201222424 六、發明說明： 【發明所屬之技術領域】 、 本發明是有關於一種影像辨識方法，特別是指一種利用物 件特徵相異性(Identification by Objects’ characteristics)進行多 個物件之動態影像辨識方法及使用該方法之系統。 【先前技術】 目前電視遊戲(TV game)或電腦遊戲(PC game)已是常見 的休閒娛樂方式，以一般電腦遊戲為例，多是在電腦裝置上安 • 裝遊戲軟體’再搭配電腦裝置的螢幕及例如鍵盤、滑鼠或搖桿 等輸入介面對遊戲軟體進行控制；然而亦有搭配廠商銷售之遊 戲軟體的互動式道具’為方便說明起見，以美國專利公開號 2004006348號所揭示的一種互動式遊戲裝置為例來說明其構 件及作用原理。 如圖1所示，一互動式遊戲裝置700具有二啞鈴狀的指示 道具71及72、一踏步台720、一螢幕裝置730、一視訊攝影機 750、一遊戲盒760及一主機裝置770。其中，主機裝置770安 • 裝有遊戲軟體；指示道具71、72分別由一玩家705的左右手 握持，且指示道具71、72上分別在其端部附有光源(Light s〇urce)711、712及721、722 ;螢幕裝置73〇可顯示遊戲軟體 中例如一虛擬的舞者731在跳舞的影像；主機裝置770可以是 電腦裝置或遊戲機台；螢幕裝置730與遊戲盒76〇係分別連接 在主機裝置770上。 J 使用上述的互動式遊戲裝置700進行跳舞遊戲時，玩家705 需開啟(Turn on)指示道具71、72的電源以使光源711、712及 721、722付以發光，讓視訊攝影機750可彳貞測光源711、712 201222424 及721、722的影像，進而由遊戲盒760計算該光源711、712 及721、722的位置等參數，最後將其輸入到主機裂置77〇來 完成玩家705手持指示道具71、72之光源7Π、712及72卜 722位置的追蹤，並顯示於螢幕裝置730之畫面上。 然而，假設玩家705在任意揮動現有道具71、72時，由 於光源71卜712及72卜722的特徵相同(例如均是等面積之圓 形)旦’因此任二光源7U、712及72卜722在遊戲盒所判讀 5V像同為m圖案’則若有任二圓形圖案執跡重疊後分離 ^視訊彡機750讀取影像後在賴盒·進行影像處理 刹。’ Γ易分辨出二者之差異而造成移動位置或移動軌跡的誤 與松此’有必要提出—種_更精確·影像巾物件的系統 【發明内容】 性進行多個::：二:’即在提供一種利用物件特徵相 於是以==態::辨使用該方法之系統， 來區分，因此不會因“分 個物件之動鮮彡像件特翻異性進行 物件ΐ=:的先線，加觸物=中藉⑴ 多個物件之動態影像= 徵相卿 一暫存器之使用，用射配合—影像感測⑸ 地進行辨識，影像感測器且令具有之至少-物件即日; 感挪器以該等感應像素“‘订列式感應像素，且該影ί! 这測該物件而形成複數影像區段，該；ί 201222424 法包含下述步驟：（A)投射光線以產生一影像，該光線帶有一圖 案；（B)根據一組曝光參數來感測該影像；〇設定該影像相對 於該組曝光參數之灰階閾值；(D)依序擷取該影像中每列之像素 值；（E)利用該灰階閾值判斷背景區域及識別出該物件具有之影 像區段；（F)利用相鄰兩列中影像區段之空間相關性分辨未知^ 件之影像區段屬於何物件；(G)匯集該等影像區段所累計之資訊 至其所屬之物件；以及(H)依據一判斷法則區分該物件之特徵屬 性’並該特徵屬性分辨不同的物件。201222424 VI. Description of the Invention: [Technical Field] The present invention relates to an image recognition method, and more particularly to a method for performing dynamic image recognition of a plurality of objects by using Identification by Objects' characteristics and A system using this method. [Prior Art] At present, TV games or PC games are common leisure entertainment methods. For example, in general computer games, most of them are installed on computer devices, and then with computer devices. Screens and input devices such as keyboards, mice or joysticks are controlled by the game software; however, there are also interactive props that match the game software sold by the manufacturer. For convenience of explanation, one disclosed in U.S. Patent Publication No. 2004006348 The interactive game device is taken as an example to illustrate its components and the principle of action. As shown in FIG. 1, an interactive game device 700 has two dumbbell-shaped indicating items 71 and 72, a stepping table 720, a screen device 730, a video camera 750, a game box 760, and a host device 770. The host device 770A is equipped with a game software; the indicator items 71, 72 are respectively held by the left and right hands of a player 705, and the indicator items 71, 72 are respectively provided with a light source (Light s〇urce) 711 at their ends, 712 and 721, 722; the screen device 73 〇 can display an image of a virtual dancer 731 dancing in the game software; the host device 770 can be a computer device or a game console; the screen device 730 and the game box 76 are respectively connected. On the host device 770. J. When the dance game is performed using the interactive game device 700 described above, the player 705 needs to turn on the power of the indicator items 71, 72 to cause the light sources 711, 712 and 721, 722 to emit light, so that the video camera 750 can be turned on. The images of the light sources 711, 712 201222424 and 721, 722 are measured, and then the parameters of the light sources 711, 712, 721, and 722 are calculated by the game box 760, and finally input to the host cleavage 77 完成 to complete the player 705 hand-held instruction items. The tracking of the 710 positions of the light sources 71, 72, and 72 is displayed on the screen of the screen device 730. However, it is assumed that when the player 705 arbitrarily swings the existing props 71, 72, the characteristics of the light source 71 712 and 72 722 are the same (for example, the circle is equal in area), so the two light sources 7U, 712, and 72 are 722. In the game box, the 5V image is the same as the m pattern, and if any of the two circular patterns are superimposed, the video camera 750 reads the image and then performs the image processing brake. ' Γ 分辨 分辨 分辨 分辨 分辨 分辨 分辨 分辨 分辨 分辨 分辨 分辨 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动 移动That is to provide a system that utilizes the feature of the object to be used in the == state:: to distinguish the use of the method, so that it does not cause the object ΐ =: the first line because of the special entanglement of the object , plus touch object = medium borrowing (1) dynamic image of multiple objects = use of the sacred phase of a register, using the ray fit - image sensing (5) to identify, the image sensor and at least - the object of the day; The sensor senses the pixel with the sensing pixels, and the image is formed by the image, and the image is formed by the image; the 201222424 method includes the following steps: (A) projecting light to generate a Image, the light has a pattern; (B) sensing the image according to a set of exposure parameters; 〇 setting a grayscale threshold of the image relative to the set of exposure parameters; (D) sequentially capturing each column of the image Pixel value; (E) use the grayscale threshold to determine the background area Identifying the image segment that the object has; (F) using the spatial correlation of the image segments in the adjacent two columns to distinguish the image segment of the unknown component; (G) assembling the image segments Information to the object to which it belongs; and (H) distinguishing the feature attribute of the object according to a judging rule and distinguishing the different objects.
本發明景>像辨識系統係利用物件影像特徵相異性對於一 中具有之至少—物件即時地進行辨識，該影像辨識系統 广-光源、-影像感測器、—類比數位轉換器、—影像處理 具暫存^ ’該光源投射具錢㈣紐；該影像感測器 渺祕仃狀感應像素，且該影像躺11㈣等感應像素感 = 區段;該類比數位轉換器連接該影: 該等感應:紊位轉換器1影像處理單元係逐列讀取 、並设定有該影像之灰階閾值及一用以八#你 =寺徵屬性的判斷法則;該暫存器連接該影二= Μ暫存該影像處设场俅處理早7C，用 藉此，該影像用等訊。 識別出該物件以，⑥7_喊階輯背景區域並 間相關性分辨未:像11段’利_鄰兩列中影像區段之空 像區段所累計之資^影像區段屬於何物件，再匯集該等影 該物件之特徵屬§、所屬之物件，並依據該酬法則區分 辨識出該影像中的該物件之特徵屬性。 201222424 具有圖案的光線有多種方式可以產生。例如，光源可以包 含多個發光元件，藉由發光元件的排列方式、不同發光時序、 不同發光頻譜、或以上兩者或更多者之組合，來產生具有圖案 的光線。或是，光源可以包含一或多個發光元件加上一個折射 光學元件(DGE，Diffiaetive Optical Element)及/#—個微鏡面 (MEMS mirror)’而發光元件透過該光學元件及/或該微鏡面 投射光線。 底下藉由具體實施例詳加說明，當更容易瞭解本發明之目 的、技術内容、特點及其所達成之功效。 【實施方式】 在本發明被詳細描述之前，要注意的是，在以下的說明内 ，中，類似的元件是以相同的編號來表示。此外，必須說明的 是，由於第一較佳實施例是用以區分實心、空心之特徵屬性， 而第一較佳實施例是用以區分長形、短形之特徵屬性，但是在 其他實施财’或可將上述實❸心、長形及短形之特徵屬 性加以混合使収朗，因此只要有上述特徵雜之應用，均 應屬於本發明概念欲保護之範疇。 如圖2所示，根據本發明，利用物件特徵相異性進行多個 物件之動態影像辨識方法可使用—影像處理系統3來實施，該 影像處理系統3具有一影像感測器(Imagesensor)31、一類比數 位轉換器（A/D C〇nverter)32、一影像處理單元（Image pr〇Cessor)33、一暫存器(Register)34 及一介面模組 ％。 其中，影像感測器31是CCD或CMOS元件製成，具有複 ，行列式感應像素，用以感應拍攝物(圖未示)反射的光^而成 。像像感測H 31以該等感應像素感測該物件而形成 201222424 用容後再述)’並轉換為類比訊號;接著，輸 ==:: 一轉換為數： S ^早33負屬大部分訊號的計算處理；影儍虛 素感32，舰_取料感應像 元======峨理單 藉此，影像處理單元33 件的影像資訊。 出物件之影像區段，利用閑值判斷背景區_ 性，而於擷取完該影像所有之像素值後，影像處理屬 辨識出影像令的物件之特徵屬性。 即 ^像處理系統3的介面模組35連接該影像處 識後之特徵屬性相關資訊輸出為符合 邊The image recognition system utilizes object image feature dissimilarity to instantly identify at least one object in the image, the image recognition system wide-light source, image sensor, analog-like digital converter, image The processing device temporarily stores ^ 'the light source projecting money (four) button; the image sensor is sensitive to the sensory pixel, and the image is lying 11 (four) and other sensed pixel sense = segment; the analog digital converter is connected to the shadow: Induction: the image processing unit of the turbulence converter 1 is read column by column, and the gray scale threshold of the image is set and a judgment rule for the attribute of the eight #你=寺征 property; the register is connected to the shadow two= Μ Temporarily store the image at the location of the field and process it 7C early, with which the image is equal. Recognizing the object, the 67_ shouting background area and the correlation between the two are not resolved: like the object segment of the image segment accumulated by the empty image segment of the image segment in the 11 segments Then, the feature of the object is §, the object to which it belongs, and the feature attribute of the object in the image is identified according to the reward law. 201222424 Patterned light can be produced in a variety of ways. For example, the light source may comprise a plurality of light-emitting elements, the patterned light being produced by the arrangement of the light-emitting elements, different illumination timings, different illumination spectra, or a combination of two or more thereof. Alternatively, the light source may include one or more light emitting elements plus a refracting optical element (DGE, Diffiaetive Optical Element) and a MEMS mirror, and the light emitting element transmits the optical element and/or the micro mirror surface. Project light. The details, technical contents, features, and effects achieved by the present invention are more readily understood by the detailed description of the embodiments. [Embodiment] Before the present invention is described in detail, it is to be noted that in the following description, like elements are denoted by the same reference numerals. In addition, it must be noted that since the first preferred embodiment is used to distinguish between solid and hollow feature attributes, the first preferred embodiment is used to distinguish between long and short feature attributes, but in other implementations. 'Or the characteristics of the above-mentioned solid heart, long shape and short shape can be mixed to make it suitable. Therefore, as long as the above-mentioned features are applied, it should belong to the scope of the present invention. As shown in FIG. 2, according to the present invention, a method for performing dynamic image recognition of a plurality of objects by using object feature dissimilarity can be implemented by using an image processing system 3 having an image sensor 31. An analog-to-digital converter (A/DC 〇nverter) 32, an image processing unit (Image pr〇Cessor) 33, a register 34 and an interface module %. The image sensor 31 is made of CCD or CMOS components and has complex and determinant sensing pixels for sensing the light reflected by the object (not shown). The image sensing H 31 senses the object with the sensing pixels to form 201222424 and then converts it into an analog signal; then, the input ==:: a conversion to a number: S ^ early 33 negative majority Signal calculation processing; shadow silly feeling 32, ship_receiving sensing pixel ====== processing single, thereby image processing unit 33 image information. The image segment of the object is judged by the idle value, and after all the pixel values of the image are captured, the image processing is the feature attribute of the object that recognizes the image. That is, the interface attribute 35 of the processing system 3 is connected to the feature attribute related information outputted by the image recognition as the matching edge.
主機41 M腦4之主機41具有的傳輸介面41卜由個人電腦4之 機：接收並加以運算後，即可在個人電腦 U •，、、貝不出該物件影像。 器42上 之辨明，影像處理系統3可用於攝錄影等取像裝置 辨識功此•，或是以安裝在電腦的辨識軟 能；另外’由於影像制器3丨、類比數式行，識功 33 主要概念是以影像處理單元33配合暫存器 〜本發月 識功能，因此以下將僅就相闕於本發明原理之部分；之辨 201222424 配合圖2、3所，說明本發明利用物件特徵相異性進行 多個物件之㈣雜_綠的第—練實酬。必須說明的 二在本較佳實_巾，喊潍感· 3丨係具有複數行列 式感應像素__，且該等像素M1係以逐列的方式感應各 物件1卜12，因此，將影像感測器31所感應到的—物件在每 :列中所得到的部分影像稱為—影像區段(ImageSeg_t)，各 影像區段之酬法’係逐列紀錄各财各影像區段之起始點並 儲存至暫存H 34;接著自該影像區段之起始點逐點累計該影像 區段之資訊並贿錢暫抑34;關斷各财各影像區段之 終點並儲存至暫存器34。 例如：影像處理系統3會先由影像感測器依序操取影 像1中每列像素311感應到的像素值經類比數位轉換器32轉 換為數位訊號輸入至影像處理單元33，讀取的方式是自第一列 開始:從左到右讀取該列中的每個像素值，每讀完一列再由上 而下讀取每列之各像素值，而判斷是否有物件之影像資訊的出 現，係偵測是否有大於一系統預設閾值之像 在讀取的同時’即可一併判斷各列中該等物件U、12之 影像區段起始點及終點在何處，如此便可接著利用相鄰兩列中 影像區段之空__(容後再述)分辨未知物件 於何物件。例如··自影像】中的第4列起始有物件之影像資訊屬 且該等影像資訊分屬於二物件〗卜丨2，因此從左而右，紀錄先 出現之影像區段111之一起始點ιη，並儲存至暫存器別，再 逐點累計影像區段111之資訊並齡至暫存器34，接著判斷該 列中具有該影像區段111之終點ιη，，並儲存至暫存器34後^ 再紀錄物件12在該列的影像區段121之起始點⑵，纽點 201222424 =ι”及其逐點累計之資訊於暫存器％ 斷，依此類推。 丹進仃下一列之判 而分辨該等影像區段分別屬於何物件！ 利用相鄰兩列中影像區段之空 之方式，即 區段屬於何物件’係判斷如符像 物件影像區段屬於該物件i: 幻了判疋一未知The host 41 of the host 41 M brain 4 has a transmission interface 41 which is received by the personal computer 4: after receiving and calculating, the image of the object can be displayed on the personal computer U. It is discerned on the device 42 that the image processing system 3 can be used for image capture device identification functions such as video recording, or is to identify the soft energy installed in the computer; and 'because of the image controller 3丨, the analog number, the line The main concept of the work 33 is that the image processing unit 33 cooperates with the temporary register to the present moonlight function, so that the following will only be part of the principle of the present invention; the identification 201222424, together with Figs. 2 and 3, illustrates the use of the object of the present invention. The characteristic dissimilarity carries out the (four) miscellaneous_green's first practice. The two must be described in the preferred embodiment, the screaming sensation 丨 具有 has a plurality of determinant sensing pixels __, and the pixels M1 sense each object 1 12 in a column-by-column manner, therefore, the image is The partial image obtained by the sensor 31 in each column is called an image segment (ImageSeg_t), and the image segment of each image segment is recorded in a row by column for each image segment. Start and save to the temporary storage H 34; then accumulate the information of the image segment point by point from the starting point of the image segment and bribe the money to suppress 34; turn off the end of each image segment and save to the temporary Memory 34. For example, the image processing system 3 firstly reads, by the image sensor, the pixel value sensed by each column of pixels 311 in the image 1 is converted into a digital signal input to the image processing unit 33 by the analog digital converter 32, and the manner of reading is performed. Starting from the first column: read each pixel value in the column from left to right, and read each pixel value of each column from top to bottom after reading one column, and judge whether there is any image information of the object. , detecting whether there is an image larger than a preset threshold of the system while reading, and then determining where the starting point and the ending point of the image segment of the objects U and 12 in each column are, so that Then, the space __ (described later) of the image segments in the adjacent two columns is used to distinguish the unknown object from the object. For example, in the fourth column of the self-image, the image information of the object starts and the image information belongs to the two object 丨 丨 2, so from the left to the right, the record begins with one of the image segments 111 that appear first. Click ιη, and save to the scratchpad, then accumulate the information of the image segment 111 point by point and age to the register 34, and then determine that the column has the end point ιη of the image segment 111, and store it to the temporary storage. After the device 34, the object 12 is recorded at the starting point (2) of the image segment 121 of the column, the dot 201222424 = ι" and its point-by-point information is broken in the register %, and so on. According to a list of judgments, the objects of the image segments belong to each other! The method of using the space of the image segments in the adjacent two columns, that is, the object belonging to the segment is judged as the image object segment belongs to the object i: Illusion
Seg-LSPrdine-〇bjrR ;且Seg-LSPrdine-〇bjrR; and
Seg-R^Preline-Objj-L ； 公式 ^ 表示例如在讀取至’的第Y列資料時； ㈣丄表轉取第丫列出現的該未知物 : 座標值；Seg.R表示讀取帛γ _ ； ^ Prdine-〇bji-R^-f Y J,- 即γ·1列出現之各該物件i之影像區 表示第w列出現之各該物件二= ^之左方起始點X座標值，若是符合Seg L他line傅R eg-R2Prdme-〇bjrL^判斷式，即表示該 段與第Y-1列出現之該物件i之影像區段屬於同—物^「像Q 如圖4所示’說明本發明利用物件特徵相異性進行多個物 件之動態f彡像_方法的二較佳實施财初料如何判斷影 像區段屬於何物件，其具有之步驟及作縣理詳述如下： 首先，設定該影像之灰階閾值(步驟10” ；接著依序操取 ，衫像中每列之像素值(步驟1Q2) ;利用該灰賴值判斷背景 區域(步驟103);識別出該物件之影像區段(步驟1〇句，其子步 驟包，紀錄此列巾該未知物件之影像區段之起始點並儲存至 暫存器（104a)，接著自該影像區段之起始點逐點累計該影像區 201222424 段^資訊並儲存至暫存器(嶋）；及判斷此列中該未 影像區段終點並儲存至暫在s 物件的 段之蝴_分辨該 1〇5)，其中，在進糾^ 之域區段屬於何物件(步驟 '、 μ二間相關性之辨識時，宜至少平行谁件 決疋下-影像區段起始點之步驟， 進仃 該影像區段所累計之、、、、 、a〗，匯集 ^ 其所屬之物件(步驟106);同理， 進订b列下-衫像區段的判斷(步驟贿）。 I ® 4 5所不’說明本發 多個物件之動細__㈣—晴:相^生進订 下： ^方法具有之步驟及作用原理詳述如 首先使用步驟101〜107來判斷久 先判斷辨，屮的輪姓β 。或工〜特徵屬性’首 =斷辨識出的物件是否圍繞任何背景(步驟 在步驟112中判定該物件特徵屬 ）·右否，則 斷辨識出的物件圍射背景二::::件。若步驟108判 該:件的空心區域並計算該空二:：判=:屬： :?(空，積)除以(實體與空二 传的商是否大於-閾值？若 匕场由積和）所 若大於該閾值，則進行步驟113 驗驟112， 值約為。，。5〜_之間。步驟113 =!=該間 一空心物件.另牛赖^ β π頰垓物件之特徵屬性為 件，如此便完成了物件之;卸類二玄物件之特徵屬性為-實心物 114)。 實心、空心之特徵屬性辨識功能(步驟 201222424 上述實施例再參照圖6進一步說明。首先以灰階閑值 像6予以二分（binarized)。接著，逐·測影像區段（料 根據_步驟104〜106)，而判斷出影像6包含物件61，和62，。 接巧，根據物件61’和62,是否包含背景，在物件61 尚什算背景611’除以整體面積的商值，於是觸出物件6 空心物件而物件62’為實心物件。Seg-R^Preline-Objj-L ; The formula ^ indicates, for example, when reading the data of the Yth column; (4) The unknown object appearing in the third column of the 转 table: coordinate value; Seg.R indicates reading 帛γ _ ; ^ Prdine-〇bji-R^-f YJ,- that is, the image area of each object i appearing in the γ·1 column indicates that each object 2 in the w column appears = ^ the left starting point X coordinate The value, if it conforms to the Seg L line R E-R2Prdme-〇bjrL^ judgment formula, means that the segment and the image segment of the object i appearing in the Y-1 column belong to the same object - "like Q". Illustrated in the present invention, the present invention utilizes the feature dissimilarity of the object to perform dynamic f-images of a plurality of objects. The second preferred implementation method of the present invention determines how the image segment belongs to the object, and has steps and conditions as detailed below. First, the grayscale threshold of the image is set (step 10); then, the pixel value of each column in the shirt image is sequentially processed (step 1Q2); the background region is determined by the gray value (step 103); The image segment of the object (step 1 ,, its sub-step package, record the starting point of the image segment of the unknown object and store it in the scratchpad (104a) Then, the image area 201222424 segment information is accumulated point by point from the starting point of the image segment and stored in the temporary register (嶋); and the end point of the non-image segment in the column is determined and stored to the temporary object. The butterfly of the segment _ distinguishes the 1〇5), wherein, in the domain segment of the correction, what object belongs to (the identification of the correlation between the steps and the μ, at least parallel to the segment - the image segment The step of starting point, the sum of the , , , , and a accumulated in the image segment is collected, and the object to which it belongs is collected (step 106); similarly, the judgment of the b-under-shirt segment is entered ( Step bribe. I ® 4 5 does not 'describe the dynamics of many objects in this hair __ (four) - sunny: the same as the following: ^ The method has a detailed description of the steps and the principle of action, such as using steps 101 to 107 first Judging the long-term judgment, the round name of the round β. or the work-feature attribute 'first=breaking whether the identified object surrounds any background (the step determines the object characteristic in step 112) · right no, then the identified The object is surrounded by a background of two::::. If the step 108 determines: the hollow area of the piece and calculates the empty two:: ==: genus: :? (product) divided by (whether the quotient of the entity and the null second pass is greater than - the threshold value? If the field is the product sum) if the threshold is greater than the threshold value, then step 113 is performed, the value is approximately ., between 5 and _ Step 113 =! = a hollow object in the room. The characteristic attribute of the β π cheek object is the piece, so that the object is completed; the characteristic property of the unloading element is - solid 114). Hollow feature attribute identification function (step 201222424) The above embodiment is further described with reference to FIG. First, binarized with a grayscale idle value like 6. Next, the image segment is measured one by one (based on _steps 104 to 106), and it is judged that the image 6 contains the objects 61, 62, and 62. Coincidentally, depending on whether the objects 61' and 62 contain the background, the object 61 is calculated as the background 611' divided by the quotient of the overall area, so that the object 6 hollow object is touched and the object 62' is a solid object.
配合® 4、7所^朗本發_用物件賴相異 多個物件之動祕像辨齡法的第二較佳實施财，是如何^ 識出長形或短形物件，該方法具有之步驟及作用原理詳述如 ”用步驟ΗΠ〜1G7來判斷各影像區段屬於何物件， =而依據另—觸法_分料物件之長形或短形 =二實:例之判斷法則係先判斷並操取獅^ ;^(步驟12〇);接著計算該物件之長邊、短邊向量(步驟 121);再計算(該物件之長邊 您驟 -閾值？（步驟122)若大卿^ 該物件之面積)是否大於 該間值，_步^4 ^驟⑵則f行步驟123，若不大於 是判斷該物件之特徵屬性為一短形 驟125);較佳地，（該物件 賊屬崎識功此(步 經過實驗後，得到邊2長3^^該物件之面積)之值 是屬於長型或短形物件，像2中的二物件21 ' 22 判斷屬於短形物件，而圖中〗的==件21經辨識後即 J長方开>物件22經辨識後即判斷 201222424 屬於長形物件。 圖9顯示本發明的另一個實施例。在圖1的先前技術中’ 指示道具71、72具有光源711、712及72卜722。根據本發明， 光源711、712、721、722之一（或更多）所投射的光線可帶 有圖案。在其中一種實施型態中，左方的指示道具71所投射 的光線圖案與右方的指示道具71所投射的光線圖案不同。在 另一種實施型態中，指示道具71之光源711和712所投射的 光線圖案彼此不同、指示道具72之光源721和722所投射的 光線圖案彼此不同。在另一種實施型態中，所有光源71卜712、 72卜722所投射的光線圖案彼此皆不相同。圖9所示實施例係 舉例顯示’光源721、722所投射的光線圖案彼此不同。帶有 圖案的光線有助於更精確辨識並區分物件，因為影像處理系統 3可更精確判斷出其係自哪一光源接收光線。有關光線帶有圖 案的優點，在後文中將再詳細說明。 圖10示出光源711、712、721、722的一個實施例（舉例 標示為721)，此光源包含一或多個發光元件725，以及一個折 射光學元件728。折射光學元件728將發光元件725所發出的 光線折射成帶有特定圖案的線性或平面光。有關圖案的細節， 在後文中將再詳細說明。 匕一事實上，光源711、712、72卜722並不絕對必須設置在 ，示道具71與72之上；指示道具71與72僅需能夠反射光線 p可。光源可設置在其他任何地方，向指示道具71盥72浐射 =。本技術業者當可了解，此種變換並不會影響前文所^對 於物件的辨識。在此種情況下，甚至連指示道具71與72也都 可省略’例如’可使用人體的某個部位來取代指示道且盘 201222424 72，只要該人體部位所反射的光線強度足夠辨識即可。 圖11顯示光源安裝於別處的實施例。為辨識並區分影像 中的物件，在本實施例中，光源80投射帶有圖案的光線。該 帶有圖案的光線被投射往指示道具72或使用者的一個人體部 位706，被其反射而為影像處理系統3所接收。影像處理系統 3中的衫像感測器31 (未示於圖11 )接收該反射光。所述光線 圖案例如可包含不同的亮度、顏色、雜、大小、紋理、或密 度等，藉由光源80中發光元件81的排列方式（例如，如圖所 示’多個發光元件81依預定圖案排列）、不同發光時序（亦即， 可藉由分別㈣各個發統件81，使受到級射的多個位置分 別於相同或㈣時點被照射）、不同發絲譜（亦即發光元二 81可發出不_譜的光線，可為可見或不可見光）、或以上兩 者或更多者之組合，來產生。 帶有圖案的光線有助於更精確辨識並區分物件，說 下。請參閱第12〜丨3圖，光線被指示道具72 (或圖 體部位7〇6)反射而為影像處理系統3中的影像感測器31^ 收。因此，根據影像感測器31上光的反射 = f具72與影《_31在z方向上的距離'然而，如 =面，這可料致產生錯誤的輯資訊，另-方面=： 的物件’例如將兩物件誤認為同-物件。 免發生此種誤判，如果路徑P1 為避 像處理系統3便可判斷出其係自^料同，影 圖MA〜HC舉例顯示光線圖案 HA所示’可在圖案中安排多個不同大小的亮區B;或^ 201222424 所示’可在圖案中安排多個不同大小的暗區D ;或如圖14C所 不’圖案中可包括不同顏色、形狀、次序、亮度的區域，等等。 圖15示出本發明的另一個實施例。除了設計發光元件81 的排列方式、發树序、發光頻譜以產生圖案之外，還可以別 的方式產生酵。如圖所示，光源8G中另包含有微鏡面82。 在本實施例中，發光元件81投射線性光至微鏡面82，而微鏡 面82將光線反射至指示道具72或人體部位。微鏡面幻可 化=軸-_動，藉由其轉動，將線性光轉換成為娜光束， 乂掃吗^日示道具72或人體部位706。在本實施例中，不但可藉 由發光το件81的安排來產生圖案，還可藉由控制微鏡面82的 轉動來產生圖案。 圖16示出本發明的另一個實施例。在本實施例中光渴 8〇另包含有折射光學元件83。在光源8G中可以只具有一麵 ^元件81 (但當純可以），此縣元件81投㈣狀光束而 折射光子元件83將其轉換成雜或平面光。轉換後的線壯 或平面光被投射至指*道具72或人體部位·。在本實施伯 ^不但可藉由發光姑81的發光時序來產生圖案（或由潑 二元件81的其他女排方式來產生，如光源⑽具有不只一個發 光元件81)還可藉由折射光學元件83 @設計來產生圖案。如 圖16右側所示，折射光學元件83例如可以將發光元件幻所 技射的點狀光束轉換成線狀或平面光，而成為點陣形狀、字母 形狀、變密度圖案、等等。 圖Π示出本發_另—個實_。在本實施财，光源 8〇 光元件81 (但當然也可更多)，且光源 8〇中另包3有可沿X軸和γ軸二維轉動的微鏡面.微鏡面 201222424 82將光源8G發出的光線反射並轉換成為伽光束，以掃猫指 示道具72或人體雜7〇6。在本實_ t，^但可藉由發光元 件的發光時序來產生醜（或由發光树81的其他安排方 式來產生’如光源80具有不只―個魏元件81)，還可藉由控 制微鏡面82的二維轉動來產生圖案。 圖18〜19示出本發明的另兩個實施例，其中光源8〇除包 含一或多懈光元件81外’另包含微鏡面82和折射光學元件 83之組合。折射絲元件83可設置在發光元件81與微鏡面 82之間、或設置在微鏡面82與指示道具72或人體部位7〇6之 間圖20 *出本發明的另一個實施例，其中微鏡面包含有 多個鏡面單元，可以個別一維（如圖示）或二維（圖未示）轉 動。這些實施例都可以產生帶有_的光線。With the 4,7, and the lang, the second best implementation of the object-like method, how to identify long or short objects, the method has The steps and the principle of action are as follows: "Use step ΗΠ~1G7 to judge what object belongs to each image segment, = and according to the other - touch method _ the shape of the object is elongated or short = two real: the rule of precedence is Judge and listen to the lion ^ ; ^ (step 12 〇); then calculate the long side, short side vector of the object (step 121); then calculate (the long side of the object you threshold - threshold? (step 122) Ruo Qing ^ Whether the area of the object is greater than the value, _step ^4 ^ (2) then f line step 123, if not greater than is to determine the feature attribute of the object is a short step 125); preferably, (the object The thief is a saga, and the value of the object is 2 or 3^^. The value of the object is a long or short object. The two objects 21 like 22 are judged to be short objects. In the figure, the == member 21 is recognized, that is, the J is squared and the object 22 is recognized, and the 201222424 is determined to be an elongated object. Fig. 9 shows another embodiment of the present invention. In the prior art of Fig. 1, the indicator props 71, 72 have light sources 711, 712, and 72. 722. According to the present invention, one (or more) of the light sources 711, 712, 721, 722 can be projected with light. In one embodiment, the light pattern projected by the left indicator 71 is different from the light pattern projected by the right indicator 71. In another embodiment, the light source 711 of the indicator 71 is indicated. The light patterns projected by 712 and 712 are different from each other, and the light patterns projected by the light sources 721 and 722 of the props 72 are different from each other. In another embodiment, all the light patterns 71, 72, 72, 722 are projected with each other. The embodiment shown in Fig. 9 shows that the light patterns projected by the light sources 721, 722 are different from each other. The patterned light helps to more accurately identify and distinguish objects, because the image processing system 3 can more accurately determine Which light source receives light from which the light has a pattern, which will be described in more detail later. Figure 10 shows an embodiment of the light sources 711, 712, 721, 722 (illustrated by way of example) 721), the light source comprises one or more light-emitting elements 725, and a refractive optical element 728. The refractive optical element 728 refracts light emitted by the light-emitting element 725 into linear or planar light with a specific pattern. In the following, it will be described in more detail. In fact, the light sources 711, 712, 72 722 are not absolutely necessary to be placed above the items 71 and 72; the indicators 71 and 72 need only be able to reflect the light p. Can be set anywhere else, to the indicator props 71 盥 72 = = =. The skilled artisan can understand that this transformation does not affect the identification of the object. In this case, even the indicator items 71 and 72 can be omitted. For example, a certain part of the human body can be used instead of the indicator track and the disk 201222424 72 as long as the light intensity reflected by the body part is sufficiently recognized. Figure 11 shows an embodiment in which the light source is mounted elsewhere. To identify and distinguish objects in the image, in this embodiment, light source 80 projects the patterned light. The patterned light is projected onto the indicator item 72 or a human body portion 706 of the user and is reflected by the image processing system 3 for reflection. The shirt image sensor 31 (not shown in Fig. 11) in the image processing system 3 receives the reflected light. The light pattern may include, for example, different brightness, color, impurity, size, texture, or density, etc., by the arrangement of the light-emitting elements 81 in the light source 80 (eg, as shown) 'the plurality of light-emitting elements 81 are in a predetermined pattern. Arrangement, different illumination timings (that is, by respectively (4) each of the hairs 81, so that a plurality of positions subjected to the stage emission are respectively irradiated at the same or (four) time points), different hair spectrum (ie, the light element II 81 It can be generated by emitting a non-spectral light, which can be visible or invisible, or a combination of two or more of them. Patterned light helps to more accurately identify and distinguish objects, say below. Referring to Figures 12 to 3, the light is reflected by the indicator item 72 (or the picture portion 7〇6) and is received by the image sensor 31 in the image processing system 3. Therefore, according to the reflection of the light on the image sensor 31 = f with 72 and the distance "_31 in the z direction", however, such as = face, which can be expected to produce erroneous information, another aspect =: object 'For example, mistake two objects for the same object. In order to avoid such a misjudgment, if the path P1 is the image avoidance processing system 3, it can be judged that the system is the same, and the images MA to HC are displayed by the light pattern HA as shown in the figure, and a plurality of different sizes of light can be arranged in the pattern. The area B; or ^ 201222424 shows that a plurality of differently sized dark areas D may be arranged in the pattern; or a pattern of different colors, shapes, orders, brightnesses, etc. may be included in the pattern as shown in FIG. 14C. Figure 15 shows another embodiment of the present invention. In addition to designing the arrangement of the light-emitting elements 81, the tree-like order, and the light-emitting spectrum to produce a pattern, the leaven may be produced in another manner. As shown, the light source 8G additionally includes a micro mirror surface 82. In the present embodiment, the light-emitting element 81 projects linear light to the micro-mirror 82, and the micro-mirror 82 reflects the light to the indicator prop 72 or the body part. The micro-mirror illusion = axis--movement, by which the linear light is converted into a nano-beam, and the smear is displayed. In the present embodiment, not only the pattern can be produced by the arrangement of the light-emitting elements 81, but also by controlling the rotation of the micro-mirror 82. Figure 16 shows another embodiment of the present invention. In the present embodiment, the light-diffusing optical element 83 is further included. In the light source 8G, there may be only one side element 81 (but purely pure), and the county element 81 casts a (four) beam of light and refracts the photon element 83 to convert it into a hetero or planar light. The converted line or flat light is projected onto the finger *prop 72 or body part. In the present embodiment, not only can the pattern be generated by the light-emitting timing of the light-emitting prism 81 (or by other female rows of the second element 81, such as the light source (10) having more than one light-emitting element 81) but also by the refractive optical element 83. @Design to create patterns. As shown on the right side of Fig. 16, the refracting optical element 83 can, for example, convert a point beam of a illuminating element into a linear or planar light into a lattice shape, a letter shape, a variable density pattern, and the like. The figure shows the hair _ another - a real _. In this implementation, the light source 8 is light-emitting element 81 (but of course more), and the light source 8A has a micro-mirror surface that can be rotated two-dimensionally along the X-axis and the γ-axis. Micro-mirror 201222424 82 Light source 8G The emitted light is reflected and converted into a gamma beam, which sweeps the cat to indicate the prop 72 or the human body 7〇6. In this case, the ugly can be generated by the illumination timing of the illuminating element (or by other arrangements of the illuminating tree 81), such as the light source 80 having more than one Wei element 81, and by controlling the micro The two-dimensional rotation of the mirror 82 produces a pattern. Figures 18-19 illustrate two other embodiments of the present invention in which the light source 8 includes a combination of a micromirror 82 and a refractive optical element 83 in addition to one or more of the light-scratching elements 81. The refractive wire element 83 can be disposed between the light-emitting element 81 and the micro-mirror surface 82, or between the micro-mirror surface 82 and the indicator prop 72 or the human body portion 7〇6. FIG. 20 shows another embodiment of the present invention, wherein the micro-mirror surface It contains multiple mirror elements, which can be rotated in one dimension (as shown) or two dimensions (not shown). These embodiments can produce light with _.
除了投射帶有圖案的光線外，請參閱圖2卜可設計使影像 處理系統3 整其曝光參數，以更精確地辨識及區分物 件。在步驟91中，影像處理系統3根據-組曝光參數感測影 像中的像素。步驟92 +，影像處理系統3判斷是否有許多像 f值落在範H外（太亮或太暗）；所謂「許多」可妓設計者 ^使用者§忍為合適的數值，例如>，>75%，〉齡，等等。 :判斷、纟Q果為疋’騎行步驟93，調整曝光參數。若判斷結果 ’則摊處理系統3處理影像以辨識並區分物件（步驟 I並賴轉前㈣光參數來制次一影像 。藉由調整曝 . 首先’可過濾掉超過高閾值（太亮）與低於低閾值（太 二效^雜I其次，如光線圖案具有不同亮度的區域，則透過 如像處理系統3可以更清晰地辨識圖案，以更 精確地辨識並區分物彳彳：。 、 201222424 本發明已就較佳貫施例敘述如上，但以上所述者，僅為本 發明之較佳實補而已，當不能以此限定本發施之範圍， 即大凡依本發明巾請專職發日魏㈣容所作之簡 等效變化與修飾，皆仍屬本發明專利涵蓋之範 【圖式簡單說明】 圖1是-示意圖，說明-種習知的互動式遊戲裝置。 圖2是-電路方塊圖，說明使用本發明之動態影像辨識方 法的影像辨識系統，用以將辨識後之特徵屬性相關資訊輸出至 一現有的個人電腦主機具有的傳輸介面。 圖3是-示意圖’說明本發明利用物件特徵相異 ::件：影像辨識方法的第一較佳實施例，可辨識： 中的二物件是屬於實心或空心物件。 扣以像 圖41流賴，說明使用本發明利用物 行多=件之動態影像辨識方法之二實施例，係如何在 初始利用各影像區段之識別來觸其所屬物件。5、 中二 圖6舉例說明如何觸實心或空心物件。 此夕f ^ ^ ® ’本發明物件賴相舰進行多個物 的二㈣==:::佳實施例，如何__ 的二= ===較佳實施例可辨識出影像中 圖9示出本發明的另-個實施例。 圖10不出光源的-個實施例，此光源包含一或多個發光 201222424 元件，以及一個折射光學元件。 圖η示出本發明的另一個實施例，其中光源安裝於別處。 圖12與13說明為何會造成誤判。 圖14A〜14C舉例顯示數種光線圖案。 圖15〜20示出本發明的另外數個實施例。 圖21舉例示出調整曝光參數的步驟。 【主要元件符號說明】 1、2影像 1卜12、21、22 物件影像 111、121 影像區段 111’、121’影像區段之起始點 111”、121” 影像區段之終點 91 〜94、101 〜107、110〜114、120〜125 步驟 3 影像處理系統 31影像感測器 311像素 32類比數位轉換器 33影像處理單元 34暫存器 35 介面模組 4個人電腦 41 主機 411傳輸介面 42顯示器 700遊戲裝置 201222424 706 人體部位 71、72指示道具 7U、712、721、722 光源 · 720 踏步台 730螢幕裝置 750視訊攝影機 760 遊戲盒 770 主機裝置 80光源 Φ 81 發光元件 82 微鏡面 83折射光學元件In addition to projecting the patterned light, please refer to Figure 2 to design the image processing system 3 to adjust its exposure parameters to more accurately identify and distinguish objects. In step 91, image processing system 3 senses the pixels in the image based on the set of exposure parameters. Step 92 +, the image processing system 3 determines whether there are many f-like values falling outside the norm H (too bright or too dark); the so-called "many" can be used by the designer ^ user to endure appropriate values, such as >, >; 75%, 〉 age, and so on. : Judgment, 纟Q fruit is 疋 'riding step 93, adjust the exposure parameters. If the result is judged, the processing system 3 processes the image to identify and distinguish the object (step I and the previous (four) light parameters to make the next image. By adjusting the exposure. Firstly, the filter can be filtered out of the high threshold (too bright) and low. At a low threshold (Taiwan effect), if the light pattern has regions of different brightness, the pattern can be more clearly recognized through the image processing system 3 to more accurately identify and distinguish objects: 20122424 The invention has been described above in terms of preferred embodiments, but the above description is only a preferred embodiment of the present invention, and the scope of the present invention cannot be limited thereto. (4) The equivalent changes and modifications of the contents are still covered by the patent of the present invention. [Simplified illustration of the drawings] Fig. 1 is a schematic diagram showing a conventional interactive game device. Fig. 2 is a circuit block diagram. The image recognition system using the dynamic image recognition method of the present invention is used to output the identified feature attribute related information to a transmission interface of an existing personal computer host. FIG. 3 is a schematic diagram. The invention utilizes the features of the object to be different:: a first preferred embodiment of the image recognition method, which can be identified: the two objects in the middle are solid or hollow objects. The buckle is shown in Figure 41, which illustrates the use of the present invention. The second embodiment of the moving image identification method is how to initially touch the object by the identification of each image segment. 5, Figure 2 illustrates how to touch a solid or hollow object. ^ ^ ® 'The object of the invention is two (four) of the plurality of objects ==::: good embodiment, how to __ two ==== preferred embodiment can identify the image Figure 9 shows the invention Another embodiment of Figure 10. Figure 10 shows an embodiment of a light source comprising one or more illumination 201222424 elements, and a refractive optical element. Figure n shows another embodiment of the invention in which the light source is mounted Figures 12 and 13 illustrate why misjudgment can occur. Figures 14A-14C illustrate several light patterns. Figures 15 through 20 illustrate additional embodiments of the present invention. Figure 21 illustrates the steps of adjusting exposure parameters. [Main component symbol description 1, 2 images 1 Bu 12, 21, 22 Object images 111, 121 Image segments 111', 121' The starting point of the image segment 111", 121" The end of the image segment 91 ~ 94, 101 ~ 107, 110 ～114, 120~125 Step 3 Image Processing System 31 Image Sensor 311 Pixel 32 Analog Digit Converter 33 Image Processing Unit 34 Scratchpad 35 Interface Module 4 Personal Computer 41 Host 411 Transmission Interface 42 Display 700 Game Device 201222424 706 Human body parts 71, 72 indicate props 7U, 712, 721, 722 Light source · 720 Step 730 Screen device 750 Video camera 760 Game box 770 Host device 80 Light source Φ 81 Light-emitting element 82 Micro-mirror 83 refracting optical element
S 18S 18