505820 經濟部智慧財產局員工消費合作社印製 A7 B7 五、發明說明（1 ) 發明背景 1.發明領域 本發明關於影像系統，特別是有關具有水平搖攝，垂直 搖攝及變焦調整能力之照相機自動化。 2 .相關技藝敘述 當照相機在捕獲一影像時，照相機觀視領域中物件之三 維座標係映射爲該物件之影像二維座標。物件或"世界” (world )座標映射爲影像座標可運用於多種用途，其可控 制照相機之觀視範圍以包含特定之世界座標，或決定以影 像座標爲基礎之世界座標。舉例而言，桑哥塔（S.. Sengupta)，黎昴（D. Lyons)以及墨菲（T. Murphy)於 1997 年12月31曰錄案之第23,353號委任備審案（Attorney docket)，US第09/002,105號共同專利申請案，名稱爲”複 數個照相機系統中用以選取並調整一照相機之最佳使用者 界面"（An Optimized User Interface for Selecting and Adjusting a Camera in a Multiple Camera System)，其揭示 一使用者界面，其中使用者在例如地平面之顯示影像上選 定一位置，對應至此選定動作，一控制器調整一或多個照 相機，因此其觀視範圍包括對應至所選定位置之世界座 標，該申請案爲本案説明·書所參考列舉者。同樣的，桑哥 塔，黎昂，墨菲及里茲（D尺⑶代^於1998年8月7日錄案之 第PHA 23,489號委任備審案，US第09/131,243號共同專利 申請案，名稱爲π複數個照相機系統中用於圖像追蹤之自 動化照相機離子系統"（An Automated Camera Handoff -4- 本纸張尺度適用中國國家標準（CNS)A4規格（210 χ 297公釐） ---·---^--------------訂*--------- (請先閱讀背面之注意事項再填寫本頁) 505820 A7 五、發明說明（2 System for Figure Tracking in a Multiple Camera System) ^ 其揭示一技·術，該技術藉由世界座標爲基礎自動控制一照 相機而自動追蹤一個人或物件，此時該個人或物件由一照 相機觀視範圍移至另一照相機觀視範圍，世界座標係衍生 自另一照相機觀視範圍之影像，該申請案爲本案說明書所 參考列舉者。 爲了在世界座標與影像座標之間提供照相機所執行之正 確映射模式，其須具有代表照相機設定.與捕獲影像之間關 係的參數，最常使用之模式爲投射照相機模式且其數學程 式爲： sm = Q(RP + T), (1) 其中m表示世界點Ρ之影像座標，S表示投射尺寸係數，Q 爲内郅照相機校正矩陣，以及R及了爲校準照相機及座標 系統之轉動及平移。爲了利用照相機在世界點與影像點之 間量測，其須決定矩陣Q且一般術語爲，，照相機校正"。矩 陣Q之一般程式爲： Q. f 〇 W 〇 / y〇 0 0.1 (2) (請先閱讀背面之注意事項再填寫本頁) ------^—訂--------線. 握濟部智慧財產局員工消費合作社印製 其中（x-0，y〇)照相機影〖象面上光學中心或."主點"，£表示 焦矩。爲便於瞭解，其已顯示共同之X及y焦距，即矩陣^ 第 列中f "可表示爲"f X ’’，X焦距，第二列中"£ ”可表示 馬，y焦距。照相機之光學中心爲照相機光學軸線與 照相機映射平面之交叉點，矩睁Q之參數通常表示爲映像 -5- 本紙張尺度適用中關家標準（CNS)A4規格⑵G χ 29?公楚 經濟部智慧財產局員工消費合作社印製 505820 A7 -------___Β7 五、發明說明（3 ) 照相機之校正通常藉由不同之照相機設定蒐集資料，之 後處理孩資料以決定或測定照相機之光學中心或焦距。若 ”、、相機/、有’4焦说力’該方法重複以決定每個變焦設定之 對應焦距。藉由鏺別參考點之3_D座標以建立一校正標的 物’配合不同之照相機而決定照相機影像中參考點之對應 座標。曲線密合技術係用以決定上述方程式（i)及（2)之 參數’其基礎在於映射照相機影像座標及每個參考點之實 際3_D座標。1999年7月27日錄案之us第5,930,740號專 利揭示一照相機校正方法，其步驟包括建立3 _ D空間中參 考點以及三角測量相對於參考點之照相機位置及方位。在 該參考專利中，一資料庫使照相機設定對應著不同之照相 機位置及/或方位映射。 爲了校正變焦照相機，其須判定改變焦距之控制能力與 此變化所產生在影像之間的關係。舉例而言，若利用步進 馬達改變照相機中透鏡間距以增加或縮小有效焦距，馬達 單一步進與焦距變化總量之間爲非線性關係。校正提供一 演算映射以決定步驟馬達或其他裝置每個動作時照相機之 焦距’該裝置可改變照相機之焦距。爲便於參考，此處所 引用之一般性術語，•變焦·設定，，係用以控制照相機，其矸 提供重複之變焦調整。雖然在校正前不知道焦距，"7之 變焦設定"將重複的提供相同之焦距。一般而言，照相機 之變焦設定與焦距成反比。 例如變焦設定，，8，，之焦距小於變進係數7 (較小之倍 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） . ^ ------^-I -------- (請先閱讀背面之注意事項再填寫本頁) 505820505820 Printed by A7 B7, Consumer Cooperatives, Intellectual Property Bureau, Ministry of Economic Affairs 5. Description of the Invention (1) Background of the Invention 1. Field of the Invention The present invention relates to imaging systems, and in particular, to camera automation with horizontal pan, vertical pan and zoom adjustment capabilities . 2. Relevant technical description When the camera is capturing an image, the three-dimensional coordinate of the object in the field of view of the camera is mapped to the two-dimensional coordinate of the image of the object. Objects or "world" coordinates are mapped to image coordinates and can be used for a variety of purposes. It can control the camera's viewing range to include specific world coordinates, or determine world coordinates based on image coordinates. For example, Attorney docket (S .. Sengupta), D. Lyons, and T. Murphy, No. 23,353, recorded on December 31, 1997, US No. 09 / 002,105 common patent application, entitled "An Optimized User Interface for Selecting and Adjusting a Camera in a Multiple Camera System", which Reveal a user interface in which the user selects a position on a display image on the ground plane, for example, and a controller adjusts one or more cameras, so its viewing range includes world coordinates corresponding to the selected position This application is the one listed in the description and book of this case. Similarly, Sangota, Leon, Murphy, and Leeds (D & D No. PHA No. 23,489, filed on August 7, 1998 and filed for trial, US Patent No. 09 / 131,243 , Named "π Automated Camera Handoff System for Image Tracking in π Multiple Camera Systems" (An Automated Camera Handoff -4- This paper size applies to the Chinese National Standard (CNS) A4 specification (210 χ 297 mm)- -· --- ^ ---------- Order * --------- (Please read the notes on the back before filling this page) 505820 A7 V. Invention Description (2 System for Figure Tracking in a Multiple Camera System) ^ It reveals a technique that automatically tracks a person or object by automatically controlling a camera based on world coordinates. At this time, the person or object is viewed by a camera. The view range is moved to another camera view range, and the world coordinate is an image derived from the view range of another camera. This application is the reference listed in the description of this case. In order to provide the camera between the world coordinate and the image coordinate, the implementation is performed. Correct mapping mode, which must have a representative camera The relationship between the captured image and the captured image. The most commonly used mode is the projection camera mode and its mathematical formula is: sm = Q (RP + T), (1) where m represents the image coordinates of the world point P, and S represents Projection size coefficient, Q is the intrinsic camera correction matrix, and R is the rotation and translation of the camera and the coordinate system. In order to use the camera to measure between the world point and the image point, it must determine the matrix Q and the general term is The camera's general formula for matrix Q is: Q. f 〇W 〇 / y〇0 0.1 (2) (Please read the precautions on the back before filling in this page) ------ ^ — Order -------- line. Printed by the Consumer Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs, where (x-0, y〇) the camera image 〖Optical Center or. &Quot; Principal Point " on the image plane, £ means focus For easy understanding, it has shown the common X and y focal lengths, that is, f " in the matrix ^ column can be expressed as " f X '', X focal length, " £ '' in the second column can represent horses, y focal length. The optical center of the camera is the intersection of the camera optical axis and the camera mapping plane. The number is usually expressed as the image -5- This paper standard is applicable to the Zhongguanjia Standard (CNS) A4 specification ⑵G χ 29? Printed by the Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Public Economy and Economics 505820 A7 -------___ B7 V. Invention Explanation (3) Camera calibration usually collects data by different camera settings, and then processes the data to determine or measure the optical center or focal length of the camera. If "", "camera", "4 focal power", the method is repeated to determine the corresponding focal length of each zoom setting. The 3_D coordinate of the reference point is used to establish a calibration target, and the camera is determined with different cameras. Correspondence coordinates of reference points in the image. The curve close technology is used to determine the parameters of the above equations (i) and (2). The basis is to map the camera image coordinates and the actual 3_D coordinates of each reference point. July 27, 1999 Japanese Patent No. 5,930,740 discloses a camera calibration method, the steps of which include establishing a reference point in 3-D space and triangulating the camera position and orientation relative to the reference point. In this reference patent, a database makes the camera The settings correspond to different camera position and / or orientation mappings. In order to correct a zoom camera, it is necessary to determine the relationship between the control ability to change the focal length and the image generated by the change. For example, if a stepper motor is used to change the camera The lens pitch is used to increase or decrease the effective focal length. There is a non-linear relationship between the single step of the motor and the total amount of focal length change. Correction Provides a calculation map to determine the focal length of the camera at each action of the step motor or other device. This device can change the focal length of the camera. For ease of reference, the general terms cited here, • Zoom · Setting, are used to control the camera It provides repeated zoom adjustments. Although the focal length is not known before correction, "7's zoom settings" will repeatedly provide the same focal length. Generally speaking, the camera's zoom setting is inversely proportional to the focal length. For example, the zoom setting, ， 8 ， ， The focal length is less than the factor of 7 (smaller multiples. The paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm). ^ ------ ^-I ---- ---- (Please read the notes on the back before filling this page) 505820
五、發明說明（4 率），變焦設定” 6 "具有較大之焦距。對照相機系統而言 具-有大量之變焦設定，故有利於焦距之微調。其並麥決定 及儲存每個可能變焦設定之觀視範園，一演算法常用以數 學方式決定每個可能變焦設定之焦距。 爲判足照相機變焦設定與對應焦距之間的演算映射，其 蒐集一組z(i)及f(z(i))之量測。對每個變焦設定z⑴而言， 其與上述校正技術範例決定對應之焦距値f«切。由於焦 距與變焦設定爲反比（較大之變焦設定對應著較短之隹 距），用以決定變焦設定對應焦距之傳統模式方程爲： f(Z) = a0 1 + alz + alz2 + α3ζ3 (3) 方私式（3)刀母中的多.項式之次數通常是未知的，經驗顯 示2次方對大部份之照相機變焦構造爲最佳的。最適合照 相機量測値Z(i)，f(z(i))之係數aG，uU由物件函數最小 化加以決定： (4) .y 二 : ·%·. » r* * « - 5 .1 Γ · ·♦··- · · i T b iII·I—丨1丨丨丨丨^|^—— (請先閱讀背面之注意事項再填寫本頁) -- ·*···* i· * - · *-: i :. ·.:**··:.·*. 1 :Γί .:*·. -—^-------l 經濟部智慧財產局員工消費合作社印製 然而:此項焦距量測之直接判定仍有問題存在。如方程式 件函數爲非線性的且叠代法須使該方程式最 小化。由於變焦拉近之映射中 r A 才甲即較少.（資訊較少）.，低變 焦汉疋（較大焦距）時焦距判定 、 s、si A , 到疋心可靠度較差.，且照相機設 疋或f測中少許之變異對隹 A 對焦距判疋具有實質之影響。相反Fifth, the invention description (4 rate), the zoom setting "6" has a larger focal length. For camera systems, there are a large number of zoom settings, so it is helpful to fine-tune the focal length. It decides and stores every possible In the viewing range of zoom settings, an algorithm is often used to mathematically determine the focal length of each possible zoom setting. To determine the computational mapping between the camera's zoom setting and the corresponding focal length, it collects a set of z (i) and f ( z (i)). For each zoom setting z⑴, its focal length 値 f «cut corresponding to the above example of the correction technique is determined. Since the focal length is inversely proportional to the zoom setting (larger zoom settings correspond to shorter Distance), the traditional mode equation used to determine the corresponding focal length of the zoom setting is: f (Z) = a0 1 + alz + alz2 + α3ζ3 (3) The number of terms in the square private formula (3) of the knife mother It is usually unknown, and experience has shown that the power of 2 is the best for most camera zoom structures. It is most suitable for camera measurement of the coefficients aG, z (i), f (z (i)), and uU is minimized by the object function. (4) .y two: ·% ·. »R * *«-5 .1 Γ · · ♦ · ·-· · I T b iII · I— 丨 1 丨 丨 丨 丨 ^ | ^ —— (Please read the precautions on the back before filling in this page)-· * ··· * i · *-· *- : i:. ·.: ** ··:. · *. 1: Γί.: * ·. -— ^ ------- l Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs However: This focal length There are still problems in the direct determination of measurement. For example, the equation function is non-linear and the iteration method must minimize the equation. Because r A is less in the map that is zoomed in. (Less information). At low zooming distances (larger focal lengths), the focal length determination, s, si A, and the reliability of the center are poor. And a small variation in the camera setting or f measurement has a substantial impact on the determination of the focal length of A. in contrast
經濟部智慧財產局員工消費合作社印製 五、發明說明（5 ) 的’變焦拉遠之映射具有較多之資訊，且照相機設定及量 測之變異對焦距判定具有較小之影響，即低變焦設定z(i) 時較大之f(Z(i))値準確度低於高變焦設定z⑴時較小之 f(z(i))値。如方程式所示，f(z(i》之較大及較不準確量 測對物件函數之影響大於f(z(i))之較小及較準確量測，因 此係數aO，a 1及a2之判定受到低變焦設定時不準確量測 之影響。 發明概要 本發明一目的爲提供水平搖攝/垂直搖攝/變焦照相機之 权正方及系統，其有利於照相機之準確校正。本發明另一 目的爲提供水平搖攝/垂直搖攝/變焦照相機之自動校正方 法及系統。本發明另一目的爲提供水平搖攝/垂直搖攝/變 焦照相機之校正方法及系統，其係計算後簡化的。 藉由提供水平搖攝/垂直搖攝/變焦照相機之校正方法及 系統而達成這些及其他目的，其無需物件之3 _ D座標以決 定照相機設定至照相機光學特徵之映射。當照相機在最小 變焦設定時（儘可能的變焦拉近）捕獲第第一影像，其畫面 包含可鑑別之細節。在取得該影像之後照相機保持在相同 之方位，其設定在最大之變焦時（儘可能的變焦拉遠）並取 得第二影像。利用照相機最小與最大焦距之比率將第一影 像之尺寸縮小以形成一樣本影像，此樣本影像在第二影像 中之位置由找尋第二影像中心附近庳域而加以決定，其對 應至樣本影像。由於照相機在第一影像及第二影像之間的 方位是固定的，該位置界定了照相機影像面之光學中心或 本紙張尺度適用中國國家標準（CNS)A4規格（21〇 x 297公釐） ---Ί I — --I--· I---丨丨丨訂·--丨丨丨丨丨線 (請先閱讀背面之注音？事項再填寫本頁) A7 A7 B7 五、發明說明（6 主點。 在決定照相機影像面光學中心之後 距可加以設定，其#葬士丁 傻…相機各種變焦焦 其係猎由不同焦距設定時已知之角度而景 心處對應至照相機輕微調整之位移。傳統之曲= 二：技：：用以決定變焦設定與對應觀視範圍之間的演算 之計㈣較佳具體實施例中’曲線密合係基於透鏡倍率 °透鏡倍率之量測有利於映射方程 式係數 < 線形測定，並提供_ 杈仏物件函數，其對低變焦設定 時不準確焦距之測定具較差之靈敏度。 圖説概述 本發明以範例進一步説明並參考至附圖，並中： 圖1顯示一利用根據本發明之樣本影像決定照相機光學 中心之範例。 圖2Α及2Β顯示-決定照相機透鏡系統焦距之技術，其 係基於根據本發明影像中物件之位移。 圖3顯示照相機焦距與照相機轉動夕時物件影像移動之間 的對應關係。 圖4Α及4Β顯示根據本發明之照相機校正程序流程圖。 圖5顯示根據本發明之照相機系统範例。 在這些圖説中相同之參數表示類似或對應之特徵或功 發明詳述 圖1顯示利用根據本發明—樣本影像13〇紋照相機系 (圖5)光學中心’y0)101之範例。影像ιι〇及12〇係 -9 本紙張尺度適用中國國家標準（CNS)A4規格(210x 297iJ" ^ —-----. (請先閱讀背面之注音？事項再填寫本頁} 經濟部智慧財產局員工消費合作社印製 統 對 ^05820 經濟部智慧財產局員工消費合作社印製 A7 五、發明說明（7 ) 應照相機系統在兩不同之變焦設定所捕獲之影像，惟其照 相機万位相同（固定之水平及垂直搖攝設定）。影像12〇對 應至照相機系統之低變焦設定（長焦距），且包含一或多個 清楚鑑別之物件，其具有高程度之影像資訊，例如明^之 邊緣及頂點。影像110對應至照相機系統之較高變焦&定 (較短之焦距），其與影像120比較而言爲，，變焦拉遠之影 像。影像110係在最小焦距時捕獲，影像12〇係在最大焦 距時取得，故吾人已知兩影像j j 〇及丨2〇之放大倍率，即 變焦透鏡以最大及最小變焦之焦距比率加以表示。若照相 機系統具有，，16x”變焦性能，影像11〇及12()分別設定爲最 小及最大焦距，影像120之尺寸爲影像11〇之16倍。技藝 中其他技術亦可用以決定影像12〇對影像11〇之尺寸比 率。 藉由比例係數將放大之影像120依比例縮小產生樣本影 像13 0其比例與放大效果較小之影像11 〇相同。之後樣 本影像130與影像丨10比較以決定影像丨1〇中區域14〇之位 其與樣本影像匹配。吾人瞭解宜自影像i 1〇中心開始 尋找匹配之區域140 ’由此影像中心向外進展直到發現匹 配區域爲止，此時樣本影像13〇在影像11〇中其他位置發 現類似區域之可能性最低。由於影像12〇經過選取且具有 南影像'資訊内容，利用技藝中常見之圖樣匹配技術，其可 在影像11〇中準確的判定樣本影像13〇之位置。由於照相 機在影像110及120之間的方位相同，樣本影像13〇在影像 110中之位置係對應至照相機系統之光學中心。因此，藉 一 10- 公釐） .1 / ------—^------- (請先閱讀背面之注意事項再填寫本頁) 線脅. 505820 第090108053號專利申請案 A7 中文說明書修正頁(91年6月） B7 五、發明説明（8 ) 由變焦拉近至含有可鑑別物件影像之照相機觀視範圍内任 一區域，其可準確的判斷照相機系統之光學中心。如後文 所述，之後光學中心之判定可加以自動化。如技藝中常見 之技術，照相機觀視範圍中垂直搖攝以鑑別物件亦可加以 自動化，故其提供完全自動之照相機校正方法。 一種發現樣本至圖樣匹配方法亦可用以決定本發明較佳 具體實施例中照相機系統之焦距。圖2 A至2 B及圖3顯示 一種方法，其根據本發明原則決定照相機之焦距。圖2A 顯示一照相機影像200，其中包含多種物件影像211，212 等。圖2 B顯示第二個照相機影像200 f，當照相機略微向 下傾斜時產生該影像。圖2B中物件影像211 ’及212 1係對 應至圖2 A中物件影像211及212，且相對於圖2 A中位置略 微的移動。圖3顯示一透鏡250，一影像面260以及一焦點 270，焦點270位於距透鏡250 —個焦距F 255位置處，透鏡 250之光學軸線以直線251表示。如圖3所示，透鏡250略 微向下傾斜為透鏡250 ·，其光學軸線對應著改變以直線 251，表示，且對應著改變焦點為270 ’。假設影像面260相 對於透鏡250固定，故其亦隨著轉動，影像面260新的位置 以260’表示。距離D 265表示焦點之垂直位移，其對應著 影像面260 1上影像之位移d 220。影像面260 1上之影像 2001係對應至較原先觀視略低之觀視點，與原先影像200 比較而言，影像面光學中心處表物件已經偏移了距離d 200。距離D及焦距F 255皆以實際距離之單位表示，為配 合映射程序，這些實際距離須轉換為影像單位。如圖所 -11 - 本紙張尺度適用中國國家標準(CNS) A4規格(210 X 297公釐) 505820 可判定影像之位移 明之目的，影像之： A7 B7 五、發明說明（9 ) · 示.，距離d 220表示距離D轉換爲咚系統對稱爲基礎-之影 像單位，影像焦距f 255表示實際焦距F 255轉換爲影像單 元。如圖所示，角度A之正切爲距離D對焦距F之比。 若透諱傾斜之角度爲A，影像焦距f 255可爲：、、 · (5) 其中f爲影像單位f焦距，d爲影像單位中影像面上影像 之位移，以及A爲透鏡傾斜產生位移d之角度。參見圖2 A 及2B，假設影像®|係對應至所顯杀之影像200及20(Τ，藉 由量測物件影像211，. 212等在圖2 Α及2Β之間的移動，其 圖中所示距離dl 221及d2 223僅爲説 位移可爲個別物件移動距離d 1，d 2等 之組合，例如平均丨，_最小平方誤差估算等。 ' I . - 上述焦.距之決遙需由一影像至另一影像對應點（211- * Ϊ . 2111，212-212,等）之自動鑑別及定位測定，已測定距離 dl，d2等之變異對.上述焦距之決定具反面效果。若照相機 影像200含有物件影像或附近區域中類似物件影像，且不 含尖銳之邊緣，該變異係明顯的。 在一較佳具體實施例中.，影像200之樣本區域230係用以 . * - — ^ 替化整個影像，以便減少計算之複雜性，並使連續影像對 應點之間已測定距離之變異降至最小·。爲了在一已知變焦 設定時找到焦距，·其捕獲第一影像200，之後照相機略微 傾斜一已.知角度Α並取得一新的影像200’。自第一影像 12- 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） -J-----11-----ill----^-------- (請先閱讀背面之注意事項.#填寫本頁) 經濟部智慧財產局員工消費合作社印製 ·. * 505820 A7 ___B7 _ 五、發明說明（1G ) (請先閱讀背面之注意事項再填寫本頁) 200-處抽取出一較小之影像區域形成樣本影像230，其中 β約對準照相機影像200之光學中心。在此較佳具體實施 例中，由角度Α傾喬_所產.生之位移量係完全取決於新影像 200·中樣本影像230之位置。新影像200*中樣本影像230位 \ - * 置與原始第一影像2〇〇中樣本影像230位置之間的差異係 對應至上述圖3之位移d 220及方程式（5 )。 在光學中心101處選定一小區域做爲樣本影像230係有利 於快速準確的決定距離d 220，在新影像20(Τ中自光學中 心101沿著位移方向尋找將很容易的發現樣本影]象230，若 所選定·之樣本影像230小於整個影像200，2001，：其存在著 些許模糊影像。吾人亦瞭解傾斜判'定之焦距f係對應至垂 直方向中之焦距。若須測定水平及垂直焦距之可能差異 fx及fy，樣本匹配方法宜藉由照相機系統輕微之水平移 動取代垂直移動重覆執行以‘決定f X。 經濟部智慧財產局員工消費合作社印製 傳統之照相機校正方法在照相機系統變焦設定與該變異 .設定所產生之對應焦距之間提供一演算映射。爲決定焦距 演算映射至根·據本發明照相機系統之變焦設定，上述傾斜 位移樣本匹配技術適用於各種變焦設定以使蒐集多組資 料，其配合後續曲線密合以決定一多項式之係數，該多項 式爲變焦設定與焦距之随關係之模式。 根據本發明之”透鏡倍率”術語P係用以決定變焦設定與 焦距之間的演算映射，其中P爲f爲倒數： ⑹ P(z(〇) /⑽） -13- 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 505820 五、發明說明（11 ) 由於透鏡倍率爲焦距之倒數以及焦距與變焦設定成反比， 透鏡倍率與變焦設定呈一定岑例關係，其模式爲： ρ(ζ(ί)) = 60 4- blz(i) + blziif^r... ⑺ 吾人發現對典型之照相機系統而言，此模式之次數2 夠。以透鏡倍率取代此模式中焦距，物件函數^表示爲： C ⑷=艺(p(z(〇) - (δ0 + 61ζ(ζ·) + &2ζ(ί)2)2 (δ) 若考量條件sp(z⑴偷）-P(z(i)max )=〇，其表示最I大變 爲已知的。如本文所述，.物件函數（8)爲—\ k率 Μ ~綠性函餐 完整之解答爲： .〜 ：:.I ’其 b = ^0α22 ~~ ^1^12 aua22-ana2l b 二 gpn-？οα2ΐ α\\α22 ^~α\2α2\ b2 = cli0+ c2bl , (9) (tf'先閱讀背面之注意事項再填寫本頁) 訂： (10) (11) 其中： •费 經濟部智慧財產局員工消費合作社印製 cl： •s 一 KOrnax 以及 ，一泣(〇_ - Z(〇max C ^ ^(〇L ** ^(OLn ： aH=X(1 + C«Z(/)2)2 -14 - (12) (13) (14) 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 經濟部智慧財產局員工消費合作社印製 505820 A7 '"""' ---— --— B7 五、·發明說明（12 ) °\2 = =Zz(〇(1 + ^2：(〇2)(Uc2z(〇) /«1 - (15) \ .an : • = Z^(〇2(l + c?z(〇)2 ，:1 m 、。: 抒 \ = 11^(0)(1 + 0^(/)2)2 _________ (17)； 9丨= η Σζ(〇2ρ(ΦΧ\^α2ζ(ί)) 1 (18) 。其中z(i)min爲最小之照相機變焦設定，Z(i)max爲最大之變 ;焦設疋’以及S爲照相機系統最小與最大變焦能力比率之 :倍率係數。這些方程式假設照相機最小與最大變焦能力 ;("倍率係數”）爲已知的且爲一般之數値。若倍率係數係未 丨知的，其可利用傳蜣之曲線密合技術加以決定。 i·如前文所述，在較大之變焦設定時位移量測以及倍率量 :測之準確度較高。在低變焦設定時（高倍率），微小之傾斜 :角度A酱形成一較大之位移.d 22〇。‘角度a之調整或位移d • ·22〇之量測中誤差對焦距f之判定具相當之影響，但對透鏡 •丨倍率.P之影響輕微。因此，由於變焦設定與倍率量測係成 比例關係，倍率量測之高準確度對係數bo，M&b2之判 定具較大之影響，倍率量測之低準確度則影響性較低。 圖4 A及4 B之流程圖範例説明根據本發明之照相機校正 方法，圖4 A顯.示照相機光學中心之決定方式，圖4]B説明 以知、相機變焦設定函數判定透鏡倍率。 圖4A中410說明照相機已設定方位，故其含有在最大變 -15- 本紙張尺度適用中國國家標準（CNS)A4規格（21〇 x 297公釐） --- ^ —*1—I ------^I—I— (請先閱讀背面之注意·事項再填，寫本頁) 505820 A7 B7 五、發明說明（ 13 經濟部智慧財產局員工消費合作社印製 焦設足時較多之影像資訊，例如物件影像具有清楚界定之 邊緣’該影像做爲-參考影像。雖然可運用自動之水 垂直搖攝程序’此方法—般係手動完成，故照相機將持络 尋找直到其定位並捕獲-足夠判定之影像爲止。42〇心 參考影像以照相機之最大變焦比率縮小尺寸形成—樣本與 像，舉例而言，若照相機爲"16X"變焦照相機，參考影; 尺寸縮小1 6倍形成樣本影像。 430説明在最小變焦設定捕獲一影像時，用以取得參考 影像之方位係固定的’由於方位保持固定，所取得影.像之 光學中心與參考影像之光學中心相同。例説明在最小 焦設足時影像圖樣匹配技術用以^所捕獲影像中樣本 像之位置，此位置界定照相機之光學中心。 圖4B之回路450至464説明透鏡倍率之測定或校正範 例：其係對應至複數個變焦設定。配合透鏡倍率校正所 取之變焦設定總數係取決於所需精確度，一般而古，5至 1〇次之歧包含了校正之變焦設定範圍，以便提;一組 至10個點配合透鏡倍率變焦設定之多項式曲線密合，其 説明於47G中。若多項式之次數爲N，其須提供至>、N + 個校正點。纟-較佳具體實施例中，多式之次數爲2， 故須提供至少3個校正點.。 4 5 2説明照相機之方位指向一充滿資訊之中心影像， 影像做爲當時變焦設定之參考影像。爲考量便利性，其 利用某些方位以取得參考影像，參考影像係用以判定照 機t光學中心，如4丨〇所述。4 5 4説明參考影像中心部 -16- 本紙張尺皮迥用T國國豕知準（CNS)A4規格（210 X 297公釐） 變 影 選 5 該 可 相 位 · · ---------.--訂-------->^AW— (請先閱讀背面之注意事項再填寫本頁) 505820 經濟部智慧財產局員工消費合作社印製 A7 ------------ 五、發明說明（14 ) 之選定可形成一樣本影像。如前文所述，由於照相機之光 學中心係已卸的，焦距之判定完全取決於光學中心之位 一移，反之亦然，且依尺寸比例縮放之樣本影像儘可能小至 單一之圖像點。在一較佳具體實施例中，樣本影像之尺寸 .與變焦設定成反比，爲考量圖樣匹配，其尺寸至少爲5χ 5 圖像點。 4 5 6説明照相機在垂直方向中偏移一已知角度a，使多 維位移所產-生之變數最小，以及捕獲位移後之影像，46〇 説明移動後影像·中樣本影像之位置。影像光學中心與樣本 影像移動後位置之間的距離界定影像位移D，其對應至已 知已知之角位移A。若照相機之位移受限於垂直角度且影 像之光學中心界定爲原點（0，0)，位移〇爲樣本影像位置/ 之垂直座大小。4 6 2説明以前述方程式（5 )之倒數決定對 應至特·定變焦設定之透鏡倍率p : 上· ' (14) -回路450至464説明對應.至每個選定校正變焦設定之透 鏡倍率以類似之方式計算，47〇説明曲線密合方程式（9) 至（13)係用以決定二次多項式之最適合係數，其將變焦 •設疋映-射至透鏡倍率。因此，若變焦設定已知，利用方程 式（7)決定對應之透鏡倍率，且係數“，^與。以及焦 距以此透鏡倍率之倒數加以決定。.. · · 圖5揭示根據本發明構造之照相機系統5〇〇範例，其包 -17- 本、.我張尺度翻中國國家標準（CNS)A4規格咖X挪公爱 ----；---,---丨丨----丨丨訂----!線等· (請先閱讀背面之注意事項再填寫本頁) 州5820 A7 五、發明說明（15 ) 口照相機510，影像處理系統55〇，控制器56〇以及一組 校正參數5 7 0。照相機之方位藉由垂直搖攝馬達5 2 〇及水 平搖攝馬達5 3 0加以調整，照相機之焦距利用變焦馬達及 透鏡5 4 0加以調整。在校正階段控制器5 6 〇控制著馬達 5 2 0，5 3 0及5 40以執行圖4A至4B所揭示之功能及程 序，以便產生校正參數5 70。影像處理系統55〇執行多種 功旎，其包括：取得影像；依比例縮小參考影像形成樣本 影像，其用以決定光學中心；在一對影像中測定每個影像 之樣本影像位置，其係配合每個變焦設定；對應至每個變 焦設定之透鏡倍率及/或焦距判定；以及變焦設定演算映 射至透鏡倍率及/或焦距。在校正參數決定之後，控制器 5 6 0接收所需觀视範圍5 8 〇之需求，並運用校正參數及其 他係數以決定適當之指令與馬達52〇，530及540連通， 例如變焦馬達5 4 〇之適當變焦設定，以產生所需之觀視範 圍。 上述僅説明本發明之原則，雖然此處並未明確的揭示， 惟技藝中技術可用以設計出各種配置以具體實施本發明之 原則，其皆在下列申請專利範圍之精神及範疇中。 . — --------I ^---------^_wl (請先閲讀背面之注意事項再填寫本頁) 經濟部智慧財產局員工消費合作社印製 -18- 本紙張尺度適用中國國家標準（CNS)A4規格（210 X 297公釐） 505820 第090108053號專利申請案 中文說明書修正頁(91年6月） 五、發明説明（15a ) 元件符號簡要說明 10 1 光學中心 260， 260* 影像面 110, 120 影像 265 距離（D):焦 1 30 樣本影像 點之垂直位移 140 區域 270，270* 焦點 200, 200* 照相機影像 500 照相機系統 2 11， 2 12 物件影像 5 10 照相機 2 11，， 2 12, 物件影像 520 垂直搖攝馬達 220 位移（d) 530 水平搖攝馬達 22 1 距離（d 1) 540 變焦馬達及透 223 距離（d2) 鏡 225 焦距（f) 550 影像處理系統 23 0 樣本區域 560 控制器 2 5 0, 2 5 01 透鏡 570 校正參數 25 1， 25 Γ 直線 5 80 觀視範圍 255 焦距（F)Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. 5. Description of the invention (5) The 'Zoom Zoom Mapping' has more information, and the camera setting and measurement of the variable focus distance determination have a smaller effect, that is, low zoom. The larger f (Z (i)) 値 at z (i) is less accurate than the smaller f (z (i)) 値 at high zoom setting z⑴. As shown in the equation, the larger and less accurate measurement of f (z (i) has more influence on the object function than the smaller and more accurate measurement of f (z (i)), so the coefficients aO, a 1 and a2 The judgment is affected by the inaccurate measurement at the low zoom setting. SUMMARY OF THE INVENTION An object of the present invention is to provide a right square and a system for a horizontal pan / vertical pan / zoom camera, which facilitates accurate camera calibration. Another aspect of the present invention The purpose is to provide an automatic correction method and system for a horizontal pan / vertical pan / zoom camera. Another object of the present invention is to provide a method and a system for correction of a horizontal pan / vertical pan / zoom camera, which are simplified after calculation. These and other objectives are achieved by providing a calibration method and system for horizontal pan / vertical pan / zoom cameras, which do not require the 3_D coordinate of the object to determine the mapping of the camera settings to the camera's optical characteristics. When the camera is at the minimum zoom setting The first image is captured (zoomed in as close as possible), and its frame contains discernible details. After the image is acquired, the camera remains in the same orientation and its setting is at the most When zooming (zoom as far as possible) and obtain a second image. Use the ratio of the minimum and maximum focal lengths of the camera to reduce the size of the first image to form a sample image. The position of this sample image in the second image can be found by The area near the center of the second image is determined and corresponds to the sample image. Since the orientation of the camera between the first image and the second image is fixed, this position defines the optical center of the camera image surface or the paper scale is applicable China National Standard (CNS) A4 Specification (21x 297 mm) --- Ί I---I-- · I --- 丨 丨 丨 Order ---- 丨 丨 丨 丨 丨 Line (Please read the back first Note: Please fill in this page for matters) A7 A7 B7 V. Description of the invention (6 main points. After determining the optical center of the camera image surface, the distance can be set. The camera's various zoom focal lengths are different depending on the camera. The focal angle is set at a known angle and the center of the field corresponds to the slight adjustment of the camera. Traditional song = 2: Technique :: The calculation used to determine the calculation between the zoom setting and the corresponding viewing range. In a preferred embodiment, The curve close measurement based on the lens magnification ° lens magnification facilitates the measurement of the mapping equation coefficient < linearity measurement, and provides a _ object function, which has poor sensitivity for the measurement of inaccurate focal lengths at low zoom settings. The invention is further illustrated by examples and referred to the accompanying drawings, and in which: Fig. 1 shows an example of determining the optical center of a camera using a sample image according to the present invention. Figs. 2A and 2B show-a technique for determining the focal length of a camera lens system, which is based on Object displacement in the image of the present invention. Fig. 3 shows the correspondence between the focal length of the camera and the object image movement when the camera is turned. Figs. 4A and 4B show the flowchart of the camera calibration procedure according to the present invention. Fig. 5 shows the camera according to the present invention. System example. In these illustrations, the same parameters indicate similar or corresponding features or functions. Detailed description of the invention Figure 1 shows an example of using the 13-line camera system (Figure 5) optical center 'y0) 101 according to the present invention. Image ιι〇 and 12〇 series-9 This paper size is applicable to China National Standard (CNS) A4 specification (210x 297iJ " ^ —-----. (Please read the note on the back? Matters before filling out this page) Ministry of Economy Wisdom Printed by the Consumers' Cooperative of the Property Bureau ^ 05820 Printed by the Consumers ’Cooperative of the Intellectual Property Bureau of the Ministry of Economic Affairs A7 V. Description of the invention (7) The image captured by the camera system at two different zoom settings, but the cameras are the same (Horizontal and vertical panning settings). The image 120 corresponds to the low zoom setting (long focal length) of the camera system, and contains one or more clearly identified objects, which have a high degree of image information, such as bright edges and Vertex. Image 110 corresponds to the higher zoom & (shorter focal length) of the camera system. Compared with image 120, the zoomed image is farther away. Image 110 is captured at the minimum focal length and image 12 It is obtained at the maximum focal length, so we know the magnifications of the two images jj 〇 and 丨 20, that is, the zoom lens is expressed by the focal length ratio of the maximum and minimum zoom. The system has, 16x "zoom performance, images 11 and 12 () are set to the minimum and maximum focal lengths respectively, and the size of image 120 is 16 times that of image 11. Other techniques in the art can also be used to determine image 12 to image 11 The size ratio of 〇. The scaled-down image 120 is scaled down to produce a sample image 13 0. The scale is the same as that of the image 11 〇. The sample image 130 is compared with the image 丨 10 to determine the image 丨 1 〇 The middle area 14o matches the sample image. We understand that it is advisable to search for a matching area 140 'from the center of the image i 10, and the image center progresses outward until a matching area is found. At this time, the sample image 13 is in image 11 The probability of finding similar areas elsewhere in 〇 is the lowest. Since image 120 has been selected and has the information content of south image, using pattern matching technology commonly used in the art, it can accurately determine the sample image 13 in image 11 Position. Since the camera has the same orientation between images 110 and 120, the position of sample image 13 in image 110 corresponds to the camera system. Optical Center. Therefore, borrow a 10-mm). 1 / ---------- ^ ------- (Please read the precautions on the back before filling this page) Thread threat. 505820 No. 090108053 Patent Application No. A7 Revised Chinese Manual (June 91) B7 V. Description of the Invention (8) From zooming to any area within the viewing range of the camera containing the image of the distinguishable object, it can accurately judge the camera system Optical center. As described later, the determination of the optical center can be automated later. As is common in the art, vertical panning in the camera's viewing range to identify objects can also be automated, so it provides fully automatic camera calibration method. A method for matching samples to patterns can also be used to determine the focal length of the camera system in the preferred embodiment of the present invention. Figures 2A to 2B and Figure 3 show a method for determining the focal length of a camera in accordance with the principles of the present invention. FIG. 2A shows a camera image 200 including various object images 211, 212, and the like. Figure 2B shows a second camera image 200 f, which is produced when the camera is tilted slightly downwards. The object images 211 'and 212 1 in FIG. 2B correspond to the object images 211 and 212 in FIG. 2A, and are slightly moved relative to the positions in FIG. 2A. FIG. 3 shows a lens 250, an image plane 260, and a focal point 270. The focal point 270 is located at a focal length F 255 from the lens 250. The optical axis of the lens 250 is represented by a straight line 251. As shown in FIG. 3, the lens 250 is slightly tilted downward to the lens 250 ·, and its optical axis is represented by a line 251 corresponding to the change, and corresponds to a change of the focal point to 270 '. Assuming that the image plane 260 is fixed relative to the lens 250, it also rotates with it, and the new position of the image plane 260 is represented by 260 '. The distance D 265 represents the vertical displacement of the focal point, which corresponds to the displacement d 220 of the image on the image plane 2601. The image 2001 on the image plane 2601 corresponds to a viewpoint that is slightly lower than the original viewpoint. Compared with the original image 200, the surface object at the optical center of the image plane has been shifted by a distance d 200. The distance D and the focal length F 255 are both expressed in units of actual distances. To fit the mapping procedure, these actual distances must be converted into image units. As shown in Figure 11-This paper size is in accordance with Chinese National Standard (CNS) A4 specification (210 X 297 mm) 505820 The purpose of the image can be determined, the image: A7 B7 V. Description of the invention (9) · Show., The distance d 220 indicates that the distance D is converted into an image unit called a base pair, and the image focal length f 255 indicates that the actual focal length F 255 is converted into an image unit. As shown, the tangent of the angle A is the ratio of the distance D to the focal length F. If the angle of the oblique tilt is A, the image focal length f 255 can be: ,, (5) where f is the focal length of the image unit f, d is the displacement of the image on the image plane in the image unit, and A is the displacement of the lens tilt d Angle. Referring to FIGS. 2A and 2B, it is assumed that the image ® | is corresponding to the displayed images 200 and 20 (T, by measuring the object image 211,. 212, etc., moving between FIGS. 2A and 2B, the figure of which The distances dl 221 and d2 223 shown are merely that the displacement can be a combination of the distances d 1, d 2, etc. of individual objects, such as the average, _ least square error estimation, etc. 'I. From one image to the other image corresponding points (211- * Ϊ. 2111, 211-212, etc.) for automatic identification and positioning measurement, the measured distances dl, d2, etc. have the opposite effect. The determination of the above focal length has the opposite effect. The camera image 200 contains an object image or a similar object image in a nearby area, and does not include sharp edges. This variation is obvious. In a preferred embodiment, the sample area 230 of the image 200 is used. *-- ^ Substitute the entire image in order to reduce the computational complexity and minimize the variation of the measured distance between corresponding points of consecutive images. To find the focal length at a known zoom setting, it captures the first image 200, After that, the camera tilted slightly. A new image 200 '. Since the first image 12- This paper size applies the Chinese National Standard (CNS) A4 specification (210 X 297 mm) -J ----- 11 ----- ill --- -^ -------- (Please read the notes on the back. #Fill this page) Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs. * 505820 A7 ___B7 _ V. Description of Invention (1G) (Please First read the precautions on the back and then fill out this page.) 200- place to extract a smaller image area to form a sample image 230, where β is approximately aligned with the optical center of the camera image 200. In this preferred embodiment, the angle The amount of displacement produced by Α 倾 乔 _ depends entirely on the position of the sample image 230 in the new image 200 ·. The sample image 230 in the new image 200 * \-* is set to the sample image in the original first image 200. The difference between 230 positions corresponds to the displacement d 220 and equation (5) in Figure 3 above. Selecting a small area at the optical center 101 as the sample image 230 is helpful for quickly and accurately determining the distance d 220 in the new image. 20 (Searching from the optical center 101 along the displacement direction in T will easily find the sample image) image 230, if The selected sample image 230 is smaller than the entire image 200, 2001: there are some blurred images. I also understand that the focal length f determined by the tilt judgment corresponds to the focal length in the vertical direction. If you need to determine the possible difference between horizontal and vertical focal lengths fx And fy, the sample matching method should be performed by the camera system ’s slight horizontal movement instead of vertical movement to determine 'f X. The traditional consumer camera calibration method printed by the Employees ’Cooperatives of the Intellectual Property Bureau of the Ministry of Economy prints the camera system zoom setting and the variation. Provide a calculation map between the corresponding focal lengths generated by the settings. In order to determine the focal length calculation map to the zoom setting of the camera system according to the present invention, the above-mentioned tilt displacement sample matching technology is applicable to various zoom settings to enable the collection of multiple sets of data, which are closely matched with subsequent curves to determine the coefficient of a polynomial, the polynomial This is a mode for setting the relationship between zoom and focal length. The term "lens magnification" according to the present invention is used to determine the arithmetic mapping between the zoom setting and the focal length, where P is the inverse of f: ⑹ P (z (〇) / ⑽) -13- This paper scale applies to China Standard (CNS) A4 specification (210 X 297 mm) 505820 V. Description of the invention (11) Since the lens magnification is the inverse of the focal length and the focal length is inversely proportional to the zoom setting, the lens magnification is related to the zoom setting to a certain degree. Its mode is : Ρ (ζ (ί)) = 60 4- blz (i) + blziif ^ r ... 发现 I have found that for a typical camera system, the number of times in this mode is enough. The focal length in this mode is replaced by the lens magnification, and the object function ^ is expressed as: C ⑷ = 艺 (p (z (〇)-(δ0 + 61ζ (ζ ·) + & 2ζ (ί) 2) 2 (δ) If we consider The condition sp (z⑴ steal)-P (z (i) max) = 0, which means that the largest I becomes known. As described in this article, the object function (8) is-\ k 率 Μ ~ green function The complete solution is:. ~ ::. I 'its b = ^ 0α22 ~~ ^ 1 ^ 12 aua22-ana2l b two gpn-? Οα2ΐ α \\ α22 ^ ~ α \ 2α2 \ b2 = cli0 + c2bl, (9 ) (tf 'read the notes on the back before filling this page) Order: (10) (11) Among them: • Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs cl: • s a KOrnax and, a cry (〇_ -Z (〇max C ^ ^ (〇L ** ^ (OLn: aH = X (1 + C «Z (/) 2) 2 -14-(12) (13) (14) This paper size applies to China Standard (CNS) A4 specification (210 X 297 mm) Printed by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs 505820 A7 '" " "' ----- --- B7 V. · Explanation of invention (12) ° \ 2 = = Zz (〇 (1 + ^ 2: (〇2) (Uc2z (〇) / «1-(15) \ .an: • = Z ^ (〇2 (l + c? Z (〇) 2 ,: 1 m,.: Ly \ = 11 ^ (0) (1 + 0 ^ (/) 2) 2 _________ (17); 9 丨 = η Σζ (〇2ρ (Φχ \ ^ α2ζ (ί)) 1 (18). Where z (i) min is the minimum camera zoom setting, and Z ( i) max is the maximum change; focus setting 疋 'and S are the ratio of the minimum and maximum zoom capabilities of the camera system: magnification factor. These equations assume that the minimum and maximum zoom capabilities of the camera; (" magnification factor ") are known and It is a general number. If the magnification factor is unknown, it can be determined by using the curve close-fit technique. I. As mentioned above, displacement measurement and magnification measurement: The accuracy is higher. At the low zoom setting (high magnification), the slight tilt: the angle A will form a large displacement. D 22 °. 'The adjustment or displacement of the angle a d • · 22 ° during the measurement The determination of the error focal length f has a considerable impact, but has a slight effect on the lens • 丨 magnification.P. Therefore, because the zoom setting is proportional to the magnification measurement system, the high accuracy of the magnification measurement has a coefficient bo, M & The judgment of b2 has a large influence, and the low accuracy of the magnification measurement has a low influence. 4A and 4B are flowchart examples illustrating a camera calibration method according to the present invention, and FIG. 4A shows a method for determining the optical center of the camera, and FIG. 4] B illustrates the lens magnification determined by the camera zoom setting function. 410 in FIG. 4A shows that the camera has been set to its orientation, so it contains a maximum change of -15- This paper size applies the Chinese National Standard (CNS) A4 specification (21 × x297 mm) --- ^-* 1-I- ---- ^ I—I— (Please read the notes and notes on the back before filling in this page, and write this page) 505820 A7 B7 V. Invention Description (13 Employees' Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs often print coke when it is fully set. Image information, such as an object image with clearly defined edges 'this image is used as a reference image. Although the automatic water vertical panning process can be used' this method is generally done manually, so the camera will keep looking until it is positioned and Capture-until the image is sufficient for judgment. The 42-heart reference image is formed by reducing the size of the camera's maximum zoom ratio—samples and images. For example, if the camera is a "16X" zoom camera, the reference image is reduced by 16 times. Sample image. 430 indicates that when capturing an image at the minimum zoom setting, the orientation used to obtain the reference image is fixed. 'Since the orientation remains fixed, the optical center of the acquired image. The optical center of the image and the optics of the reference image The example shows that when the minimum focus is set, the image pattern matching technology is used to ^ the position of the sample image in the captured image, which defines the optical center of the camera. The loops 450 to 464 of Figure 4B illustrate examples of measuring or correcting the lens magnification. : It corresponds to a plurality of zoom settings. The total number of zoom settings taken in conjunction with the lens magnification correction depends on the required accuracy. Generally, the difference between 5 and 10 times includes the range of the zoom setting for correction in order to improve; A set of up to 10 points that fit the polynomial curve of the lens magnification zoom setting is explained in 47G. If the degree of the polynomial is N, it must be provided to > and N + correction points. 纟 -preferred embodiment The number of times of the multi-mode is 2, so it is necessary to provide at least 3 correction points. 4 5 2 Note that the orientation of the camera points to a central image full of information, and the image is used as the reference image for the zoom setting at that time. For convenience, its Use some orientations to obtain a reference image. The reference image is used to determine the optical center of the camera, as described in 4 丨 〇. 4 5 4 Explains the center of the reference image. National Standard (CNS) A4 Specification (210 X 297 mm) Variation Selection 5 This phase can be · · ---------.-- Order -------- > ^ AW— (Please read the notes on the back before filling out this page) 505820 Printed A7 by the Consumer Cooperatives of the Intellectual Property Bureau of the Ministry of Economic Affairs ------------ V. The choice of invention description (14) can be formed A sample image. As mentioned above, since the optical center of the camera is removed, the determination of the focal length depends entirely on the position of the optical center, and vice versa, and the sample image scaled according to the size ratio is as small as possible to a single image. Image points. In a preferred embodiment, the size of the sample image is inversely proportional to the zoom setting. For consideration of pattern matching, its size is at least 5 × 5 image points. 4 5 6 shows that the camera is offset by a known angle a in the vertical direction to minimize the number of variables produced by the multi-dimensional displacement and capture the displaced image. 46 0 describes the position of the sample image in the moved image. The distance between the optical center of the image and the position of the sample image after moving defines the image displacement D, which corresponds to the known angular displacement A. If the displacement of the camera is limited by the vertical angle and the optical center of the image is defined as the origin (0, 0), the displacement 0 is the position of the sample image / the vertical seat size. 4 6 2 indicates that the reciprocal of the aforementioned equation (5) is used to determine the lens magnification p corresponding to the special fixed zoom setting p: up · (14)-Circuits 450 to 464 indicate the correspondence. The lens magnification for each selected corrected zoom setting is Calculated in a similar way, 47 ° shows that the curve close-fitting equations (9) to (13) are used to determine the most suitable coefficient of the quadratic polynomial, which sets the zoom-set mapping to the lens magnification. Therefore, if the zoom setting is known, use equation (7) to determine the corresponding lens magnification, and the coefficients “, ^ and... And the focal length to determine the reciprocal of this lens magnification... Figure 5 discloses a camera constructed according to the present invention. System 500 example, its package -17- this, our standard Chinese translation (CNS) A4 specifications coffee X Norongai ----; ---, --- 丨 丨 ---- 丨丨 Order ----! Line etc. (Please read the precautions on the back before filling this page) State 5820 A7 V. Description of the invention (15) Port camera 510, image processing system 55, controller 56 and a group Calibration parameter 5 7 0. The orientation of the camera is adjusted by the vertical pan motor 5 2 0 and the horizontal pan motor 5 3 0, and the focal length of the camera is adjusted by the zoom motor and lens 5 4 0. During the correction phase, the controller 5 6 〇 Controls the motors 5 2 0, 5 3 0, and 5 40 to perform the functions and procedures disclosed in FIGS. 4A to 4B in order to generate a correction parameter 5 70. The image processing system 55 0 performs a variety of functions, including: acquiring images; Reduce the reference image in proportion to form a sample image, which is used to determine the optical center; The position of the sample image of each image in a pair of images is determined with each zoom setting; the lens magnification and / or focal length determination corresponding to each zoom setting; and the zoom setting calculation is mapped to the lens magnification and / or focal length. After the correction parameter is determined, the controller 5 60 receives the demand of the required viewing range 5 8 〇, and uses the correction parameter and other coefficients to determine the appropriate command to communicate with the motors 52, 530, and 540, such as the zoom motor 5 4 〇 Appropriate zoom settings to produce the desired viewing range. The above only illustrates the principles of the present invention. Although not explicitly disclosed here, the techniques in the art can be used to design various configurations to implement the principles of the present invention. All are within the spirit and scope of the following patent application scopes. — — -------- I ^ --------- ^ _ wl (Please read the notes on the back before filling this page) Ministry of Economic Affairs Printed by the Intellectual Property Bureau's Consumer Cooperatives-18- This paper size applies to the Chinese National Standard (CNS) A4 (210 X 297 mm) 505820 No. 090108053 Patent Application Chinese Manual Correction Sheet (June 91) 5 Description of invention (15a) Brief description of component symbols 10 1 Optical center 260, 260 * Image plane 110, 120 Image 265 Distance (D): Focus 1 30 Vertical displacement of sample image point 140 area 270, 270 * Focus 200, 200 * Camera Image 500 Camera system 2 11, 2 12 Object image 5 10 Camera 2 11, 2, 2 12, Object image 520 Vertical pan motor 220 Displacement (d) 530 Horizontal pan motor 22 1 Distance (d 1) 540 Zoom motor and lens 223 Distance (d2) Lens 225 Focal distance (f) 550 Image processing system 23 0 Sample area 560 Controller 2 5 0, 2 5 01 Lens 570 Correction parameter 25 1, 25 Γ Straight line 5 80 Viewing range 255 Focal length (F)
裝 訂Binding
線 -18a-本紙張尺度適用中國國家標準(CNS) A4規格(210X297公釐)Line -18a- This paper size applies to China National Standard (CNS) A4 specification (210X297 mm)