M373507 五、新型說明： 【新型所屬之技術領域】 本創作與全景拼接裝置有， 像的全景拼接裝置。 θ一種立體影 【先前技術】 現有監控裝置大都以單一固定. 雖偵測，對移動目標物體^攝二= :施間單追踨，無法持續追踨所有移2二以面 =要追踨其他目標物體時，則必須 : 巧景’搜尋其他目標物體，如此= 間，—貝訊不足，只能得到局部影像晝面。、。夕時 習知一種由廣角鏡頭搭配攝 周圍360度視野，雖可取得監控環境之二景二1摘取 是在影像邊緣產生失真缺點及晝面解戶、二|各， 像差，況產生’需要另外搭配搖擺俯又 影像。要接出全景影像’傳統的作 ：: ，投影到某一共同曲面，然後】 = = 若是利用傳統單一攝影機進行相以 以進行影像拼接完成全景重建可得到一 晝面，但缺乏3D景深資訊；使用多A =王景 =監控的範圍’但是由於每台攝影‘的；野I:: 同且獨立，因此也衍生出多台攝影不 ，問題’因此必須解決可見視野有^ 4 ^ 育訊及晝面清晰的問題。 u'7、/衣 此鑑於習知存在上述之缺點，創作人乃針對ϋ =缺點研究改進，並提出一種立體（3D)視覺= 接裝置，利用立體攝影機擷取場景影像，得到^ ^ 3 M373507 =、景深資訊的立體影像，之後再以全景拼接影像處 進行全景影像的接合，進而達到3D場景建構的 ^果。透過本裝置所見夠的3D場景，使用者可連 ^取得目標物影像之具有立體景深且清晰可識別影 此*亦即具有場景功效’彌補習用技術的各項 失。 【新型内容】 與入^ 了達到上述之目的，本創作提供一種立體視 ίίίϊίϊ置，包括二攝影機，其中該二攝影機 屬於凝視型雙眼視覺系統。 ^ 7述的裝置，其中所述凝視型雙眼視覺系 、’’疋心該一攝影機的光軸交會於同一點。 虚述的裝置，更包含一訊號處理器，用以 處理來自該凝視型雙眼視覺系統的訊號。 =上所述的裝置，其中該二攝影機係左右設 ^左左右影像’而該訊號處理器則接收 ζ Μ 透過該左右影像產生用以校正攝影機 鏡頭的鏡頭校正參數。 将心预 >Α如上所述的裝置，其中該左右影像包含一左A 像與一右影像。 /丨私匕3左衫 通姑述的裝置，其中該訊號處理器透過該鏡 頭校f參數將該左右影像合成為-立體影ί 鏡 至該以2ί裝置，係有複數個該立體影像輸入 影像合全ί拼=號處理器將該複數個立體 選自的裝置，其中該攝影機的感光元件係 體（CMOS)中的一種。 飞金屬乳化物半導 為了達到上述之目的，本創作 視覺全景拼接裝置，句括一怒」^再棱供一種立體 罝匕括凝視型立體攝影系統； 4 M373507 =及了訊號，理器，與該凝視型立體攝影系統電連 ，並接收來自該凝視型攝影系統的初步赘像。 2所=襄置，，中該凝視型立體攝影系 統，包括一左攝影機與一右攝影機，且 的光軸與該右攝影機的光軸交會。 左攝〜機 傻鱼如斤ί的裝置，其中該初步影像包含-左影 像與一右影像。 與傻裝Λ’其中該訊號處理器透過該左 〜象_該右汾像，未付該攝影系統的鏡頭校正來數。 如前所述的裝置，其中該訊號處理器將‘左卑 2 ΐ:右影像透過該鏡頭校正參數合成而得到-： 如前所述的裝置，其中該訊號處理器 2處=為一立體影像’並再將 ：；像 接合而成為一全景影像。 m〜像 視*述之目❾’本創作更提出一種立體 ，境，用以建構3〇全景立體 55】置、一訊號傳輸裝置及監控中：。 請參閱圖1，為本創作的裝置示音^^ & 的立體^見覺全景拼接裝置大致上包括二立體岑 攝影機12更包含—左攝㈣12匕 右攝衫機12R，以模擬人眼的視 2 言本創作的立體攝影機 一步而 影設傷，故左攝影機12L與=體攝 軸交會於一點疋以此兩者的光 對焦之所在。 ”’卩為此一攝影機 5 M373507 請繼續參閱圖I，！ 攝影機12之外，尚勺义、一T本創作除了包含一立體 理來自左攝影機i2°LH一訊號處理器14，用以處 立體攝影機12拍攝的〃、攝影機12R訊號，意即由 左攝影機12L所拍攝的初步^景^象’此初步影像包含了 拍攝的右影像，而 :像’與右攝影機12JR所 器Μ所使用來求出用工影像即被訊號處理 數，訊號處理器正鏡頭用的鏡頭校正參 體（m影像1〇且亦由1左入1像^右影像合成為一立 圍，立體攝影機12场11是相當大的範 是以多部立體攝影機而f2朝來不拍同摄的：向拍攝，或甚至 更可進-步的將複=立顯體；象一 13上，且 接影像，並透過一傳輸f 7^ 10 s成一全景拼 至監控巾4、16。輸裝置15將全景拼接影像傳輸 言月參閱圖2 ’為本創作的運作 程大致上分為兩個階段，第一階】=二= 成部分，第二階段則是全景拼形=八/ 此漭葙昔沬β本碰，了 ^饮办1豕的开> 成部分。 ϊ 藉由左攝影· 12[與右攝影 2的仟if像與右影像求出用以校正鏡 ?立：2 影像與右影像經過影像匹配而形成 盘古，二通常此一影像匹配的過程是將左影像 像技衫至同一場景以進行匹配。在步驟2之 後疋^驟3,將複數個立體影像接合而成為一全景 拼接影像（Panoramic Image Mosaics)。因砼士功 μ 二 言’步驟！、2屬於第一階段，而步以以2 f二此外，在步驟3中’所述的複數個立體影像是 :系列的影像，亦即以一特定範圍的場景為主的連 哨拍攝的影像。該系列的每張影像都經過同一點光M373507 V. New description: [New technical field] The creation and panoramic splicing device has a panoramic splicing device. θ A kind of stereoscopic image [Prior Art] Most of the existing monitoring devices are fixed by a single. Although the detection, moving the target object ^ 2 =: Shi Jian single tracking, can not continue to track all the movements 2 to the surface = to track other When the target object is used, it must be: Qiaojing 'search for other target objects, so = between, - the lack of Beixun, can only get the partial image. ,. On the eve of the evening, a wide-angle lens is used to take a 360-degree view of the surrounding area. Although the surveillance environment can be obtained, the second scene is extracted. The distortion is caused by the distortion at the edge of the image, and the aberrations are generated. Also with the swing and image. To take out the panoramic image 'traditional work::, project to a common surface, then】 = = If you use a traditional single camera to perform image mosaic to complete the panoramic reconstruction, you can get a face, but lack 3D depth information; Use more A = Wang Jing = range of monitoring 'but because of each photography'; wild I:: same and independent, so also derived from multiple photography, the problem 'Therefore, we must solve the visible field of view ^ 4 ^ A clear question. U'7, / clothing Because of the above-mentioned shortcomings of the above, the creator is aimed at ϋ = shortcomings research improvement, and proposed a stereo (3D) visual = connection device, using a stereo camera to capture the scene image, get ^ ^ 3 M373507 =, the stereoscopic image of the depth of field information, and then the panoramic image is joined by the panoramic stitching image to achieve the 3D scene construction. Through the 3D scenes seen by the device, the user can obtain the stereoscopic depth of field of the target image and clearly recognize the shadow*, which has the effect of the scene' to make up for the loss of the conventional technology. [New Content] In order to achieve the above objectives, the present invention provides a stereoscopic image, including two cameras, wherein the two cameras belong to a gaze binocular vision system. The device described in the above, wherein the gaze-type binocular vision system, ‘the heart of the camera, intersects at the same point. The imaginary device further includes a signal processor for processing signals from the gaze binocular vision system. = The device described above, wherein the two cameras are provided with left and right images left and right and the signal processor receives 镜头 产生 through the left and right images to generate lens correction parameters for correcting the camera lens. The device is as described above, wherein the left and right images include a left A image and a right image. / 丨 匕 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左 左The device is a device selected from a plurality of stereoscopic devices, one of which is a photosensitive element system (CMOS) of the camera. In order to achieve the above-mentioned purpose, the three-dimensional splicing device of the present invention provides a three-dimensional gaze-type stereo photography system; 4 M373507 = and the signal, the device, and The gaze-type stereo camera system is electrically connected and receives preliminary artifacts from the gaze-type photography system. The gaze-type stereo photography system includes a left camera and a right camera, and the optical axis intersects the optical axis of the right camera. Left camera ~ machine silly fish as a device, the preliminary image contains - left image and a right image. And the stupid device, wherein the signal processor passes the left-to-right image of the right image, and the lens of the camera system is not corrected. The device as described above, wherein the signal processor obtains 'left 2 2 ΐ: the right image is synthesized through the lens correction parameter -: the device as described above, wherein the signal processor 2 is a stereo image 'And then:; like a joint into a panoramic image. m~ Image 视 * 目 目 目 ' This creation also proposes a three-dimensional, environment, used to construct 3 〇 panoramic stereo 55] set, a signal transmission device and monitoring:. Referring to FIG. 1 , the stereoscopic splicing device of the device for displaying the sound of the device ^^ & generally comprises two stereo cameras 12 further comprising a left camera (four) 12 匕 right camera 12R to simulate the human eye. Depending on the 2nd camera, the stereo camera is wounded in one step, so the left camera 12L and the body camera axis meet at a point, and the light of both is focused. "'For this camera 5 M373507 Please continue to refer to Figure I!! In addition to the camera 12, the original creation, in addition to a stereoscopically from the left camera i2 °LH a signal processor 14, used to stereo The camera 12R signal captured by the camera 12, that is, the preliminary image taken by the left camera 12L, this preliminary image contains the right image of the shot, and: like the use of the right camera 12JR The image of the work is processed by the signal, and the lens of the signal processor is used to correct the lens (m image 1〇 is also composed of 1 left into 1 image ^ right image into a vertical circumference, stereo camera 12 field 11 is quite large The fan is a multi-camera camera and the f2 is not taking the same photo: the shooting, or even the more step-by-step will be the complex; the image is like a 13 and connected to the image, and through a transmission f 7^ 10 s into a panoramic view to the monitoring towel 4, 16. The transmission device 15 will transmit the panoramic mosaic image. See Figure 2 'The operation process of this creation is roughly divided into two stages, the first order] = two = Part, the second stage is a panoramic mosaic = eight / this time 沬 β Touch, the opening of the drink is 1 part. 藉 By left photography · 12 [with the right image of the 仟if image and the right image to find the correct lens: 2 image and right image through the image Matching to form Pangu, two usually the process of image matching is to match the left image to the same scene to match. After step 2, step 3, the plurality of stereo images are joined to form a panoramic mosaic image (Panoramic) Image Mosaics). Because of the gentleman's work, the two steps 'step!, 2 belong to the first stage, and the step is to 2 f 2, in addition, the multiple stereo images described in step 3 are: series of images, ie An image of a whistle based on a specific range of scenes. Each image of the series passes through the same light.