第二實施例][Second Embodiment]
第二實施例的大部分的硬體元件相同於第一實施例,相同的硬體元件的細節不再贅述。Most of the hardware components of the second embodiment are the same as those of the first embodiment, and the details of the same hardware components will not be repeated.
圖11是根據本發明的第二實施例所繪示的影像辨識系統的方塊示意圖。請參照圖11,第二實施例與第一實施例的不同之處在於,於第二實施例中,電子裝置10僅包括了一個可變焦的影像擷取裝置150(1)。換句話說,第二實施例中的影像擷取裝置150需要快速地切換焦距來對監控場景進行拍攝。FIG. 11 is a block diagram of the image recognition system according to the second embodiment of the present invention. Referring to FIG. 11, the difference between the second embodiment and the first embodiment is that in the second embodiment, the
圖12是根據本發明的第二實施例所繪示的影像擷取方法的流程圖。請參照圖12,第二實施例所提供的影像擷取方法的流程步驟大部份相似於第一實施例所提供的影像擷取方法的流程步驟(如,步驟S1203~步驟S1209相同於步驟S303~S309),其中步驟S1201不同於步驟S301。FIG. 12 is a flowchart of the image capturing method according to the second embodiment of the present invention. 12, the process steps of the image capturing method provided in the second embodiment are mostly similar to the process steps of the image capturing method provided in the first embodiment (eg, step S1203 to step S1209 are the same as step S303 ~S309), where step S1201 is different from step S301.
在步驟S1201中,經由可變焦的影像擷取裝置,根據多個焦距依序對監控場景連續地進行多次拍攝,以擷取對應所述多個焦距的多個影像。In step S1201, through the zoomable image capturing device, the surveillance scene is sequentially shot multiple times according to multiple focal lengths to capture multiple images corresponding to the multiple focal lengths.
具體來說,為了方便說明,假設影像擷取裝置150具有焦距FL1~焦距FL4(如圖4A所示)。影像擷取裝置150可根據焦距FL1~焦距FL4的順序先利用焦距FL1對監控場景SC1進行拍攝以獲得影像IMG1、再使用焦距FL2對監控場景SC1進行拍攝以獲得影像IMG2、再使用焦距FL3對監控場景SC1進行拍攝以獲得影像IMG3、最後再使用焦距FL4對監控場景SC1進行拍攝以獲得影像IMG4。此外,處理器110可根據對應的焦段FS1~FS4相應地獲得目標子影像TIMG1~TIMG4(步驟S1203~1205),並且根據所述目標子影像TIMG1~TIMG4產生且輸出對應的重建影像RIMG(步驟S1207~1209)。Specifically, for the convenience of description, it is assumed that the
圖13A與圖13B是根據本發明的第二實施例所繪示的多個影像擷取操作的時序示意圖。請參照圖13A,在時間點T1,處理器110獲得了影像擷取裝置150所擷取的對應焦段FS1的目標子影像TIMG1(1)、在時間點T2,處理器110獲得了影像擷取裝置150所擷取的對應焦段FS2的目標子影像TIMG2(1)、在時間點T3,處理器110獲得了影像擷取裝置150所擷取的對應焦段FS3的目標子影像TIMG3(1)、在時間點T4,處理器110獲得了影像擷取裝置150所擷取的對應焦段FS4的目標子影像TIMG4(1);在時間點T5,處理器110根據所述目標子影像IMG1(1)~IMG4(1)產生重建影像RIMG(1),並且處理器110獲得了影像擷取裝置150所擷取的對應焦段FS1的目標子影像TIMG1(2)、在時間點T6,處理器110獲得了影像擷取裝置150所擷取的對應焦段FS2的目標子影像TIMG2(2)、在時間點T7,處理器110獲得了影像擷取裝置150所擷取的對應焦段FS3的目標子影像TIMG3(2)、在時間點T8,處理器110獲得了影像擷取裝置150所擷取的對應焦段FS4的目標子影像TIMG4(2);在時間點T9,處理器110根據所述目標子影像IMG1(2)~IMG4(2)產生重建影像RIMG(2),並且處理器110獲得了影像擷取裝置150所擷取的對應焦段FS1的目標子影像TIMG1(3)、在時間點T10,處理器110獲得了影像擷取裝置150所擷取的對應焦段FS2的目標子影像TIMG2(3)、在時間點T11,處理器110獲得了影像擷取裝置150所擷取的對應焦段FS3的目標子影像TIMG3(3)、在時間點T12,處理器110獲得了影像擷取裝置150所擷取的對應焦段FS4的目標子影像TIMG4(3);在時間點T9,處理器110根據所述目標子影像IMG1(3)~IMG4(3)產生重建影像RIMG(3),以此類推。請參照圖13B,所產生的重建影像RIMG(1)~RIMG(3)可在時間點T5、T9、T13分別被輸出至AI影像辨識裝置20。FIGS. 13A and 13B are schematic diagrams of the sequence of multiple image capturing operations according to the second embodiment of the present invention. 13A, at time T1, the
也就是說,假設所述多個時間點T1~T4各自皆對應一個時間框(Time Frame),第二實施例僅需耗費四個時間框,便可輸出一個重建影像給AI影像辨識裝置20。雖然第二實施例產生重建影像的速度為第一實施例的四分之一,但是第二實施例中用以設置影像擷取裝置的硬體成本也為第一實施例的四分之一。在整體上,不論第一或第二實施例,相較於傳統做法,皆可以節省用於擷取合格影像的大量的時間與資源,進而增進了影像辨識系統的整體效率。In other words, assuming that each of the multiple time points T1 to T4 corresponds to a time frame, the second embodiment only needs four time frames to output a reconstructed image to the AI
綜上所述,本發明的實施例所提供的影像擷取方法與電子裝置,可根據多個焦距對監控場景進行拍攝,以擷取多個影像,根據分別對應所述多個焦距的多個焦段來保留所述多個影像各自的目標子影像,以直接根據所述多個目標子影像來產生且輸出重建影像。如此一來,可在不分析所述多個影像及不貼合/縫合所述多個目標子影像的情況下,快速且有效率地使所產生的所述重建影像中位於多個深度的多個物件影像皆為對焦且清晰的,進而增進使用所述重建影像的影像辨識操作的精確度。In summary, the image capturing method and electronic device provided by the embodiments of the present invention can shoot a monitoring scene according to multiple focal lengths to capture multiple images, according to the multiple focal lengths corresponding to the multiple focal lengths. The focal length retains the target sub-images of each of the multiple images, so as to directly generate and output reconstructed images based on the multiple target sub-images. In this way, without analyzing the multiple images and not fitting/stitching the multiple target sub-images, multiple images located at multiple depths in the generated reconstructed image can be quickly and efficiently made. Each object image is focused and clear, thereby improving the accuracy of the image recognition operation using the reconstructed image.
雖然本發明已以實施例揭露如上,然其並非用以限定本發明,任何所屬技術領域中具有通常知識者,在不脫離本發明的精神和範圍內,當可作些許的更動與潤飾,故本發明的保護範圍當視後附的申請專利範圍所界定者為準。Although the present invention has been disclosed in the above embodiments, it is not intended to limit the present invention. Anyone with ordinary knowledge in the relevant technical field can make some changes and modifications without departing from the spirit and scope of the present invention. The protection scope of the present invention shall be subject to those defined by the attached patent application scope.
SC1:監控場景 FC1~FC4、AFC1~AFC4:目標物件的影像 U1~U4:使用者 1:影像辨識系統 10:電子裝置 20:AI影像辨識裝置 110、210:處理器 120、220:主記憶體 130、230:通訊電路單元 140、240:儲存裝置 141:監控場景資料庫 142:鏡頭規格資料庫 143:程式碼 150:影像擷取裝置陣列/影像擷取裝置 150(1)~150(N):影像擷取裝置 S301、S303、S305、S307、S309:第一實施例與第二實施例的影像擷取方法的流程步驟 S1701、S1703、S1705、S1707、S1709:第二實施例的影像擷取方法的流程步驟 FL1~FL4:焦距 FS1~FS4:焦段 FS1_F~FS4_F、FS1_B~FS4_B:距離 FVL:視野深度 TA1~TA4:目標區域 SA1:拍攝區域 EA1~EA4:移除區域 TA1(1)~TA(4):頂點 TIMG1~TIMG4、TIMG1(1)~TIMG1(4)、TIMG2(1)~TIMG2(4)、TIMG3(1)~TIMG3(4)、TIMG4(1)~TIMG4(4):目標子影像 RIMG、RIMG(1)~RIMG(4):重建影像 T1~T13:時間點SC1: Monitoring scene FC1~FC4, AFC1~AFC4: the image of the target object U1~U4: User 1: Image recognition system 10: Electronic device 20: AI image recognition device 110, 210: processor 120, 220: main memory 130, 230: communication circuit unit 140, 240: storage device 141: Surveillance Scene Database 142: Lens Specification Database 143: Code 150: Image capture device array/image capture device 150(1)~150(N): image capture device S301, S303, S305, S307, S309: the process steps of the image capturing method of the first embodiment and the second embodiment S1701, S1703, S1705, S1707, S1709: process steps of the image capturing method of the second embodiment FL1~FL4: Focal length FS1~FS4: focal length FS1_F~FS4_F, FS1_B~FS4_B: distance FVL: Depth of Field TA1~TA4: target area SA1: shooting area EA1~EA4: Remove area TA1(1)~TA(4): vertex TIMG1~TIMG4, TIMG1(1)~TIMG1(4), TIMG2(1)~TIMG2(4), TIMG3(1)~TIMG3(4), TIMG4(1)~TIMG4(4): target sub-image RIMG, RIMG(1)~RIMG(4): Rebuild the image T1~T13: time point
圖1是根據本發明的第一實施例所繪示的監控場景的示意圖。 圖2是根據本發明的第一實施例所繪示的影像辨識系統的方塊示意圖。 圖3是根據本發明的第一實施例所繪示的影像擷取方法的流程圖。 圖4A是根據本發明的第一實施例或第二實施例所繪示的對應多個焦距的多個焦段的示意圖。 圖4B是根據本發明的第一實施例所繪示的監控場景的拍攝區域與對應的多個焦段的示意圖。 圖5A是根據本發明的第一實施例所繪示的所擷取的對應焦段FS1的影像IMG1的示意圖。 圖5B是根據本發明的第一實施例所繪示的辨識影像IMG1中的對應焦段FS1的目標區域TA1的示意圖。 圖5C是根據本發明的第一實施例所繪示的保留對應焦段FS1的目標區域TA1內的目標子影像TIMG1的示意圖。 圖6A是根據本發明的第一實施例所繪示的所擷取的對應焦段FS2的影像IMG2的示意圖。 圖6B是根據本發明的第一實施例所繪示的保留對應焦段FS2的目標區域TA2內的目標子影像TIMG2的示意圖。 圖7A是根據本發明的第一實施例所繪示的所擷取的對應焦段FS3的影像IMG3的示意圖。 圖7B是根據本發明的第一實施例所繪示的保留對應焦段FS3的目標區域TA3內的目標子影像TIMG3的示意圖。 圖8A是根據本發明的第一實施例所繪示的所擷取的對應焦段FS4的影像IMG4的示意圖。 圖8B是根據本發明的第一實施例所繪示的保留對應焦段FS4的目標區域TA4內的目標子影像TIMG4的示意圖。 圖9是根據本發明的第一實施例所繪示的根據多個目標子影像所產生的重建影像的示意圖。 圖10是根據本發明的第一實施例所繪示的多個影像擷取操作的時序示意圖。 圖11是根據本發明的第二實施例所繪示的影像辨識系統的方塊示意圖。 圖12是根據本發明的第二實施例所繪示的影像擷取方法的流程圖。 圖13A與圖13B是根據本發明的第二實施例所繪示的多個影像擷取操作的時序示意圖。Fig. 1 is a schematic diagram of a monitoring scene according to the first embodiment of the present invention. FIG. 2 is a block diagram of the image recognition system according to the first embodiment of the present invention. FIG. 3 is a flowchart of the image capturing method according to the first embodiment of the present invention. 4A is a schematic diagram of multiple focal lengths corresponding to multiple focal lengths according to the first embodiment or the second embodiment of the present invention. 4B is a schematic diagram of a shooting area and corresponding multiple focal lengths of a surveillance scene according to the first embodiment of the present invention. FIG. 5A is a schematic diagram of the captured image IMG1 corresponding to the focal length FS1 according to the first embodiment of the present invention. FIG. 5B is a schematic diagram of the target area TA1 corresponding to the focal length FS1 in the recognition image IMG1 according to the first embodiment of the present invention. 5C is a schematic diagram of retaining the target sub-image TIMG1 in the target area TA1 corresponding to the focal length FS1 according to the first embodiment of the present invention. 6A is a schematic diagram of the captured image IMG2 corresponding to the focal length FS2 according to the first embodiment of the present invention. 6B is a schematic diagram of retaining the target sub-image TIMG2 in the target area TA2 corresponding to the focal length FS2 according to the first embodiment of the present invention. FIG. 7A is a schematic diagram of the captured image IMG3 corresponding to the focal length FS3 according to the first embodiment of the present invention. FIG. 7B is a schematic diagram of retaining the target sub-image TIMG3 in the target area TA3 corresponding to the focal length FS3 according to the first embodiment of the present invention. FIG. 8A is a schematic diagram of the captured image IMG4 corresponding to the focal length FS4 according to the first embodiment of the present invention. FIG. 8B is a schematic diagram of retaining the target sub-image TIMG4 in the target area TA4 corresponding to the focal length FS4 according to the first embodiment of the present invention. FIG. 9 is a schematic diagram of a reconstructed image generated from multiple target sub-images according to the first embodiment of the present invention. FIG. 10 is a schematic diagram of a sequence of multiple image capturing operations according to the first embodiment of the present invention. FIG. 11 is a block diagram of the image recognition system according to the second embodiment of the present invention. FIG. 12 is a flowchart of the image capturing method according to the second embodiment of the present invention. FIGS. 13A and 13B are schematic diagrams of the sequence of multiple image capturing operations according to the second embodiment of the present invention.
S301、S303、S305、S307、S309:影像辨識方法的流程步驟S301, S303, S305, S307, S309: Process steps of image recognition method
Claims (12)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW109101963A TW202129597A (en) | 2020-01-20 | 2020-01-20 | Image capturing method and electronic device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
TW109101963A TW202129597A (en) | 2020-01-20 | 2020-01-20 | Image capturing method and electronic device |
Publications (1)
Publication Number | Publication Date |
---|---|
TW202129597A true TW202129597A (en) | 2021-08-01 |
Family
ID=78282730
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW109101963A TW202129597A (en) | 2020-01-20 | 2020-01-20 | Image capturing method and electronic device |
Country Status (1)
Country | Link |
---|---|
TW (1) | TW202129597A (en) |
-
2020
- 2020-01-20 TW TW109101963A patent/TW202129597A/en unknown
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108898567B (en) | Image noise reduction method, device and system | |
WO2018201809A1 (en) | Double cameras-based image processing device and method | |
JP6682550B2 (en) | System and method for continuous autofocus | |
CN102891960B (en) | Method and camera for determining an image adjustment parameter | |
JP2010021943A (en) | Imaging apparatus | |
US8942509B2 (en) | Apparatus and method creating ghost-free high dynamic range image using filtering | |
WO2023024697A1 (en) | Image stitching method and electronic device | |
CN1971398A (en) | Imaging device and image processing method | |
CN111629146B (en) | Shooting parameter adjusting method, shooting parameter adjusting device, shooting parameter adjusting equipment and storage medium | |
WO2014187265A1 (en) | Photo-capture processing method, device and computer storage medium | |
CN104754227A (en) | Method and device for shooting video | |
WO2008147724A2 (en) | Methods, systems and apparatuses for motion detection using auto-focus statistics | |
JP4640198B2 (en) | Electronic camera, multi-point simultaneous focusing frame display method, and program | |
TW202236840A (en) | Image fusion for scenes with objects at multiple depths | |
US8681235B2 (en) | Apparatus for processing digital image signal that obtains still image at desired point in time and method of controlling the apparatus | |
WO2016082590A1 (en) | Picture processing method and apparatus, and storage medium | |
JP2003179798A (en) | Digital camera | |
CN110855957B (en) | Image processing method and device, storage medium and electronic equipment | |
TW202129597A (en) | Image capturing method and electronic device | |
JP2011090410A (en) | Image processing apparatus, image processing system, and control method of image processing apparatus | |
US11343424B1 (en) | Image capturing method and electronic device | |
JP2018007156A (en) | Image search device, image search method, and program | |
CN116051736A (en) | Three-dimensional reconstruction method, device, edge equipment and storage medium | |
CN115701121A (en) | Image acquisition method and electronic device | |
JP2007266664A (en) | Thumbnail sorter and imaging apparatus |