201126189 六、發明說明： t發明戶斤屬之技術領域3 發明領域 本發明係有關估計當攝下相片時，相機正面對的方 向。特別係關於包括或連結衛星定位接收器的相機。 c先前技術;1 發明背景 「地理標籤」指的是紀錄所關切的一個地方或事件的 地點。地理標籤是註記與組織如影像和影片的紀錄資訊的 一種越來越普遍的方式。藉由對媒體加上地點元資料，使 用者可以有趣且直觀的方式來搜尋或瀏覽存檔—例如以 在一張地圖上標出地點組的方式。為了要避免在產生這樣 的註記中所必須的手動鍵入地點名稱或地理座標等耗費勞 力的工作，正逐漸地使用衛星定位來提供精確地點元資 料。諸如全球定位系統（GPS)的衛星定位系統使接收器能 夠依據來自於一組軌道衛星（其被稱為太空船——SV)的 信號之量測抵達時間來計算其位置。 【發明内容】 發明概要 依據本發明之一觀點，提供一種用於估計具有附加天 線之相機之方位角定向之方法，其中該天線具有一個預定 方向放射模式，該方向放射模式在該相機被持於針對攝影 的一個正常定向中時，在方位角平面上為非等向的，該方 法包含下列步驟：量測於該天線從各個多個發送器所接收 201126189 之數個信號之強度， ·獲得各個該等多個發送器之位置〈獲 得該相機之位置；預測所接收的各個該等信號之期望強 度；以及依據所獲得之該相機之位置以及該等多個發送器 之位置、所量測之該等信號強度、該天線放射模式、以及 所預測的該等㈣強度，而估計該相機之财位角定向。 此方法使得相機之指向方向可以利用一個單一的指向 天線來估計。此天線應固定在相機上，贿其在相^正 常地持著拍Μ，具有在水平（方位幻方向巾的方向特 性。這種正常定向通常會包括縱向定向或橫向定向中的一 者或二者。此方法利用來自於已知地點的發送器或信標之 信號強度之量測，來推斷此指向天線之定向（亦稱為指 向）。其避免了提供電子羅盤或其他硬體之需要；因此，對 於已經備有-個定位系統的相機來說，此方法允許在些許 額外花費上的方位角方向估計。所使用之天線可為簡單指 向型，且並不需要多個元件或是一個相位陣列。因為方向 估計係基於信號強度，所以並不需要一般所知地例如量測 相位或明確判定到達角度。例如，並不需要如同一些其他 無線電測向方法常見地旋轉天線，以判定到達角度或是最 強接收方向。藉由利用相機位置來判定方位角定向，此方 法使用已由傳統地理標籤解決方法所產生的資訊。因此最 小化額外的運算複雜度。相機之方位角方向或羅盤方向可 被用來區分在相同有利地點，但面向不同方向所拍攝的照 片。因此，其有助於判別不同相片之題材。這代表簡化量 測於拍攝此照片時的相機地點的一個顯著進步。 201126189 發送器可為衛星，在這種情況中，所接收的信號為衛 星信號。 衛星定位普遍用於傳統地理標籤應用中。其通常是可 靠準確的，並且提供全球性的覆蓋範圍。衛星之位置為用 於此位置估計處理之必要輸入。因此，於衛星定位接收器 中實施本方案所需的唯一額外量測，為所接收之對應於各 個衛星的信號強度之量測。 獲得各個該等多個衛星之位置之步驟可能包含下列至 少一個步驟：對所接收的一個衛星信號的一個資料訊息解 碼；以及透過一個通訊網路而獲得衛星位置資訊或衛星軌 道資訊。 在即時衛星導航應用中，可藉由解碼在衛星發送中所 包含的星曆及/或年曆資訊，來獲得衛星位置。在這個方案 中，本方法所使用的衛星位置為已在傳統衛星定位接收器 中所實施之處理的副產品。在其他方案中，較佳的是從一 個獨立來源，如一個接收器，來獲得衛星星曆及/或年曆資 料。在這種情況中，所獲得的位置可能會較為可靠，因為 集中資料庫可維持準確且完整的衛星健康紀錄，並校正任 何錯誤或異常。 獲得該相機之位置之步驟可包含下列步驟：依據從各 個該等多個發送器接收而來之數個信號，而估計該相機之 位置。 這是很有效率的，因為用來獲得相機位置的相同信號 組亦被用來估計方向。這表示在這些硬體部件中，有許多 201126189 是可以在作業之間共享的。發送器可為衛星，如上面所討 論的，或可為陸上的信標，如手機基地台或無線區域網路 (WLAN)存取點（AP)。在其中之任一情況中，用來從發 送器接收信號的接收器硬體皆可適於透過指向天線來量測 信號強度。 選擇性地，用於估計該相機之位置之該等信號係於一 個更添天線接收的，該更添天線具有一個預定放射模式， 該放射模式在該相機被持於針對攝影的一個正常定向中 時，在該方位角平面上為等向的。 具有在水平方位角平面上變化的放射/增益模式的指 向天線對於方向估計處理而言是必須的。然而，為了成功 估計相機位置之信號接收動作通常會需要跨越所有發送器 地最大化所有的接收信號力量。易言之，對於位置估計來 說，通常較佳為不要在方位角平面上具有方向選擇性。提 供一個更添的天線因而使方向估計免於降低位置估計之穩 健性。指向天線係用於方向估計；更添天線係用於位置估 算，其可為等向的（全向性的）或弱指向性的但指向天空。 請注意，從發送器所接收並用於各個處理的信號較佳為相 同信號，以允許再使用相同的接收器硬體。確實，更添天 線是可能包含與用於方向估計之天線相同，但具有更改過 的組態的天線。 當供有一個更添之等向天線時，此方法可更包含量測 於該更添天線所接收之該等信號之強度之步驟，以及依據 針對該更添天線之所量測的數個信號強度而估計該相機之 6 201126189 該方位角定向。 、在更添之天線處之信號強度 :處。雖然其係全方位地接=方向估計上有所 其可，例如在與針對此& =對t位角作差別對 可提供比2= _準給方向估計處理。這 準確度。^所預測的信_度所可能提供的更大的 X相機可具有多個附加天線， 一個不同的財方向放射模式該等多個天線具有 機被持於針對攝影的—個正方向放射模式在該相 上為非等向的，其中該方法包含〗彳纟4方位角平面 線中選擇-個或多個天線；以^步^:從該等多個天 而量測於所選擇的各個該等、各個5亥等多個發送器 度，並且ο料助狀料mm㈣號之強 對所選擇的各個該等天線所量 線放射模式、以及所預測的該等信號強度。 A天 雖然此方法僅適用於一個天線，^可 來增進穩健度及/«確度。方向料演算法，立應Z 各個天線，或者信號強度之量測可跨越數個天線與庫用於 匯總資料之演算法而結合。在其中之任_情財，^需要 知道此等多個天線的相對相互定向。此資訊是由方向 模式來提供的。 該方法可更包含下列步驟：感測該相機及/或該天線的 一個仰角或一個轉動角；以及依據所感測到的角度而估竹 201126189 該天線之該方位角定向。 感測仰角，或傾斜角，與轉動（如繞著相機之光軸轉 動）角度之步驟可用來改善此方法。此資訊可與天線之放 射/增益模式結合，以更準確地調和所預測的和所量測的信 號強度。 此方法亦可包含依據該天線的一個預定偏振而估計該 天線之該方位角定向。 指向天線之偏振知識可有助於在將實際接收的信號強 度與期望信號強度作比較時，對反射信號打折。因此，此 即亦可有助於增進此方法之穩健度。假設信號之原始（發 送）偏振亦為已知。 估計該相機之該方位角定向之步驟可係在接收該等信 號後的一個有意的非零延遲之後執行的。 離線式地，即在實際接收信號之後的某個時候，執行 方向估計，是有可能的。在這種情況下，可簡單地為稍後 的分析紀錄信號或信號強度量測。這可，例如，最小化在 捕捉相片時所需的處理數量。其亦特別適於地理標籤應 用，此時，位置及/或方向資訊在典型上並非立即需要的。 藉由延遲處理可最小化一個可攜式裝置的複雜度，因為資 料可由一個更有利的處理器在不同的時間與地點來分析； 亦可節省在此可攜式裝置中的電池電力。 依據本發明之更一觀點，提供有一種適於估計相機之 方位角定向之裝置，該裝置包含：可附加於該相機的一個 天線，該天線並具有一個預定方向放射模式，該方向放射 201126189 模式在該相機被持於針對攝影的一個正常定向中時，在該 方位角平面上為非等向的；一個接收器，該接收器電氣式 地連接至該天線，並適於從多個發送器中之各個發送器接 收數個信號；以及處理構件，該處理構件適於：量測所接 收之該等信號之強度；獲得各個該等多個發送器之位置； 獲得該相機之位置；預測所接收的各個該等信號之期望強 度；以及依據所獲得的該相機之位置與該等多個發送器之 位置、該天線放射模式、與所預測之該等信號強度，而估 計該相機之該方位角定向。 圖式簡單說明 本發明現將經由參考隨附圖式的範例來說明，其中： 第1圖為依據一實施例的一個方向估計方法之流程圖； 第2圖為依據一實施例的一個裝置之方塊圖； 第3圖繪示依據利用一個指向天線之一實施例的方向 估計方法與裝置之原理； 第4圖繪示依據利用兩個指向天線之一實施例的方向 估計方法與裝置之原理； 第5圖為針對兩個指向天線之更進一步的範例組態之 圖示； 第6圖繪示依據利用一個指向天線與一個更添的等向 天線之一實施例的方向估計方法與裝置之原理。 C實施方式3 較佳實施例之詳細說明 本發明已識出，在地理標籤攝影中，照片題材之地點 201126189 比相機本身的地點更重要。例如，一個攝影家可能會在站 在相同地點時拍攝兩張照片，但這些照片的内容可能會非 常不同，因為相機在這兩個時候是面向不同的題材。傳統 的地理標籤將會指定給這些相片相同的地點。相反地，一 個攝影家可能會拍攝所感興趣的單一題材之許多照片，例 如，繞著一個雕像行走來獲得不同視角。傳統的地理標籤 將會指定不同的地點給各個影像。 為了要區分這些不同情況，必須要不只紀錄捕捉影像 的地點，也紀錄當時相機正面對的方向。這將會豐富這個 攝影師的地理標藏經驗，並允許更進步的針對影像收集的 搜尋與瀏覽功能。相機在某個給定位置的定向或面對方向 具有兩個自由度：在水平面上的方位角；以及在垂直面上 的仰角。方位角對於攝影應用來說是更重要的，因為這就 是「羅盤方向」，或在地球表面上的平面定向。對方位角的 知識因此對於判定相機正從某個給定已知地點面對什麼來 說是必要的。在任何情況中，若有需要，仰角皆可輕易地 藉由加速儀感測器來判定。 本發明允許基於相機所接收的信號之量測來判定定向 之方位角。用於方向估計之信號較佳為與用來判定相機位 置之信號相同的信號；因此，一個内建在相機裡或附加於 相機上的位置接收器便可除了提供地點資訊以外亦提供位 置資訊。此外，並不需要提供額外的硬體部件來使一個分 立的電子羅盤包含磁域感測器，或使一個慣性導航系統 (INS)包含迴轉感應器。這允許在提供增進的地理標藏特 10 201126189 徵時，使相機或相機配件之成本降到最小。例如，可在一 張地圖上展示拍攝各個相片時的方向，或者是可自動地藉 由參考一個地理資訊系統（GIS)資料庫來辨別相片題材。 可藉由亦紀錄關於從相機到題材間之距離的資訊，來使相 片題材之辨別更為準確。此資料可直接獲得，例如由在相 機中的測距儀（用來自動定焦）來獲得，或間接獲得，例 如由鏡頭焦距資訊來獲得。 第1圖為依據本發明之一實施例的一個方法之流程 圖。在此範例中，GPS衛星定位信號兼用於獲得位址與估 計定向之方位角。此方法適於判定備有指向GPS天線之相 機的定向。 在步驟10中，量測從數個衛星接收而來的信號強度。 G P S星群包含在不同軌道中的持續發送資料信號的至少2 4 個衛星。若GPS接收器可成功地自四個衛星計算出其範 圍，則其可產生在地球上任何地方的定位。定位之準確度 隨著可見衛星之數量增加而增進。量測各個衛星信號之強 度比完成一個範圍之計算要來得容易；因此，針對比用於 位置估計所用之更多衛星而量測信號強度係有可能的—— 例如在解碼衛星資料訊息中之錯誤可免於範圍計算，但仍 可獲得信號強度估計。可藉由任何適當的構件來量測信號 強度，例如利用功率信號雜訊比（PSNR)。絕對信號強度 值在方向估計處理上並非必要的；重要的是跨越衛星組的 相對強度。 在步驟20中，獲得衛星位置。這是以傳統GPS位置估 11 201126189 計程序中的-部分來完成的，並且如此-來，此資訊將可 在得知相機位置之前獲得。不同的GPS接收器可以不同方 式來獲得衛纽置。-種常見作法是在衛星廣播本身中解 碼資料訊息，以獲得描述衛星軌跡的星曆及年曆資訊。或 者，在「輔助」GPS (AGPS)中，此資訊係經由諸如手機 網路等分開的通道來提供的。在稱為「擷取並稍後處理」 (此後以「擷取並處理」稱之）的另一個Gps實作中，並 不-定要提供星潛與年曆資料給接收器。然而，衛星發送 之短序列係由接收器來接收並儲存的；稍後，它們便可被 上傳到另-個裝置來處理，以得出位置估計。執行摘後處 理之裝置可例如透過網路而從歷史絲資料的—個中央資 料庫來獲得衛星位置資訊。本方向估計方法並非局限於步 驟20中之任何獲得衛星位置的特定構件，並因此可應用於 任何GPS實作。 當要從衛星資料中獲得衛星位置資料時，有多種允許 不同程度的準確度與穩健度的選擇。在最佳情況中若可 取得星磨紐細節的轨道軌„訊，那麼各個衛星資料信 號皆可被完全地解碼 '然而’某些衛星信號對於完成解碼 來說可能會太虛弱（雜吵）（雖然信號強度可能仍用於方向 估&十處理中，如上所述 > 即使星曆不能從—個給定衛星資 料訊息中解碼’其仍有可能解碼切。這提供近似的轨跡 資料’並:能在實務上料來支援方向估計來說係足夠準 確的。確貫，關於所有魅的年料訊是由所有的衛星廣 播的’以使強健地接收給定單_星信號錢過其他衛星 12 201126189 的年曆訊息來告知其所在不為必須的。在步驟 触器之位置（即相機位置）。再次地，這是在Gps處= 正^列中完成的，並且如此—來，便不f任何額外處理 =藤相細位置估計方法亦廣為人知，且可廣泛獲得。 再:人聲明’本發明並不局限於GPS實作之任何特㈣型。 雖然步驟1〇、20與30不需以任何特 在即時GPS實作中，弈勃V-产订但 貫作中先執仃仏號強度量測步驟H)，接著是 彳：=Γ…交為有利。這是因為信號強度量測 攸相機接合可以比從定位魏結合更_。數位相機血 里上會在影像肺與儲相記憶體的處理中產 (ΕΜ)介面’其可降低所接收的衛星信號之灌。同樣地， 在-個娜並處理實作t，（t相機「吵雜」時）在操取斑 儲存稱後之位置估計所需要的錢樣本前，練行信號強 度量測可為有利的。 在步驟40中，此方法預測當給定各個衛星之地點時， 在攝下相時在接收器所在之地點可期望的觀察到的作 號強度。㈣發送的信號功率已知，且/或假設對所有_ 星來"兒都π樣的。信號之衰減將會取決於從各個衛星到 接收益的距離’且大氣效應亦將會使得衰減針對在天空中 之不同仰角的術星而改變，如接收器所觀察到的。用於預 測所期待的錢功率之模組可料密或更祕密。模組越 精確實際方向估計之準確率就越高。例如，—個平凡的 模型可假設應從财魅接收約略相等的功率。更複雜的 核型可能會將各個衛星在天空中的仰角納入考慮 。可將實 13 201126189 際信號量測納入考慮，或自己使用，以作為單純理論上之 預測的一個選擇。在任何情況中，目標皆為提供來自於指 向天線的實際信號強度量測可與之比較的一組參考信號位 準。 在步驟50中，估計接收器（相機）天線所指向的方向。 期望要觀察到的信號強度以及實際觀察到的信號強度係藉 由指向天線之接收模式或增益模式（亦稱為放射模式），以 及當信號強度作量測時所指向的方向而相關。可以多樣的 方式來獲得與表達放射模式。其包括先前明確的在不同波 束角之增益的定量量測；由一個除錯理論或數值模型（諸 如理想貼片天線）所做的逼近；或由參考所使用的特定天 線之已知設計特性。在知道放射模型之特性的情況下，剩 餘未知參數便為與相機定向相關者。定向之估計因此係由 參考放射模組而比較所預測與所量測的信號強度，以及求 解定向參數來達成。 第2圖示出適於實施上述方法的一個範例裝置。其包括 附加於一個相機100的一個GPS接收器200。接收器之附加 可為内部的（内建的）或外部的。若接收器為外部的（例 如當接收器為可拆卸地附加於相機的一個配件時），那麼相 機之熱靴可較佳為用於附著。在任何情況下，知道接收器 天線210相對於相機之定向皆係必要的。 此實施例之GPS接收器200為一個擷取並處理接收 器。其包括用於降頻轉換衛星信號與取樣結果中間頻率 (IF)信號的一個GPS射頻（RF)前端220。其亦具有一個 14 201126189 微處理器230，純前端2職⑽星㈣ :存在-個記憶體240中。梢後，當接收器2〇。與： =(抑叫接時，便為了作處理而上傳所儲存的樣本。 此方=執仃針對位置估計的—般㈣處理，以及本發明之 接著U作為—個選擇，在某些擷取並處理實作中，pc將 這種产、、只身料給—個飼服器（未示於圖中）以作處理。在 月'下，方向估計可轉而在伺服器上執行。 有住何明中會清楚看出的此方法並不局限於具 的用於在 作或㈣的GPS接收11 (除了上文所述 、—個一致定向中將天線固定於相機之需求）。 與装11_以放大細節繪林發明之原理。其示出相機 具有方^之指向天線21◦。此天線之放射模式在水平面上 時的正:性(非等向)。亦即，當相機係以其針對拍攝照片 定向广向持著時，天線之增益會隨著方位角而改變。 中線210之主瓣粗略地以扇形波束4科示於第碥 主辦種方式來滿足天蚊社之_。較佳為天線 有利同的方向。這對於GPS信號接收來說係 此外，叫攝相片時，相機之前端通常不會被遮掩。 右將此指向天線之r波束」45〇取為錐形，那麼 t 就f在相機於橫向或縱向定向中被滿足。也就是 在實務上之指广性將不與沿著相機光轴之轉動相關。當然， 式，且ΐ t無法達到的一個完美的圓形對稱天線增益模 射杈式會藉由相機的90度旋轉而改變。 個感測器，如加速儀，而將此納入考量，其可檢測相 15 201126189 機相對於水平面之轉動。此等感測器普遍地被提供在數位 相機中，以指明所擷取之影像的格式。其他相機可具有手 動的縱向模式設定：此亦可用於推論天線放射模式。 依據由第3圖所指出的幾何結構，天線21〇將接收來自 於衛星401的一個相對較強信號以及來自於衛星402的一個 較弱信號°因此，藉由比較所量測與期望的信號強度，再 參考放射模式，此方法可推斷出相機朝向衛星401之方向較 多。知道此衛星之位置（並因此知道相對位置），便可將方 位角定向估計為「北」。 第3圆提供衛星相對於天線的幾何結構的一個簡化的 平面圖。在實務上，衛星401、402在天空中之仰角將對結 果產生衝擊。垂直於上頭的衛星提供少量方向資訊或是完 王不提供方向資訊。接近地平線的衛星最有可能被任意的 陸地障礙遮蔽（樹木、山丘、建築）。因此，在估計方向上 的最佳資訊係來自於位在天空中的適當仰角的衛星之量 測。 對於相片地理標籤應用來說，以羅經點來標記影像通 书便已足夠’而非數值角度。這表示針對此方法之準確度 而求並不特別高。例如，可依據8個羅經點（北、東北、東、 東南以±22.5度的期望誤差來估計方向。 在這些應用中，具有逼近半球形的放射（增益）模式 的天線疋有益的。這表示，增益在一個三維球形的一半上 觀的’且在對面的半球形上是明顯較低的。此模式在 接收5午多信號（例如為了準確的位置估計）的需求與天線 201126189 方向性（方向估計所需）之需求間提供一個 貼片天線可提供轉_增益 的協t 這種天線模擬成具有隨著從垂直於此補釘之 偏離的角度偏離，而漸進地減少之增益。亦即=:軸 是在垂直於此貼片的方向上，且在此貼片 軸呈90度）上步像s 十面（與垂直 個給定角度Μ在t個較低的值。對於從_偏離的一 °在％者此軸轉動時，增益幾乎是$ —μ 亦即，增益模式在此貼片平面上係圓形對稱的·的。 Ζ更添的指向天線加進此褒置，如第4圖所示。這增 料的信號強度量測數目（對於各個天線來說 & d個）並且能夠因而增加準確度及/或方向估計 2:健：4不於第4圖之範例中，示出-個第二天線 八八有疋向於相對於第一天線210之波束450為90度 地的個增4式或波束45&。第二天線21Ga將從衛星402 ^收個較強^號，並從衛星401接收-個較弱信號。這些 里、·】被加進由在第—天線2iQ處的信號量測所提供的限制 組。其可破用來驗證或難方位角估計。 、要、’°。由多個指向天線所提供的資訊有許多可能的作 法°下列範例係以遞增複雜度之順序來呈現： ()利用針對從在不同定向之不同天線中，所選出來的 個天線的量刪。例如基於天線定向（橫向或縱向）而選 擇看似最佳的天線。 (—)利用（ 合這些估計， 〜）之多個結果（各為一個獨立估計）並組 以產生一個更準確或更穩健的結果。此種範 17 201126189 例包括取平均值或取中間值。組合步驟可包括一個品質或 信賴量測，諸如由這些估計來從其平均所作的標準差。 (三）聯合地利用來自於各個信號與各個天線的信號強度 量測。例如，可將一個詳盡的搜尋用於估計方向，如以下 以放大細節來描述的。在此種方法中，可針對方位角定向 的一個給定理論，而產生估計針對各個信號與各個天線的 期望信號強度。這些估計可跨越多個天線來作比較與結 合，如同單一個指向天線所做的。 可考慮多個天線的某些特定優勢組態： (1) 指向於相對於另一個天線的90度方位的天線，如上文 已述者--例如5 —個向前而一個向左。這樣的安排給予 良好的衛星涵蓋範圍與方向涵蓋的多樣性。 (2) 如在（1)之情況中處於彼此之90度，但相對於相機 光軸來說是在傾斜角度的兩個天線，例如示於第5圖中者。 一個天線210c可能是指向此軸右邊（R)45度，而另一個210b 可能是指向此軸左邊（L) 45度。這些天線各以猜測放射模 式450c與450b示出。這種組態保有（1)之優點，但將天線 增益集中至相機前端，於此處很有可能可以檢測到最強信 號。於此方案中，這兩個天線之增益模式可有很大程度的 重疊——這可幫助確定衛星不在多個天線之主辦間之低增 益區域中「遺失」。 (3) 三個天線：兩個如在（2)中者與第三個垂直向上（U) 指，如傳統GPS相機天線。在橫向模式中，這個垂直天線 提供一個基準或參考信號位準，並有助於定位（下文以放 18 201126189 大細節討論）。當在縱向模式中時，天線u變成水平的 取決於相機轉動方向（順時針或逆時針）而指向卢邊〆〜 邊。左右天線，450b 450c，之一傾斜向前上方指，或右 :傾斜向前下方指。_，或肌之結合(典型二二 刖上方指的天線，L或R)可之後用於找出方向。 當包括多個天線時，可依據在許多傳統數位相機中 感測的相機之定向（橫向、縱性旋轉）來選擇他們的用斤 因此’例如在（3)之情況中，依據相紅定向來選擇^與只· Lmi;或其中之一皆可。這樣的選擇可藉由只對所還 天線作量測而㈣作出，或藉由儲存對所有天線之' 但只使用依據所感測之定向所選擇的天線量測二 處理中離線作出。 柏後 可提供多個接收器（即，多個GPSRF前端）以夕 個天線。然而，多個天線亦可共享-個GPSRF前端二 利用-個切換電路來達成，其允許較簡單便宜的硬體實 作。可依序（即連續地’而非同時地）量測各個天線之作 號強度王里响上’當依序執行時可能會有相機被移走 重新疋位的風險。因此，應該要盡量快速地取樣所有天^ 之L波強度’以使由相機移動所造成的量測變異不顯著。 l在决Η被按下時立即做出量測時尤其如此，因為攝影師 將會穩持相機以拍下照片。舉例來說，可能會花費觸毫秒 到1❼之間的時間來針對單—個天線作信號強度量測（或紀 錄足夠允魏行錢料料線樣本）。從一做線切換到 下個天線可月b會化費例如】毫秒到丄秒的時間。 19 201126189 注意’ GPS信號之高度指向接收會與—般的接收來自 盡s多的天線之號以估§·]*位置之需求衝突。為此，除了 指向天線以外’尚提供一個更添的全向天線可能會有所助 益。在本文中，全向指的是天線之放射模式實質上至少在 方位角平面上為等向的。這項準則係由在大部分現有的相 機内GPS實作中所使用的的面朝上的天線來滿足。第6圖示 出具有此種修改模式之裝置的一個範例。在這個安排中， 第3圖之指向天線210擴增具有一個等向放射模式455的一 個更添的天線215。天線215將兼接收來自於衛星4〇1、4〇2 之信號。此天線之輸出將因此更適於在步驟2〇與30中之位 置估計。 實務上，在方向估計上兼利用指向天線21〇與更添的等 向天線215是有益處的。例如，等向或向上指的更添天線可 針對各個發送器而提供一個參考信號位準。這可用來補充 或取代如上文所述的，在步驟40的信號位準之理論上的預 測。亦即，在全向天線接收的信號強度可用作在單向天線 所里測的彳§號強度之參考或預測器。這可以是很有用處 的，因為其可避免過度依賴理論模型，在實務上，理論模 型之假設可能不會被滿足。例如，雖然理論模型不能夠預 測由於障礙物而導致的信號衰減，但於全方位天線的信號 量測則會含蓄地將此類實際接收條件納入考量。 為求完整，現在將說明用於步驟4〇的一個信號強度預 測方法之範例。依據一個實施例，此方法依賴分別在步驟 20與30中所獲得的衛星位置以及相機位置。從這些位置， 20 201126189 而計算各個衛星相對於相機的方位角與仰角是很直接的。 個1實施例中’所接收的信號強度假設是依賴在衛星仰 角上（即’高於水平線之角度）。可依據由GPS衛星所發送 的定型波束模式之已知特性，而在這兩個變數間得出一個 理論上的關係（例如請見有名的著作：Kaplan所著之 Understanding GPS . Principles and Applications j 5 Artech House出版）；然而，亦可將其他因素納入考量，諸如障礙 可能性與衛星信號在飛行中之衰減。所接收之信號能量在 仰角上的變異之樣本值如下： 仰角 (度） 增益 (dB) 0 -8 10 -6 15 -4 25 -2 30 -1 50 0 對於大於50度（即從5〇度到9〇度）的仰角，增益再從零遞 減。易言之’可對於適當仰角的衛星期望最大信號強度。 如於上表之值中可看出的，所接收之信號強度在這些適當 角度附近些微改變。 可將這些增益值應用於一個正常的信號雜訊比 (SNR)’以給予針對在變化仰角上所接收之信號強度的期 望值。以找出在實務上可給予良好結果的一個正常的（預 堍的）47dB之SNR。在這個階段中，信號強度之預測並不 -、任何天線特性相關。例如，基於上文所提供之數字對 21 201126189 於在仰角25度之衛星的預測接收SNR會是45dB。 為求完整，現在將說明適於用在步驟50中的一個估計 演算法的一個範例。在這個實施例中，天線沿著相機之視 線（光軸）而指向。天線之指向放射模式係在一個圓形對 稱的半球形上模擬，其增益依據視線從光軸之角度偏離而 變化。針對一個貼片天線之樣本天線增益值如下： 偏離角度 (度） 天線增益 (dB) 0 0.0 10 -0.5 20 -1.0 35 -2.0 50 -4.0 75 -6.0 80 -7.0 85 -8.0 100 -9.0 110 -9.5 135 -9.0 150 -8.0 165 -7.5 170 -7.0 一個0度的偏離代表向衛星之視線與相機之光軸（指向方 向）重疊。一個9〇度的偏離會表示信號係於天片天線之平 面上（即，在其邊緣）接收。在偏離角度，A，與仰角，B， 以及方位角C之間的關係由下式給定：201126189 VI. Description of the invention: Technical field of the invention of the invention 3 Field of the Invention The present invention relates to estimating the direction in which the camera is facing when a photo is taken. In particular, it relates to cameras that include or connect satellite positioning receivers. c Prior Art; 1 Background of the Invention "Geographic labeling" refers to the location of a place or event of concern. Geotags are an increasingly common way to note and organize record information such as images and movies. By adding location metadata to the media, users can search or browse archives in a fun and intuitive way—for example, by marking a group of places on a map. In order to avoid the laborious work of manually typing location names or geographic coordinates that are necessary in generating such annotations, satellite positioning is increasingly being used to provide accurate location information. A satellite positioning system, such as the Global Positioning System (GPS), enables the receiver to calculate its position based on the measured arrival time of a signal from a group of orbiting satellites (referred to as spacecraft - SV). SUMMARY OF THE INVENTION In accordance with one aspect of the present invention, a method for estimating an azimuthal orientation of a camera having an additional antenna is provided, wherein the antenna has a predetermined directional radiation pattern that is held at the camera For a normal orientation of photography, which is non-isotropic in the azimuthal plane, the method includes the steps of: measuring the intensity of the plurality of signals received by the antenna from each of the plurality of transmitters 201126189, Positions of the plurality of transmitters <obtaining the position of the camera; predicting a desired intensity of each of the received signals; and measuring the position of the cameras and the positions of the plurality of transmitters as measured The signal strength, the antenna radiation pattern, and the predicted (four) intensity are used to estimate the orientation of the camera's financial position. This method allows the pointing direction of the camera to be estimated using a single pointing antenna. The antenna should be attached to the camera and be held in a normal position with a slap in the horizontal direction (the directional direction of the directional direction of the towel. This normal orientation usually includes one or two of the longitudinal or lateral orientation). This method uses the measurement of the signal strength of the transmitter or beacon from a known location to infer the orientation (also referred to as pointing) of the pointing antenna. This avoids the need to provide an electronic compass or other hardware; Therefore, for a camera that already has a positioning system, this method allows azimuth direction estimation at a small additional cost. The antenna used can be a simple pointing type and does not require multiple components or one phase. Array. Because the direction estimate is based on signal strength, there is no need to generally know, for example, to measure the phase or to explicitly determine the angle of arrival. For example, it is not necessary to rotate the antenna as is common in some other methods of radio direction finding to determine the angle of arrival or Is the strongest receiving direction. By using the camera position to determine the azimuth orientation, this method uses the traditional geo-tag solution The resulting information, thus minimizing the extra computational complexity. The azimuth or compass direction of the camera can be used to distinguish between photos taken in the same vantage point but in different directions. Therefore, it helps to distinguish different photos. Theme. This represents a significant improvement in the location of the camera when shooting this photo. 201126189 The transmitter can be a satellite, in which case the received signal is a satellite signal. Satellite positioning is commonly used in traditional geotagging applications. Medium. It is usually reliable and accurate, and provides global coverage. The position of the satellite is the necessary input for this location estimation process. Therefore, the only additional measurement required to implement this solution in the satellite positioning receiver, The measurement of the received signal strengths corresponding to the respective satellites. The step of obtaining the locations of the plurality of satellites may comprise at least one of the steps of: decoding a data message of the received satellite signal; and transmitting through a communication network Get satellite location information or satellite orbit information on the road. In the navigation application, the satellite position can be obtained by decoding the ephemeris and/or almanac information contained in the satellite transmission. In this scheme, the satellite position used in the method is already in the conventional satellite positioning receiver. A by-product of the implementation of the process. In other scenarios, it is preferred to obtain satellite ephemeris and/or almanac data from an independent source, such as a receiver. In this case, the location obtained may be more reliable. Because the centralized database maintains an accurate and complete satellite health record and corrects for any errors or anomalies. The step of obtaining the location of the camera can include the following steps: relying on a number of signals received from each of the plurality of transmitters And estimate the position of the camera. This is very efficient, because the same signal set used to obtain the camera position is also used to estimate the direction. This means that among these hardware parts, many 201126189 are available between jobs. shared. The transmitter can be a satellite, as discussed above, or can be a land beacon, such as a cell phone base station or a wireless local area network (WLAN) access point (AP). In either case, the receiver hardware used to receive the signal from the transmitter can be adapted to measure the signal strength through the pointing antenna. Optionally, the signals used to estimate the position of the camera are received by an additional antenna having a predetermined radiation pattern in which the camera is held in a normal orientation for photography. When in the azimuthal plane, it is isotropic. A pointing antenna having a radiation/gain mode that varies in a horizontal azimuth plane is necessary for the direction estimation process. However, in order to successfully estimate the camera position, the signal receiving action typically requires maximizing all received signal power across all transmitters. In other words, for position estimation, it is generally better not to have directional selectivity in the azimuthal plane. Providing an added antenna thus frees the direction estimate from reducing the robustness of the position estimate. The pointing antenna is used for direction estimation; the antenna is added for position estimation, which can be isotropic (omnidirectional) or weakly directed but pointing to the sky. Note that the signals received from the transmitter and used for each process are preferably the same signal to allow reuse of the same receiver hardware. Indeed, the addition of the antenna is likely to include the same antenna as used for direction estimation, but with a modified configuration. When an additional isotropic antenna is provided, the method may further comprise the step of measuring the strength of the signals received by the additional antenna, and the plurality of signals measured according to the added antenna. The intensity of the camera is estimated to be 6 201126189. , at the signal strength of the added antenna: where. Although it is possible to estimate the direction in all directions, for example, a difference pair with respect to this & = pair of t-bits can provide a ratio estimation process of 2 = _. This accuracy. ^ The larger X camera that may be provided by the predicted signal may have multiple additional antennas, a different financial direction radiation mode. The multiple antennas have a machine held in a positive direction radiation mode for photography. The phase is a non-isotropic, wherein the method comprises selecting one or more antennas from the 彳纟4 azimuth plane line; and measuring the selected ones from the plurality of days And a plurality of transmitter degrees, such as each 5 hai, and the amount of the material (mm) is the line radiation pattern of each of the selected antennas and the predicted signal strength. Day A Although this method is only applicable to one antenna, ^ can improve the robustness and / / accuracy. The direction material algorithm, Li Ying Z each antenna, or the measurement of signal strength can be combined across several antennas and algorithms used to aggregate data. In any of them, it is necessary to know the relative mutual orientation of these multiple antennas. This information is provided by the direction mode. The method may further comprise the steps of: sensing an elevation angle or a rotation angle of the camera and/or the antenna; and estimating the azimuthal orientation of the antenna according to the sensed angle. The step of sensing the elevation angle, or tilt angle, and the angle of rotation (e.g., about the optical axis of the camera) can be used to improve the method. This information can be combined with the antenna's transmit/gain mode to more accurately reconcile the predicted and measured signal strength. The method can also include estimating the azimuthal orientation of the antenna based on a predetermined polarization of the antenna. The polarization knowledge of the pointing antenna can help to discount the reflected signal when comparing the actual received signal strength to the desired signal strength. Therefore, this can also help to improve the robustness of this method. It is assumed that the original (transmitted) polarization of the signal is also known. The step of estimating the azimuthal orientation of the camera may be performed after an intentional non-zero delay following receipt of the signals. It is possible to perform direction estimation off-line, that is, at some point after actually receiving the signal. In this case, the signal or signal strength measurement can be simply measured for later analysis. This can, for example, minimize the amount of processing required to capture a photo. It is also particularly suitable for geotagging applications where location and/or direction information is typically not immediately required. The complexity of a portable device can be minimized by delay processing because the data can be analyzed at a different time and place by a more advantageous processor; battery power in the portable device can also be saved. According to a further aspect of the present invention, there is provided an apparatus adapted to estimate an azimuthal orientation of a camera, the apparatus comprising: an antenna attachable to the camera, the antenna having a predetermined directional radiation pattern that radiates the 201126189 mode When the camera is held in a normal orientation for photography, it is non-isotropic on the azimuthal plane; a receiver electrically connected to the antenna and adapted to be from a plurality of transmitters Each of the transmitters receives a plurality of signals; and a processing component adapted to: measure the strength of the received signals; obtain a position of each of the plurality of transmitters; obtain a position of the camera; Estimating the desired intensity of each of the signals; and estimating the orientation of the camera based on the obtained position of the camera and the positions of the plurality of transmitters, the antenna radiation pattern, and the predicted signal strengths Angle orientation. BRIEF DESCRIPTION OF THE DRAWINGS The invention will now be described by way of example with reference to the accompanying drawings, in which: FIG. 1 is a flow diagram of a method of direction estimation according to an embodiment; FIG. 2 is a diagram of an apparatus according to an embodiment. Block diagram; FIG. 3 illustrates the principle of a method and apparatus for estimating a direction according to an embodiment using a pointing antenna; FIG. 4 illustrates the principle of a method and apparatus for estimating a direction according to an embodiment using two pointing antennas; Figure 5 is a diagram of a further example configuration for two pointing antennas; Figure 6 illustrates the principle and method of direction estimation based on an embodiment utilizing a pointing antenna and an additional isotropic antenna . C. Embodiment 3 DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT The present invention has recognized that in geography photography, the location of the photo subject 201126189 is more important than the location of the camera itself. For example, a photographer might take two photos while standing in the same location, but the content of these photos may be very different, because the camera is geared toward different themes at these two times. Traditional geotags will be assigned to the same locations as these photos. Conversely, a photographer may take many photographs of a single subject of interest, for example, walking around a statue to obtain different perspectives. Traditional geotags will assign different locations to individual images. In order to distinguish between these different situations, it is necessary to record not only the location where the image was captured, but also the direction in which the camera was facing. This will enrich the photographer's geo-collection experience and allow for more advanced search and browsing capabilities for image collection. The orientation or facing direction of the camera at a given location has two degrees of freedom: the azimuth in the horizontal plane; and the elevation angle in the vertical plane. Azimuth is more important for photographic applications because it is the "compass direction" or the plane orientation on the surface of the earth. Knowledge of the azimuth is therefore necessary to determine what the camera is facing from a given known location. In any case, the elevation angle can be easily determined by an accelerometer sensor if needed. The present invention allows the azimuth of the orientation to be determined based on the measurement of the signal received by the camera. The signal for direction estimation is preferably the same signal as the signal used to determine the position of the camera; therefore, a position receiver built into or attached to the camera can provide location information in addition to location information. In addition, there is no need to provide additional hardware components to include a separate electronic compass with a magnetic domain sensor or an inertial navigation system (INS) that includes a rotary sensor. This allows the cost of the camera or camera accessory to be minimized when providing enhanced geographic listings. For example, the direction in which each photo was taken may be displayed on a map, or the photo subject may be automatically identified by reference to a Geographic Information System (GIS) database. The identification of the subject matter can be more accurately determined by also recording information about the distance from the camera to the subject matter. This information can be obtained directly, for example by a rangefinder (for automatic focus) in the camera, or indirectly, for example by lens focal length information. Figure 1 is a flow diagram of a method in accordance with an embodiment of the present invention. In this example, the GPS satellite positioning signal is used to obtain the azimuth of the address and estimated orientation. This method is suitable for determining the orientation of a camera equipped with a pointing GPS antenna. In step 10, the signal strength received from a number of satellites is measured. The G P S constellation contains at least 24 satellites that continuously transmit data signals in different orbits. If the GPS receiver can successfully calculate its range from four satellites, it can generate a location anywhere on the earth. The accuracy of positioning increases as the number of visible satellites increases. Measuring the intensity of each satellite signal is easier than completing a range calculation; therefore, it is possible to measure the signal strength for more satellites than used for position estimation—for example, errors in decoding satellite data messages. It is exempt from range calculation, but signal strength estimates are still available. The signal strength can be measured by any suitable means, such as by using a power signal to noise ratio (PSNR). Absolute signal strength values are not necessary in the direction estimation process; what is important is the relative strength across the satellite group. In step 20, the satellite position is obtained. This is done in the - part of the traditional GPS position estimation procedure, and as such, this information will be available before the camera position is known. Different GPS receivers can be used in different ways to get the security button. A common practice is to decode the data message in the satellite broadcast itself to obtain ephemeris and almanac information describing the satellite trajectory. Alternatively, in Auxiliary GPS (AGPS), this information is provided via separate channels such as the mobile phone network. In another Gps implementation called "Capture and Process Later" (hereafter referred to as "Capture and Process"), it is not necessary to provide star dive and almanac data to the receiver. However, short sequences transmitted by the satellite are received and stored by the receiver; later, they can be uploaded to another device for processing to derive a position estimate. The device performing the post-harvest processing can obtain satellite location information from a central repository of historical silk data, for example, via the Internet. The present direction estimation method is not limited to any of the specific components of the satellite position obtained in step 20, and thus can be applied to any GPS implementation. When satellite location data is to be obtained from satellite data, there are a number of options that allow for varying degrees of accuracy and robustness. In the best case, if the orbital track of the details of the star mill is available, then each satellite data signal can be completely decoded. However, some satellite signals may be too weak for the decoding (noisy). Although the signal strength may still be used in the direction estimation & ten processing, as described above > Even if the ephemeris cannot be decoded from a given satellite data message 'it is still possible to decode the slice. This provides approximate trajectory data' And: It can be accurate enough to support the direction estimation in practice. True, the information about all the charms is broadcast by all satellites 'to make a strong receipt of the order _ star signal over other satellites 12 201126189's almanac message to inform him that it is not necessary. In the position of the stepper (ie camera position). Again, this is done in the Gps = positive ^ column, and so - come, not f Any additional processing = vine phase fine position estimation method is also widely known and widely available. Again: People declare that 'the invention is not limited to any special (four) type of GPS implementation. Although steps 1〇, 20 and 30 do not need to Any special in real-time GPS implementation, Yi Bo V-production, but the first step in the strength measurement step H), followed by 彳: = Γ ... is beneficial. This is because the signal strength measurement 攸Camera bonding can be more integrated than positioning Wei. The digital camera will produce a (ΕΜ) interface in the processing of image lung and memory memory, which can reduce the irrigation of the received satellite signals. Similarly, in- Na also handles the implementation of t, (when the camera is "noisy"), it is advantageous to measure the signal strength before the position of the spot after the spot storage is estimated. In step 40, the method predicts the observed intensity of the observations at the location of the receiver when the phase is taken, given the location of each satellite. (d) The transmitted signal power is known and/or assumed to be π for all _ stars. The attenuation of the signal will depend on the distance from each satellite to the receiving benefit' and the atmospheric effects will also cause the attenuation to change for the star at different elevation angles in the sky, as observed by the receiver. The modules used to predict the expected power of the money can be dense or secret. The more accurate the module is, the higher the accuracy of the actual direction estimate. For example, a trivial model can assume that approximately equal power should be received from the financial asset. More complex karyotypes may take into account the elevation angle of each satellite in the sky. The actual signal measurement can be taken into account or used by itself as an alternative to purely theoretical predictions. In any case, the goal is to provide a set of reference signal levels from which the actual signal strength measurements from the pointing antenna can be compared. In step 50, the direction in which the receiver (camera) antenna is pointing is estimated. The expected signal strength and the actual observed signal strength are related by the receive mode or gain mode (also known as the radiation mode) of the pointing antenna and the direction in which the signal strength is measured. The radiation pattern can be obtained and expressed in a variety of ways. It includes previously quantified gains at different beam angles; approximations made by a debug theory or numerical model (such as an ideal patch antenna); or known design characteristics of a particular antenna used by the reference. In the case of knowing the characteristics of the radiation model, the remaining unknown parameters are related to the camera orientation. The estimation of the orientation is thus achieved by comparing the predicted and measured signal strengths with the reference radiation module and solving the orientation parameters. Figure 2 shows an exemplary apparatus suitable for implementing the above method. It includes a GPS receiver 200 attached to a camera 100. The addition of the receiver can be internal (built-in) or external. If the receiver is external (e.g., when the receiver is detachably attached to an accessory of the camera), the hot shoe of the camera may preferably be used for attachment. In any event, it is known that the orientation of the receiver antenna 210 relative to the camera is necessary. The GPS receiver 200 of this embodiment is a capture and processing receiver. It includes a GPS radio frequency (RF) front end 220 for downconverting satellite signals and sampling intermediate frequency (IF) signals. It also has a 14 201126189 microprocessor 230, pure front-end 2 (10) stars (four): exist in a memory 240. After the tip, when the receiver 2 〇. And: = (when the call is terminated, the stored sample is uploaded for processing. This party = the general (four) process for position estimation, and the U of the present invention as a choice, in some capture And in the implementation of the implementation, pc will give this production, and only the body to a feeding device (not shown in the figure) for processing. Under the month, the direction estimate can be transferred to the server. This method, which will be clearly seen by He Mingzhong, is not limited to the GPS receiver 11 used in the work or (4) (in addition to the above, the need to fix the antenna to the camera in a uniform orientation). The details of the invention are shown in detail. It shows that the camera has a pointing antenna 21◦. The radiation pattern of this antenna is positive (non-isotropic) when it is in the horizontal plane. That is, when the camera is aimed at shooting When the photo orientation is held wide, the gain of the antenna will change with the azimuth angle. The main lobe of the midline 210 is roughly represented by the fan beam 4 in the third mode of the host to satisfy the Japanese mosquito net. The antenna has the same direction. This is for the GPS signal reception. In the case of a film, the front end of the camera is usually not obscured. Right, the r beam "45" pointing to the antenna is tapered, then t is satisfied in the horizontal or vertical orientation of the camera. The breadth will not be related to the rotation along the optical axis of the camera. Of course, a perfect circular symmetrical antenna gain mode that cannot be achieved by the model will be changed by the 90 degree rotation of the camera. This is taken into account, such as an accelerometer, which detects the rotation of the phase 15 201126189 relative to the horizontal plane. These sensors are commonly provided in digital cameras to indicate the format of the captured image. The camera can have a manual longitudinal mode setting: this can also be used to infer the antenna radiation pattern. According to the geometry indicated by Figure 3, the antenna 21 will receive a relatively strong signal from the satellite 401 and from the satellite 402. A weaker signal. Therefore, by comparing the measured and expected signal strengths with reference to the radiation pattern, the method can infer that the camera is moving in more direction toward the satellite 401. Knowing the satellite The position (and hence the relative position) can be used to estimate the azimuth orientation as "north." The third circle provides a simplified plan view of the geometry of the satellite relative to the antenna. In practice, the satellites 401, 402 are in the sky. The elevation angle will have an impact on the result. Satellites perpendicular to the upper head provide a small amount of direction information or the king does not provide direction information. Satellites close to the horizon are most likely to be obscured by arbitrary land obstacles (trees, hills, buildings). The best information in the estimated direction comes from the measurement of the satellite at the appropriate elevation in the sky. For photo geotagging applications, marking the video with a compass point is enough 'not a numerical angle. It is not particularly high for the accuracy of this method. For example, it can be based on 8 compass points (North, Northeast, East, and Southeast with ±22. A 5 degree expected error is used to estimate the direction. In these applications, antennas with a near-hemispherical radiation (gain) mode are beneficial. This means that the gain is seen in the half of a three-dimensional sphere and is significantly lower on the opposite hemisphere. This mode provides a patch antenna that provides a turn-gain ratio between the need to receive multiple signals for 5 noon (eg, for accurate position estimation) and the need for antenna 201126189 directionality (required for direction estimation). There is a gain that progressively decreases with an angular deviation from a deviation perpendicular to the patch. That is, the =: the axis is in the direction perpendicular to the patch, and the patch axis is 90 degrees). The upper step is s ten sides (with a vertical given angle Μ at t lower values. The deviation of _1° in the case of % rotation of this axis, the gain is almost $_μ, that is, the gain mode is circularly symmetrical on the patch plane. ΖAdditional pointing antenna is added to this device, As shown in Figure 4. The number of signal strength measurements for this feed (for each antenna & d) and can thus increase the accuracy and / or direction estimate 2: Health: 4 is not in the example of Figure 4 In the middle, it is shown that the second antennas eighty-eight are oriented with respect to the beam 450 of the first antenna 210 at a degree of 90 degrees or the beam 45 & the second antenna 21Ga will be from the satellite 402 ^ Receive a stronger ^ and receive a weaker signal from the satellite 401. These are added to the restricted set provided by the signal measurement at the antenna 2iQ. It can be broken for verification or Difficult azimuth estimation. , ,, '°. There are many possible ways to provide information from multiple pointing antennas. The following examples are in the order of increasing complexity. Presentation: () utilizes an amount of antennas selected from different antennas in different orientations, for example based on antenna orientation (lateral or vertical) to select the best-looking antenna. (-) Utilization It is estimated that ~) multiple results (each of which is an independent estimate) and grouped to produce a more accurate or robust result. Such a model 17 201126189 includes averaging or taking intermediate values. The combining step may include a quality or Reliability measurements, such as the standard deviation made from these estimates from their average. (iii) Jointly utilizing signal strength measurements from individual signals and individual antennas. For example, an exhaustive search can be used to estimate direction, such as The following is described in an enlarged detail. In this method, the desired signal strength for each signal and each antenna can be estimated for a given theory of azimuthal orientation. These estimates can be compared across multiple antennas. Combining, as a single pointing antenna does. Consider some specific advantages of multiple antenna configurations: (1) pointing to another day The 90-degree azimuth antenna, as already mentioned above - for example, 5 - forward and one to the left. This arrangement gives good satellite coverage and diversity of coverage. (2) as in (1) In the case of two degrees at 90 degrees to each other, but at an oblique angle with respect to the optical axis of the camera, for example as shown in Figure 5. An antenna 210c may be pointing 45 degrees to the right (R) of the axis, and The other 210b may be pointing 45 degrees to the left (L) of this axis. These antennas are shown in guess radio mode 450c and 450b. This configuration retains the advantages of (1) but concentrates the antenna gain to the camera front. It is very likely that the strongest signal can be detected. In this scenario, the gain modes of the two antennas can overlap to a large extent - this helps to determine that the satellite is not missing in the low gain region between the hosts of multiple antennas. "." (3) Three antennas: two as in (2) and a third vertical upward (U) finger, such as a conventional GPS camera antenna. In landscape mode, this vertical antenna provides a reference or reference signal level and aids in positioning (discussed below in the large details of 2011 26189). When in portrait mode, the antenna u becomes horizontal depending on the direction of camera rotation (clockwise or counterclockwise) and points to the side of the rim. Left and right antennas, 450b 450c, one tilted forward and upward, or right: tilted forward and downward. _, or a combination of muscles (typically the antenna above the dioxin, L or R) can then be used to find the direction. When multiple antennas are included, their orientation can be selected based on the orientation (transverse, longitudinal rotation) of the cameras sensed in many conventional digital cameras. Thus, for example, in the case of (3), depending on the red orientation Choose ^ and only · Lmi; or one of them. Such selection may be made by measuring only the returned antennas (4), or by storing the antenna measurements for all antennas but using only the antenna measurements selected according to the sensed orientation. Cypress can provide multiple receivers (ie, multiple GPSRF front ends) with eve antennas. However, multiple antennas can also be shared - a GPSRF front end 2 is achieved with a switching circuit that allows for a simpler and cheaper hardware implementation. The intensity of each antenna's strength can be measured sequentially (i.e., continuously, rather than simultaneously). When executed sequentially, there may be a risk that the camera will be removed and re-clamped. Therefore, the L-wave intensity of all days should be sampled as quickly as possible so that the measurement variation caused by camera movement is not significant. This is especially true when the measurement is made immediately when the decision is pressed, because the photographer will hold the camera steady to take the picture. For example, it may take time between milliseconds and 1 来 to measure the signal strength for a single antenna (or the record is sufficient for the sample line). Switching from one line to the next can cost a month, for example, from milliseconds to leap seconds. 19 201126189 Note that the height of the GPS signal points to the reception and the general reception of the number from the antenna that is used to estimate the position of the §·]* position. For this reason, it may be helpful to provide an additional omnidirectional antenna in addition to pointing to the antenna. In this context, omnidirectional means that the radiation pattern of the antenna is substantially isotropic at least in the azimuthal plane. This criterion is met by the face-up antenna used in most existing in-camera GPS implementations. Fig. 6 illustrates an example of a device having such a modification mode. In this arrangement, the pointing antenna 210 of Fig. 3 amplifies an additional antenna 215 having an isotropic radiation pattern 455. The antenna 215 will also receive signals from the satellites 4〇1, 4〇2. The output of this antenna will therefore be more suitable for position estimation in steps 2 and 30. In practice, it is advantageous to utilize both the pointing antenna 21 and the additional isotropic antenna 215 in direction estimation. For example, an add-on or upward pointing antenna can provide a reference signal level for each transmitter. This can be used to supplement or replace the theoretical prediction of the signal level at step 40 as described above. That is, the signal strength received at the omnidirectional antenna can be used as a reference or predictor for the intensity of the 彳§ measured in the unidirectional antenna. This can be useful because it avoids over-reliance on theoretical models, and in practice, the assumptions of theoretical models may not be met. For example, although theoretical models are not able to predict signal attenuation due to obstacles, signal measurements on omnidirectional antennas implicitly take into account such actual reception conditions. For completeness, an example of a signal strength prediction method for step 4〇 will now be described. According to one embodiment, this method relies on the satellite position and camera position obtained in steps 20 and 30, respectively. From these locations, 20 201126189 it is straightforward to calculate the azimuth and elevation of each satellite relative to the camera. The received signal strength assumption in the 1 embodiment is dependent on the satellite elevation angle (i.e., 'the angle above the horizontal line'). A theoretical relationship can be derived between these two variables based on the known characteristics of the stereotyped beam pattern transmitted by the GPS satellite (see, for example, the famous book: Understanding GPS by Kaplan). Principles and Applications j 5 Artech House); however, other factors can be taken into account, such as the likelihood of obstacles and the attenuation of satellite signals in flight. The sample values of the variation of the received signal energy at the elevation angle are as follows: Elevation angle (degrees) Gain (dB) 0 -8 10 -6 15 -4 25 -2 30 -1 50 0 For greater than 50 degrees (ie from 5 degrees At an elevation angle of 9 degrees, the gain is further decremented from zero. It is easy to say that the maximum signal strength is expected for satellites with appropriate elevation angles. As can be seen in the values in the above table, the received signal strength changes slightly around these appropriate angles. These gain values can be applied to a normal signal to noise ratio (SNR)' to give a desired value for the received signal strength at varying elevation angles. To find a normal (predicted) 47dB SNR that can give good results in practice. In this phase, the prediction of signal strength is not - and any antenna characteristics are related. For example, based on the number pair 21 201126189 provided above, the predicted received SNR for a satellite at 25 degrees elevation angle would be 45 dB. For completeness, an example of an estimation algorithm suitable for use in step 50 will now be described. In this embodiment, the antenna is pointed along the line of sight of the camera (optical axis). The pointing radiation pattern of the antenna is simulated on a circular symmetrical hemisphere whose gain varies depending on the line of sight deviating from the optical axis. The sample antenna gain values for a patch antenna are as follows: Offset angle (degrees) Antenna gain (dB) 0 0. 0 10 -0. 5 20 -1. 0 35 -2. 0 50 -4. 0 75 -6. 0 80 -7. 0 85 -8. 0 100 -9. 0 110 -9. 5 135 -9. 0 150 -8. 0 165 -7. 5 170 -7. 0 A 0 degree deviation represents the overlap of the line of sight of the satellite with the optical axis (pointing direction) of the camera. A 9 degree deviation will indicate that the signal is received on the plane of the antenna (i.e., at its edge). At the deviation angle, the relationship between A, elevation angle, B, and azimuth angle C is given by:
cos A = cos β cos C 方位角，c，為在相機之方位角定向與衛星之方位角間之水 平方位角之差值。 22 201126189 此範例方法由測試依靠此資料與此模型之假設繼續。 即，此方法在多個離散的方位角上，針對各種情況將所期 望的與所量測的信號SNR值作比較，而執行一個搜尋。給 予所量測之資料與理論模型最佳配對的角度被判定為方位 定向之實際角度。在本實施例中，此方法估算8種情況，即 8個羅經點：N、NW、W、SW、S、SE、E、NE。對於各 個情況以及各個衛星，衛星之角度偏離係由仰角，B，以及 方位差，C，來算出的，如上文所述。可之後基於角度偏離 k表中讀得天線增益。增益值被加到在步驟4〇中所判定的 預測SNR (如上文所述）。如將從上文之說明回憶起的__ 在本實施射，歸驟4〇巾所產生的SNR關與天線增益 獨立。因此，若相機正指向正受測試的方向，則將「等向」 SNR與方向增益結合可為此魅提供針對射聊的一個 明確估計。最終估計係針_同魅，而與所量測之snr 作比較，並且結合針對财魅之錯誤。於本實施例中， 在最終估計與所量.值_絕對差域加總。在這個計 算被針對_财向重叙後，可得«小總賴對差值之 方向便被選擇為方向估計。 對於夕個才曰向天線’如之前提過的，可橫越所有所使 用的天線而加總在估計與量測SNR間之絕對差值 顯然地，徹底搜尋並非為可產生方向估計的唯一最佳 ==,範_此並〜何特定最佳化演 注意，預測信號位準可能是針對實際上並未檢測之衛 23 201126189 星來產生。因此，可能無法得到所有由此模型所預測之衛 星的真實信號強度量測。可將這些衛星忽略；然而，完全 /又有所期望之接收k號的情況對於方向估計演算法來說卻 是有幫助的。可簡單地糾㈣錢衛星峰意設定一個 相對低的「臨界」SNR來涵括這種選擇性的資訊。例如， 在本實施财，制了―個測的「臨界」SNR。此方法 之後繼續像之前-樣，取代—個真實量測的值而將所 期望的接收SNR與此臨界值作比較。 目的衛星--是 ，而非加總絕對 另一種作法其含蓄地避免相異數 要計算平均絕對差值，或其他正規化的值 值。 型，盆且有㈣方法依鼓線增料—個彳目對基柄 =更細即來獲得，那麼更精確的方向估計合是有可 ::，，隨著仰角之變異可與在方位角;二 異相異--即，增益模式可不以此轴為^ 模式在這兩個維度中是不同的，那麼相==Cos A = cos β cos C Azimuth, c, is the difference between the azimuth of the camera and the azimuth of the satellite. 22 201126189 This sample method is continued by testing relying on this material and the assumptions of this model. That is, the method performs a search on a plurality of discrete azimuth angles in comparison with the measured signal SNR values for various situations. The angle at which the measured data is optimally paired with the theoretical model is determined as the actual angle of the azimuth orientation. In this embodiment, the method estimates eight conditions, namely eight compass points: N, NW, W, SW, S, SE, E, NE. For each case and for each satellite, the angular deviation of the satellite is calculated from the elevation angle, B, and the azimuth difference, C, as described above. The antenna gain can then be read based on the angular deviation k table. The gain value is added to the predicted SNR determined in step 4 (as described above). As will be recalled from the above description, the SNR generated by the wiper is independent of the antenna gain. Therefore, if the camera is pointing in the direction being tested, combining the "isotropic" SNR with the direction gain provides a clear estimate of the sneak peek for this gaze. It is finally estimated that the needle is the same as the measured snr, and combined with the error for the charm. In the present embodiment, the final estimate is summed with the quantity. After this calculation is re-stated for the _ financial direction, the direction of the small difference is chosen as the direction estimate. For the antenna, as mentioned earlier, the absolute difference between the estimated and measured SNR can be summed across all the antennas used. Obviously, the thorough search is not the only one that can produce the direction estimate. Good ==, Fan _ This and ~ specific optimization performance attention, predictive signal level may be generated for the actual 23 23 26 26 stars. Therefore, the true signal strength measurements of all satellites predicted by this model may not be available. These satellites can be ignored; however, the full/desired reception of the k-number is helpful for the direction estimation algorithm. It is possible to simply correct (4) the money satellite peak to set a relatively low "critical" SNR to cover this selective information. For example, in this implementation, a "critical" SNR is measured. This method then continues to compare the expected received SNR with this threshold as before, instead of the true measured value. The destination satellite -- yes , not the total absolute. Another way to implicitly avoid the difference is to calculate the average absolute difference, or other normalized value. Type, basin and (4) method according to drum line feeding - one eye to base handle = finer to get, then the more accurate direction is estimated to be::,, with the elevation angle can be related to the azimuth Two different differences - that is, the gain mode may not be the axis of the ^ mode is different in these two dimensions, then the phase ==
向間之旋敎將會料酬純錢· P 素，如攝影師靠近相機的頭及/或手之“二將其他 中。無論如何，已發現上文所描述 良好結果。其亦易於實施，而這是_料=。予貫務上 的：ΓΓ實施例利用指向與相機鏡頭相同之方 然而，若提# 2上文已制提㈣理由，此為較佳的 k供—個面向前方的天線是不可能的— 24 201126189 為設計限制，那麼此方法仍可與在其他定向中之天線作 用，只要在拍攝照片時，天線被安排成在方位角方向中指 向。例如，一個在相機是橫向定向時面向左邊或右邊的天 線將在此定向中具有所說明的指向性質。然而，當相機被 (順時針或逆時針地）轉成縱向定向時，因為天線將會垂 直向上指或是垂直向下指，所以這些方向性的需求將不被 滿足。在這種情況中，一種解決方法是讓攝影師在拍攝縱 向照片時拍攝至少一張橫向的照片，以允許方向估計。 如早在先前所提過的，實際量測的與預測的信號強度 係由於天線增益模式而相關。天線的一個額外的明顯參數 為其偏振。GPS衛星發送之信號為右手圓形偏振。接收器 天線增益將因此依據天線之偏振，以及信號之到達方向而 變化。由於信號偏振會因為反射而被修改這樣的事實，情 況因而變得更複雜；然而，若將此納入考量，那麼可將其 用來從方向估計計算中消除反射信號。例如，若天線是右 手圓形偏振的，那麼具有不同偏振的反射信號將會相對衰 減，而直接接收的信號將會相對增強。再一次地，如定向 選擇性一般，這種偏振選擇性與某種程度的收集足夠用於 執行位置估計的衛星信號之目標有所牴觸。如前文所述， 對於針對定位的GPS信號接收來說，這項缺點可藉由提供 一個更添的天線來克服。 除了此方向估計方法的這些優點外，不可避免地會有 使方向估計錯誤、含糊不清或不可能達成的情況。這典型 上係因為沒有檢測到任何衛星（如在室内或地底下）而發 25 201126189 生的；或是因為天線的視野被限制了，使得衛星信號只能 從一個方向接收（例如在拍攝一楝建築物窗外的相片時）。 一種特別困難的情況可能會在面對建築物時發生：在這種 情況中，天線放射模式（與攝影師之存在）將阻礙在相機 後面的信號之接收；同時，建築物阻礙向前信號之方向接 收。因此，接收到的唯一信號可能會是從後面抵達，但在 前面被建築物反射。這可導致與真正的方位角定向相反的 方向估計。 針對這些情況，並且為了普遍改善可用性，提供一個 使用者介面以允許攝影師在多種可能性之間選擇、糾正錯 誤的方向估計以及/或是在不可獲得估計時手動輸入方 向，可大有助益。 上文所說明之實施例已然全為集中在利用一個衛星定 位系統之實作。然而，如熟於此記者可迅速明顯看出的， 此方法之範圍並非局限於其應用。此方法普遍地使得能夠 從一個已知位置得來方向估計，只要所接收的信號強度可 針對多個其位置亦為已知的發送器量測。例如，可利用手 機基地台作為參考發送器，而應用此方法。在一個實施例 中，各個基地台之地點係儲存在一個資料庫中。用以檢測 在一個蜂窩狀網路中的一個行動裝置之位置之方法係廣為 人知的—例如，可藉由基地台三角定位來判定接收器之 位置。此方法然後如上文所述地繼續；預測期望信號強度 並將其與來自於基地台的所量測之信號強度作比較；以及 最後利用已知的天線模式參數估計方向。 26 201126189 如此方向估計演算法的陸上實作之另一個範例，發送 器可為WLAN存取點。再次地，假如AP與接收器之位置為 已知，則此方法為同。 亦可結合上文所述之多樣陸上與基於衛星的作法，以 增加可得信號量測的數目與多樣性’因而更增進此系統之 準確度。 如熟於此技者將可迅速明顯看出的，在純理論上，沒 有任何真實的天線係完美地等向的。因此，可以說，由於 真實放射模型中之缺陷’任何特定可實現的天線皆係某種 程度的「非等向」。在本文中’熟於此技者亦將無疑地了解， 於本說明書與所附申請專利範圍中之對於「非等向」天線 與「等向」天線之扣涉，係意欲在實務上分別被解讀為代 表「實質上非等向」及「實質上等向」。例如，一個等向天 線可被定義為’在有關平面中’具有在最大與最小值間低 於3dB之增益變異的天線。同樣的，一個非等向天線可被定 義為，例如，在有關360度的角度範圍上，具有在最大與最 小増益值間多於6dB之差值之天線。本發明將提供對於明定 非等向天線來說的達到增益變異很大之程度，以及對於明 定等向天線來說的達到增益變異很小之程度的更佳的效 能。即’非等向天線應盡量為指向的，而等向天線應盡量 為非指向的。亦應注意，此方法之效能取決於於給定實施 例中所使用的各個天線之有效增益。這將取決於諸如天線 與相機之設計，以及其他鄰近物品之存在，諸如攝影師之 手或頭，等因素。 27 201126189 對於熟於此技者來說，多種其他修改體將係明顯可見 的。 【圖式簡單說明3 第1圖為依據一實施例的一個方向估計方法之流程圖； 第2圖為依據一實施例的一個裝置之方塊圖； 第3圖繪示依據利用一個指向天線之一實施例的方向 估計方法與裝置之原理； 第4圖繪示依據利用兩個指向天線之一實施例的方向 估計方法與裝置之原理； 第5圖為針對兩個指向天線之更進一步的範例組態之 圖示； 第6圖繪示依據利用一個指向天線與一個更添的等向 天線之一實施例的方向估計方法與裝置之原理。 【主要元件符號說明】 10-50…步驟 230…微處理器 100.. .相機 240...記憶體 200.. .GPS接收器 300...個人電腦（PC) 210、210a、210b、210c、215·.· 4CU、402...衛星 天線 450、450a、450b、450c.··波束 220.. .前端 455...等向放射模式 28The inter-circle will be paid for pure money, P, such as the photographer's proximity to the camera's head and / or hand "two other. In any case, the good results described above have been found. It is also easy to implement, And this is _ material =. For the transaction: ΓΓ The embodiment uses the same direction as the camera lens. However, if #2 has been made above (4), this is the better k for a front-facing Antennas are not possible — 24 201126189 For design limitations, this method can still be used with antennas in other orientations, as long as the antenna is arranged to point in the azimuthal direction when taking a photo. For example, one is horizontal in the camera The antenna facing left or right when oriented will have the indicated pointing properties in this orientation. However, when the camera is turned (clockwise or counterclockwise) into portrait orientation, because the antenna will be vertically pointing upwards or vertically Under the finger, so these directional needs will not be met. In this case, one solution is to let the photographer take at least one horizontal photo when taking a portrait photo to allow direction estimation. As previously mentioned, the actual measured and predicted signal strength is related to the antenna gain mode. An additional significant parameter of the antenna is its polarization. The signal transmitted by the GPS satellite is a right-hand circular polarization. Receiver antenna The gain will therefore vary depending on the polarization of the antenna and the direction of arrival of the signal. The situation becomes more complicated due to the fact that the signal polarization is modified by reflection; however, if this is taken into account, it can be used Eliminate the reflected signal from the direction estimation calculation. For example, if the antenna is circularly polarized by the right hand, the reflected signals with different polarizations will be relatively attenuated, while the directly received signals will be relatively enhanced. Again, such as directional selectivity In general, this polarization selectivity is inconsistent with some degree of collection of satellite signals sufficient for performing position estimation. As mentioned earlier, this drawback can be exploited for GPS signal reception for positioning. Provide an added antenna to overcome. In addition to these advantages of this direction estimation method, it will inevitably A situation in which the direction is estimated to be erroneous, ambiguous, or impossible to achieve. This is typically because no satellites were detected (eg, indoors or under the ground); or because the field of view of the antenna is limited, Allows satellite signals to be received only in one direction (for example, when taking a photo outside a building window). A particularly difficult situation may occur when facing a building: in this case, the antenna radiation pattern (with photography) The presence of the division will hinder the reception of signals behind the camera; at the same time, the building blocks the reception of the forward signal. Therefore, the only signal received may arrive from behind but be reflected by the building in front. Leading to direction estimation opposite to true azimuth orientation. For these situations, and to generally improve usability, a user interface is provided to allow the photographer to choose between multiple possibilities, correct the wrong direction estimate, and/or not It is helpful to manually enter the direction when obtaining an estimate. The embodiments described above are all focused on the implementation of a satellite positioning system. However, as is well apparent to those skilled in the art, the scope of this method is not limited to its application. This method generally enables direction estimation from a known location as long as the received signal strength can be measured for a plurality of transmitters whose locations are also known. For example, this method can be applied using a mobile base station as a reference transmitter. In one embodiment, the locations of the various base stations are stored in a database. Methods for detecting the location of a mobile device in a cellular network are well known - for example, the position of the receiver can be determined by base station triangulation. The method then continues as described above; predicting the desired signal strength and comparing it to the measured signal strength from the base station; and finally estimating the direction using known antenna mode parameters. 26 201126189 Another example of a land-based implementation of such a direction estimation algorithm, the transmitter can be a WLAN access point. Again, this method is the same if the AP and receiver locations are known. It is also possible to combine the various onshore and satellite-based practices described above to increase the number and diversity of available signal measurements' thus increasing the accuracy of the system. As will be apparent to those skilled in the art, in pure theory, no real antenna system is perfectly isotropic. Therefore, it can be said that any particular achievable antenna is a certain degree of "non-isotropic" due to defects in the real radiation model. In this article, it will be understood by those skilled in the art that the deduction of "non-isotropic" antennas and "isotropic" antennas in the scope of this specification and the appended claims is intended to be Interpreted as representing "substantially non-equal" and "substantially equal." For example, an isotropic antenna can be defined as an antenna having 'in the relevant plane' having a gain variation of less than 3 dB between the maximum and minimum values. Similarly, an anisotropic antenna can be defined, for example, as having an antenna with a difference of more than 6 dB between the maximum and minimum gain values over a range of 360 degrees. The present invention will provide a greater degree of gain variability for a given anisotropic antenna and a better performance for a given isotropic antenna to a degree that achieves a small gain variation. That is, the non-isotropic antenna should be pointed as far as possible, and the isotropic antenna should be as non-directional as possible. It should also be noted that the effectiveness of this method depends on the effective gain of the individual antennas used in a given embodiment. This will depend on factors such as the design of the antenna and camera, as well as the presence of other nearby items, such as the photographer's hand or head. 27 201126189 A variety of other modifications will be apparent to those skilled in the art. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a flow chart of a direction estimating method according to an embodiment; FIG. 2 is a block diagram of a device according to an embodiment; FIG. 3 is a view showing one of the pointing antennas according to one embodiment; The principle of the direction estimation method and apparatus of the embodiment; FIG. 4 illustrates the principle of the direction estimation method and apparatus according to an embodiment using two pointing antennas; FIG. 5 is a further example group for two pointing antennas Illustration of the state; Figure 6 illustrates the principle of the direction estimation method and apparatus according to an embodiment utilizing a pointing antenna and an additional isotropic antenna. [Description of main component symbols] 10-50...Step 230: Microprocessor 100.. Camera 240: Memory 200.. GPS Receiver 300... Personal Computer (PC) 210, 210a, 210b, 210c 215·.·4CU, 402...satellite antennas 450, 450a, 450b, 450c.··beam 220.. front end 455... isotropic radiation mode 28