200926066 九、發明說明： 【發明所屬之技術領域】 本發明係提供一種視訊遊戲之顏色差異量影像辨識 方法，尤指一種可依據色彩與色彩平均值的差異量來辨識 影像的方法。 【先前技術】 隨著數位技術的進步及電腦視訊相關技術的發展，影 ❹ 像辨識已廣泛應用在視訊遊戲、互動裝置藝術、人群數量 的計算監控、門禁防盜系統及看護監控系統等方面，藉以 偵測與分析物體的活動以及異常事件的狀態。 在影像辨識技術的開發上，習用雖有多種辨識方法被 設計’然而通常需要設定欲辨識物體的顏色區域範圍，此 外’所需下達的程式指令多，運算效能難以提升。 Ο 【發明内容】 本發明人有鑑於此，乃悉心從事設計，於今有本發明 ^生，此，本發明之主要目的在提供色差異= 辨識處理上，藉以提供偵戲的影像 複雜情形，同時，該影像辨C體進入及偵測影像的 能迅速㈣像賴應用上用在需要簡易而效 算人群規模等料；再者’防盗、互動裝置藝術、計 景的顏色有不同時’就可::要=辨:物體與背 仃列斷因而，在影像辨識 5 200926066 的應用上至少可為大眾或工業提供一種有用的選擇。 依據本發明一實施例構成的視訊遊戲之顏色差異量 影像辨識方法，係包含以下步驟： 擷取影像； 將目前影格的影像偵測區域轉為灰階化的圖像； 計算目前影格的灰階平均值； 依據影像灰階值變動量的大小來判斷影像是否有變 動，若判斷成立，代表影像有變動，若判斷不成立，代表 〇 影像無變動。 在一實施例中，該影像灰階值變動量係包含顏色差設 定值與顏色差面積設定值，且依據影像灰階值變動量的大 ' 小來判斷影像是否有變動的步驟係包含計算目前影格有 多少個像素點與灰階平均值的絕對值超過顏色差設定 值，以及判斷超過該顏色差設定值的像素點數量有沒有大 於該顏色差面積設定值。 在一實施例中，顏色差設定值為灰階平均值的10%， ° 色差面積設定值為系統所要偵測影像是否有變動的影像 區域範圍的50%。 關於本發明之其他目的、優點及特徵，將可由以下較 佳實施例的詳細說明並參照所附圖式來了解。 【實施方式】 有關本發明的構造設計，將經由僅為例子但非用以限 制的實施例並參照所附圖式作進一步說明： 200926066 «月f閱第一圖，顯不依據本發明構成的視訊遊戲之顏 色差異量影像辨識方法包含四個主要步驟，即擷取影像 (步驟101)、灰階化（步驟1〇2)、計算灰階平均值（步 驟103)、依據影像灰階值變動量的大小來判斷（步驟 104)·’若判斷成立，則代表影像有變動（步驟1〇5)，若 判斷不成立，則代表影像無變動（步驟106)。 在步驟101之中，係利用攝影機來满取影像，並且傳 〇 送到一影像辨識系統内做影像處理及執行後續的運作；在 本實施例中，影像解析度為640 X 480，影像格式為 RGB24 ’共有三種顏色’分別為R (紅色）、G (綠色）、 ' (藍色），母一種顏色有8位元的大小，每一種顏色變化 可以有0到255的範圍變化。 在步驟102之中，係當系統接收到影像的訊號後，將 原始的彩色影像轉換成灰階影像，公式為Gray=(R+G+B) G，Gray為系統所需要的灰階值，R、G、B為原始影像 ❾ 訊號。 在步騍103之中，係當原始影像轉換成灰階影像後， 將所需要判斷的影像區域進行灰階平均值的計算動作公 式為Avg= / κ，Avg是灰階平均值，y是灰階影 像中的像素編號（影像解析度64〇χ48〇的影像有3〇72〇〇 個像素），£是系統所要偵測影像是否有變動的影像區域範 圍。 在步驟104中進行判斷時，需要有預設條件，在本實 200926066 施例中，預設條件有兩個，一為顏色差設定值，以Color A 為表示，代表的意義為灰階平均值的10%(計算公式：Color A = Avg. x 10%)，二為顏色差面積設定值，以Color B 表示，代表的意義為系統所要偵測影像是否有變動的影像 區域範圍的50% (計算公式：Color B = K X 50%)， 該二設定值為判斷偵測影像的區域是否有影像變動的依 據。 在步驟104中係依據下列計算公式判斷影像有無變 〇 動：Count = 0200926066 IX. INSTRUCTIONS: [Technical Field] The present invention provides a color difference image recognition method for a video game, and more particularly to a method for recognizing an image based on the difference between the color and the color average. [Prior Art] With the advancement of digital technology and the development of computer-visual related technologies, image recognition has been widely used in video games, interactive device art, population monitoring, access control and surveillance systems, and so on. Detect and analyze the activity of objects and the state of anomalous events. In the development of image recognition technology, although a variety of identification methods have been adopted for designing, however, it is usually necessary to set a range of color regions for which an object is to be recognized, and in addition, there are many program instructions to be issued, and it is difficult to improve the calculation performance. Ο 发明 发明 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本 本The image discriminates between the C body and the image detection. (4) The image is used in a simple and effective population size; in addition, the 'anti-theft, interactive device art, and the color of the landscape are different' ::To = Identify: Objects and backs are broken, thus providing at least a useful choice for the general public or industry in the application of Image Recognition 5 200926066. The method for recognizing a color difference image of a video game according to an embodiment of the present invention comprises the steps of: capturing an image; converting an image detection area of the current frame into a grayscale image; calculating a gray scale of the current frame The average value is determined according to the magnitude of the fluctuation of the grayscale value of the image. If the judgment is true, the representative image is changed. If the judgment is not true, the image is unchanged. In an embodiment, the grayscale value variation of the image includes a color difference setting value and a color difference area setting value, and the step of determining whether the image has a change according to a large 'small amount of the image grayscale value variation amount includes calculating the current The number of pixels and the absolute value of the grayscale average exceeds the color difference setting value, and whether the number of pixels exceeding the color difference setting value is greater than the color difference area setting value. In one embodiment, the color difference setting is 10% of the grayscale average, and the color difference area setting is 50% of the range of the image area in which the image is to be detected by the system. Other objects, advantages and features of the present invention will become apparent from the Detailed Description of the <RTIgt; [Embodiment] The structural design of the present invention will be further described by way of example only and not limitation, and reference to the accompanying drawings: 200926066 «The first picture of the month f, which is not constructed according to the present invention The color difference image recognition method of the video game includes four main steps, namely, capturing the image (step 101), graying (step 1〇2), calculating the grayscale average value (step 103), and changing according to the grayscale value of the image. The size of the quantity is judged (step 104). If the determination is satisfied, the representative image is changed (step 1〇5). If the determination is not satisfied, the representative image is not changed (step 106). In step 101, the camera is used to capture the image, and the image is sent to an image recognition system for image processing and subsequent operations; in this embodiment, the image resolution is 640 X 480, and the image format is RGB24 'there are three colors' are R (red), G (green), and '(blue), and the parent color has an 8-bit size. Each color change can vary from 0 to 255. In step 102, when the system receives the signal of the image, the original color image is converted into a grayscale image, and the formula is Gray=(R+G+B)G, and Gray is the grayscale value required by the system. R, G, and B are the original image signals. In step 103, after the original image is converted into a grayscale image, the calculation of the grayscale average value of the image region to be determined is Avg=/κ, Avg is the grayscale average, and y is gray. The pixel number in the image (the image resolution is 64〇χ48〇 has 3〇72〇〇 pixels), which is the range of the image area that the system wants to detect whether the image has changed. In the determination in step 104, a preset condition is required. In the actual example of 200926066, there are two preset conditions, one is the color difference setting value, represented by Color A, and the representative meaning is the grayscale average value. 10% (calculation formula: Color A = Avg. x 10%), and the second is the color difference area setting value, expressed by Color B, which means the system is 50% of the image area range in which the image is to be detected. The calculation formula: Color B = KX 50%), the two set values are the basis for judging whether there is image change in the area of the detected image. In step 104, the image is judged according to the following calculation formula: Count = 0
For ^ = 1 to κFor ^ = 1 to κ
If ( I Grayj - Avg.| > Color A) ' {If ( I Grayj - Avg.| > Color A) ' {
Count = Count +1; }Count = Count +1; }
If ( Count > Color B ) { ® 表示影像有變動； }If ( Count > Color B ) { ® indicates that the image has changed; }
ELSE { 表示影像無變動； } 其中，Count代表系統偵測區域範圍中像素值與灰階 平均值離異程度大於Color A之像素個數，Count初值設 200926066 定為〇;若該計算公式的判斷成立時，系統即判斷為影像 有變動’否則判斷為影像沒有變動，判斷完成後系統便可 以給於使用者相對應的回饋。 以下以該顏色差異量影像辨識法應用在視訊遊戲的 一實施例來對本發明作進一步說明： 請參閱第二圖，顯示在該視訊遊戲中，玩家的手21 尚未觸碰該視訊遊戲之按姐22前的影像；假設有以下3*3 矩陣的像素點（表格内的值為色光三原色，依序為紅、綠、 〇 藍，值為0〜255): 灰階化後，出來結果如下： 200 205 195 205 200 195 195 205 200 接著，依上述步驟103計算目前影格的灰階平均值， 將全部像素的灰階值加起來平均，全部影像的灰階平均值 為（200+205+195+205+200+195+195+205+200) / 9=200 (200, 195, 205) (200, 210, 205) (195, 200, 190) (200, 210, 205) (195, 200, 205) (190, 195, 200) (200, 195, 190) (210, 205, 200) (205, 195, 200) 首先’依上述步驟102將目前影格的影像偵測區域轉 為灰階化的圖像，轉為灰階化時，公式為：灰階值=(紅 色+綠色+藍色）/ 3ELSE { means that the image has no change; } where Count represents the number of pixels in the range of the system detection area and the gray level average is greater than the number of pixels in Color A, and the initial value of Count is set to 200926066 as 〇; if the calculation formula is judged When it is established, the system judges that the image has changed. Otherwise, it is judged that the image has not changed. After the judgment is completed, the system can give the corresponding feedback to the user. The present invention is further described below with reference to an embodiment in which the color difference image recognition method is applied to a video game: Referring to the second figure, the player's hand 21 has not touched the video game player in the video game. Image before 22; assume that there are pixels of the following 3*3 matrix (the values in the table are the three primary colors of the color, in the order of red, green, and indigo, the value is 0~255): After graying, the result is as follows: 200 205 195 205 200 195 195 205 200 Next, according to the above step 103, calculate the grayscale average value of the current frame, add the grayscale values of all the pixels to average, and the grayscale average of all the images is (200+205+195+) 205+200+195+195+205+200) / 9=200 (200, 195, 205) (200, 210, 205) (195, 200, 190) (200, 210, 205) (195, 200, 205) (190, 195, 200) (200, 195, 190) (210, 205, 200) (205, 195, 200) First, follow the above step 102 to convert the image detection area of the current frame into a grayscale image. For example, when changing to grayscale, the formula is: grayscale value = (red + green + blue) / 3
接著’依上述步驟104進行辨識；該辨識的依據係為 系統預先對顏色差的設定值和顏色差面積的設定值（設計 200926066 者可依需求及應用環境自行設定顏色差設定值，顏色差設 定值若越小，感應越靈敏，越大則感應越不靈敏，而顏色 差面積設定值是用來判斷欲偵測的物件進入辨識區域要 多大的面積才會啓動事件），在本實施例中，顏色差設定 值設定為-20〜20，顏色差面積設定值設定為4 ;當將灰階 化後的每個像素值與全部影像的灰階平均值相減，並計算 超出顏色差設定值範圍的像素點有多少個，若超過顏色差 面積設定值，則判斷為有物體進入影像中，公式為：每個 〇 像素值-全部影像灰階平均值。結果如下： 200-200=0(未超出顏色差設定值-20〜20) 205_200=5(未超出顏色差設定值_20〜20) 195-200=-5(未超出顏色差設定值-20〜20) 205-200=5(未超出顏色差設定值_20〜20) 200-200=0(未超出顏色差設定值_20〜20) 195-200=-5(未超出顏色差設定值-20〜20) 195-200=-5(未超出顏色差設定值-20〜20) ® 205-200=5(未超出顏色差設定值-20〜20) 200-200=0(未超出顏色差設定值_20〜20) 超出顏色差設定值範圍的像素點數量為0，未超過顏 色差面積設定值4,因此判斷無物體進入該影像中。 再請參閱第三圖，顯示玩家的手21進入該影像畫 面，觸碰該視訊遊戲之按鈕22後的狀態，此時，該影像 畫面的像素值會改變為如下：__ (200, 195, 205) (200, 210, 205) (195, 200, 190) 200926066 (200, 210, 205) (180, 120, 30) (180, 120, 30) (200, 195, 190) (180, 120, 30) (175, 115, 25) 依上述步驟102轉為灰階化後，像素值如下: 200 205 195 205 110 110 195 110 105 接著，依上述步驟103計算目前影格的灰階平均值， 將全部像素的灰階值加起來平均，全部影像的灰階平均值 ® 為（200+205+195+205+110+110+195+110+105) / 9=159(四 捨五入到整數） 接著，依上述步驟104進行辨識，將每個像素的灰階 值與全部影像的灰階平均值相減，並計算超出顏色差設定 值範圍的像素點有多少個。結果如下： 200-159=41(超出顏色差設定值-20〜20) 205-159=46(超出顏色差設定值-20〜20) 195-159=36(超出顏色差設定值-20〜20) ❹ 205-159=46(超出顏色差設定值-20〜20) 110-159=-49(超出顏色差設定值-20〜20) 110-159=-49(超出顏色差設定值-20〜20) 195-159=36(超出顏色差設定值-20〜20) 110-159=-49(超出顏色差設定值-20〜20) 105-159=-54(超出顏色差設定值-20〜20) 超出顏色差設定值範圍的像素點數量為9,超過顏色 差面積設定值4，因此判斷有物體進入該影像中，按紐22 11 200926066 會變亮。 依據本發明構成的視訊遊戲之顏色差異量影像辨識 法係依據色彩與色彩平均值的差異量來辨識，在實施上， 只要欲辨識物體顏色與背景顏色差異大時，就可以進行辨 識，不僅在任何背景顏色下都能有效辨識，所需下達的程 式指令不多，因此運算上的效能也很快，在影像辨識的應 用上提供一實用的選擇，故本發明已能達成預期之設計目 的及實施效果。 〇 前述是對本發明之構造作較佳實施例的說明，而依本 發明的設計精神是可作多種變化或修改的實施例。是以對 於熟悉此項技藝人士可作之明顯替換與修飾，仍將併入於 本發明所主張的專利範圍之内。 ❹ 12 200926066 【圖式簡單說明】 第一圖：係本發明視訊遊戲之顏色差異量影像辨識方 法之流程圖。 第二圖：係本發明應用在視訊遊戲之一實施例的示意 圖，圖中顯示玩家的手未進入該影像畫面觸 碰按鈕前的狀態。 第三圖：係與第二圖相同的示意圖，圖中顯示玩家的 手進入該影像晝面觸碰按鈕時的狀態。 〇 【主要元件符號說明】 (101) 擷取影像 (102) 灰階化 (103) 計算灰階平均值 (104) 依據影像灰階值變動量的大小來判斷 (105) 影像有變動 (106) 影像無變動 ° ⑼手 (22)按紐 13Then, the identification is performed according to the above step 104; the basis of the identification is that the system sets the set value of the color difference and the color difference area in advance (design 200926066 can set the color difference setting value according to the demand and the application environment, and the color difference setting) If the value is smaller, the sensitivity is more sensitive, the larger the sensitivity is, the more sensitive the area is, and the color difference area setting value is used to determine the area of the object to be detected entering the identification area to start the event), in this embodiment The color difference setting value is set to -20~20, and the color difference area setting value is set to 4; when each grayscaled pixel value is subtracted from the grayscale average value of all images, and the excess color difference setting value is calculated How many pixels are in the range? If the value of the color difference area is exceeded, it is determined that an object enters the image. The formula is: each pixel value - the average gray level of all images. The result is as follows: 200-200=0 (the color difference setting value is not exceeded -20~20) 205_200=5 (the color difference setting value _20~20 is not exceeded) 195-200=-5 (the color difference setting value is not exceeded -20 ~20) 205-200=5 (the color difference setting value _20~20 is not exceeded) 200-200=0 (the color difference setting value _20~20 is not exceeded) 195-200=-5 (the color difference setting value is not exceeded) -20~20) 195-200=-5 (not exceeding the color difference setting value -20~20) ® 205-200=5 (not exceeding the color difference setting value -20~20) 200-200=0 (no color is exceeded Difference setting value _20~20) The number of pixels exceeding the range of the color difference setting value is 0, and the color difference area setting value 4 is not exceeded, so that no object is judged to enter the image. Referring to the third figure, the player's hand 21 enters the image screen and touches the button 22 of the video game. At this time, the pixel value of the image screen changes to the following: __ (200, 195, 205 ) (200, 210, 205) (195, 200, 190) 200926066 (200, 210, 205) (180, 120, 30) (180, 120, 30) (200, 195, 190) (180, 120, 30) (175, 115, 25) After the graying is performed according to the above step 102, the pixel values are as follows: 200 205 195 205 110 110 195 110 105 Next, the gray level average of the current frame is calculated according to the above step 103, and all the pixels are The grayscale values add up to the average, and the grayscale average of all images is (200+205+195+205+110+110+195+110+105) / 9=159 (rounded to an integer). 104 performs identification, subtracting the grayscale value of each pixel from the grayscale average of all images, and calculating how many pixels exceed the range of the color difference setting value. The result is as follows: 200-159=41 (out of color difference setting value -20~20) 205-159=46 (out of color difference setting value -20~20) 195-159=36 (exceeding color difference setting value -20~20 ) 205 205-159=46 (Exceeded color difference setting value -20~20) 110-159=-49 (Exceeded color difference setting value -20~20) 110-159=-49 (Exceeded color difference setting value -20~ 20) 195-159=36 (out of the color difference setting value -20~20) 110-159=-49 (out of the color difference setting value -20~20) 105-159=-54 (exceeding the color difference setting value -20~ 20) The number of pixels beyond the range of the color difference setting value is 9, exceeding the color difference area setting value of 4, so it is determined that an object enters the image, and the button 22 11 200926066 will be illuminated. The color difference image recognition method of the video game constructed according to the present invention is identified according to the difference between the color and the color average value. In the implementation, as long as the difference between the color of the object and the background color is large, the identification can be performed, not only in It can be effectively identified under any background color. There are not many program instructions to be issued, so the computational efficiency is also very fast. It provides a practical choice in the application of image recognition, so the present invention can achieve the intended design purpose and Implementation Effect. The foregoing is a description of the preferred embodiments of the present invention, and the embodiments of the invention may be modified or modified. It is obvious that those skilled in the art can make substitutions and modifications that are still within the scope of the claimed invention. ❹ 12 200926066 [Simple description of the diagram] The first figure is a flow chart of the image recognition method for the color difference amount of the video game of the present invention. Fig. 2 is a schematic view showing an embodiment of the present invention applied to a video game, showing a state in which the player's hand does not enter the state before the image screen touches the button. The third figure is the same schematic diagram as the second figure, showing the state of the player's hand when entering the image and touching the button. 〇【Main component symbol description】 (101) Capture image (102) Grayscale (103) Calculate the grayscale average value (104) Determine according to the magnitude of the image grayscale value variation (105) Image changes (106) No change in image ° (9) hand (22) button 13