1274845 九、發明說明: 【發明所屬之技術領域】 本發明係關於一種使用單一鏡頭偵測物體邊角及距離之裝 置與方法,特別是指用於機器主體的視覺感測褽置上,只需採 .以單一電荷耦合器件(ccd)鏡頭進行對外界環境(如目標物或障 •礙物)的辨識,並藉影像分析系統之方法設計得知目標物的距 離、方向、方位等資訊,以利控制機器主體(如機器人)對外界 •環境之作動為本發明之首要創新設計,深具產業利用價值。 【先前技術】 按,目前各產業自動化程度愈來愈高,利用機器主體連設 以視覺系統,即機器視覺(MachineVision),來做輔助定位、尺 寸量測、瑕疵檢測的應用愈來愈普遍,尤其電子業使用程度最 為普遍,如黏晶機、銲線機、腳位檢查機、印字品質檢測機、 印刷電路板(PCB)檢查機等,都是很典型的應用。 • 上述之機器視覺(MachineVisi〇n),是以機器模仿人類視覺 的光學識別系統,為感測系統中非常獨特且極具潛力的一種感 /則技術,它利用攝影機和電腦,來擷取,分析及解釋影像内容, 進而對機器主體下達某些決策;關於機器視覺(MachineVisi〇n) 的研究己有三十多年的歷史,雖然仍未具有人類視覺的威力和 智慧,但是對於可控制環境之下的明確工作任務,機器視覺卻 能以100%正確的效率及一天24小時的持續工作,來完成人類所 不能完成的任務。 1274845 然而,由於傳統工業界所使用之機器視覺的工具,皆是利 用多顆電荷Μ合器件(CCD)鏡頭各以不同角度之設立,來個別操 取欲處理外界環境之目標物雜㈣杨制f彡像,使著電腦 必須接受多個不同的視覺訊號,再將多角度的影像同步進行處 方月(3什才出目軚物或障礙物的距離,故傳統之機器視覺所 運㈣硬體設傷要求較大,以致資源成本高,再加上其必須利 用多張圖片中之目標物或障礙物的位置重疊區計算出正確位 置故其處理運异及分析的過程中,往往造成資料量相當龐大, ^ 哥找母張圖片之間與物體相關性的問題,使得電腦 及機器主體的_速度過於緩慢,以致影響機器主體(如機器人) 射卜界進打作之達成效率,無法有效提昇產業利機器視覺 之貫質效益,實為傳統機器視覺之缺撼。 基於此,本案發明人為了解決前述習知使用多個電荷耗合 器件⑽)鏡頭,必需利用圖片中目標物之位置重疊區計算位 置而k成所而硬體设備成本高昂、資料分析需求量爆炸及益 法提昇使用機器視覺之工作效率的種種不便,故而亟思且成功 «出^種只需利用單—電荷叙合器件(⑽鏡頭,以及配合電 ^ X套可70整之影像分析I统,即可完成快速檢測出目 標物(或障礙物)的狀況、目標物(或障礙物)所在環境、目標物 (或障礙物)方位角、目標物(或障礙物)方向角以及目標物(或障 礙物)中所包含的角落位置,以利精準操控機器主體執行工作, 1274845 俾達提高該產業之利用價值。 【發明内容】 本!X明之目的即在於提供_種使用單一鏡頭谓測物體邊角 及距離之裝置與方法,其係用於機器主體(如機器人)的視覺感 測裝置上’只需單—電軸合器件(⑽鏡頭即可進行對外界環 t的辨識’並可得知目標物(或障礙物)的距離、方向、方位等 為訊’再利用主控雷腦肉奢6 建70正之影像分析系統,得以有效解 決傳統機器視覺應用於目標物(或障礙物)測距時,必須利用多 此付知目&物(或障礙物)距離的問題,且利用需單一電荷麵合 器件(⑽鏡頭進行辨識,即可找尋目標物(或障礙物)特徵及目 標物(或障礙物)特徵之其他資訊,其設備要求可較傳統多電荷 輕合器件(CCD)鏡頭較低,可節省多餘的設備經費,降低資源成 本’且其需分析的資料量較小,&可增加辨識速度。 本發明之另一主要目的即在於提供一種使用單一鏡頭偵測 物體邊角及距離之方法,主要湘單—電荷麵合器件(⑽)鏡頭 在擷取影像之後,由電腦内建影像分析系統採以兩條割線互相 追及比較其之間斜率的角落偵測演算法,其可正確的運算出 每個物體邊界可能的角落,因此,可應用於需藉單一電荷耦合 器件(CCD)鏡頭找尋出邊界後之物體的特徵辨識上,如物件的原 點定位及可能物件搜尋等運用。 1274845 為達本發明上揭目的所設計之使用單—鏡頭偵測物體邊角 離之衣置與方法,其I置係包含:單—電荷耗合器件(⑽) 鏡頭’負責擷取機器主體之外界環境,以無線通訊將影像回傳 至主控電腦;主控電腦,内建有影像分析系統,在以無線接收 單一電純合ϋ件(⑽鏡頭之影像後,便立即分析並計算目標 物之=訊’將欲控制機器主體之命令,亦以無線通訊系統傳送 至機器主體;機器主體(如機器人(RGbGt)),内建無線接收模 ^上為接收主控電腦所傳達之命命,以控制機器主體執行所 需之工作’以達到控制迴避物體或對外界作工之運動;其方法 係由單-電荷麵合器件(CCD)鏡頭直接萃取出具體之物體輪廊 線,藉電腦内建之影像分析系統採以兩條割線互相追隨及比較 其之間斜率的角落_演算法,來進行分析影像中的輪廓資 訊,即利用兩條割線於輪靡線段上找出影像中的所有符合角落 特徵的角落,且同時計算出目標物之正確數據,能正確獲取目 標物之辨識與鏡頭之距離、目標物與鏡頭之相對方位及障礙物 之方向角。 據此,藉由上述,本發明之設計,只需利用單一電荷耦合 器件(CCD)鏡頭,以及利用連結主控電腦内建以一套可完整之影 像分析系統,即可完成快速檢測、辨識出目標物(或障H = 狀況、目標物⑷章礙物)所在環境、目標物(或障礙物)方位角、 目標物(或障礙物)方向角以及目標物(或障礙物)中所包含的角 1274845 落位置,以利精準操控機器主體執行工作,俾達提高該產業之 利用價值。 【實施方式】 有關本發明使用單一鏡頭偵測物體邊角及距離之裝置與方 法所採以技術手段及達成功效、目的,兹以最佳實施例配:圖 示、元件符號予以詳述說明如下·· 首先,請參閱圖-所示,係為本發明使用單一鏡頭所需硬 體裝置之示意圖; 本發明使用單-鏡頭硬體裝置制物體邊角及距離之裝 置,其係包括: 單-電何搞合器件鏡頭i,負責擷取機器主體之外界環境, 並以無線通δίΐ將影像回傳至主控電腦2 · 主控電腦2,内建有影像分析系統21,在以無線接收單—1274845 IX. Description of the Invention: [Technical Field] The present invention relates to a device and method for detecting the corners and distances of an object using a single lens, in particular, for a visual sensing device for a machine body, Adopting a single charge-coupled device (CCD) lens to identify the external environment (such as targets or obstacles), and designing the distance, direction, and orientation of the target by means of an image analysis system. The control of the main body (such as robots) to the outside world and the environment is the primary innovative design of the invention, which has great industrial value. [Prior Art] According to the current level of automation in various industries, it is becoming more and more common to use the main body of the machine to connect the visual system, namely MachineVision, to assist positioning, size measurement and flaw detection. Especially in the electronics industry, the most common use, such as die bonding machine, wire bonding machine, foot inspection machine, printing quality inspection machine, printed circuit board (PCB) inspection machine, etc., are typical applications. • Machine vision (MachineVisi〇n) is an optical recognition system that mimics human vision. It is a unique and promising sensory technology in the sensing system. It uses cameras and computers to capture. Analyze and interpret the content of the image, and then make certain decisions on the machine body; the research on Machine Vision (MachineVisi〇n) has more than 30 years of history, although it still does not have the power and wisdom of human vision, but for the controllable environment Under the clear task of work, machine vision can accomplish tasks that humans can't accomplish with 100% correct efficiency and continuous work 24 hours a day. 1274845 However, due to the machine vision tools used in the traditional industry, the use of multiple charge-coupled device (CCD) lenses at different angles to individually handle the target of the external environment (4) Yang system f彡, so that the computer must accept a variety of different visual signals, and then multi-angle images are synchronized for the prescription month (3 only to see the distance of objects or obstacles, so the traditional machine vision (4) hardware The requirements for injury are large, so that the resource cost is high. In addition, it must use the overlapping position of the target or obstacle in multiple pictures to calculate the correct position. Therefore, in the process of processing and analysis and analysis, the amount of data is often caused. Quite huge, ^ The problem of the correlation between the image and the object of the mother's picture makes the speed of the computer and the machine body too slow, which affects the efficiency of the machine body (such as the robot). The industry's advantage in machine vision is the lack of traditional machine vision. Based on this, the inventor of the present invention used multiple charge dissipators to solve the above-mentioned conventional problems. (10)) Lens, it is necessary to use the position overlap area of the target in the picture to calculate the position and the cost of the hardware equipment is high, the data analysis demand explosion and the efficiency of the work to improve the work efficiency of using machine vision, so the thinking And the success of the [only] use of the single-charge recombination device ((10) lens, and with the X-ray can be 70 complete image analysis I can complete the rapid detection of the target (or obstacle) situation, The target (or obstacle) environment, the target (or obstacle) azimuth, the target (or obstacle) direction angle, and the corner position contained in the target (or obstacle) to facilitate precise manipulation of the machine body Execution work, 1274845 俾 提高 提高 提高 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 。 On the visual sensing device of the robot, 'only need to single-electrical coupling device ((10) lens can identify the outer ring t' and can know the distance of the target (or obstacle) , direction, orientation, etc. for the use of the main control of the Thunder brain luxury 6 built 70 positive image analysis system, to effectively solve the traditional machine vision applied to the target (or obstacle) ranging, must use more than the knowledge & The problem of the distance of the object (or obstacle), and the use of a single charge surface-matching device ((10) lens for identification, to find the target (or obstacle) characteristics and other information of the target (or obstacle) characteristics, The device requirements can be lower than the traditional multi-charged light-coupled device (CCD) lens, which can save redundant equipment and reduce resource cost, and the amount of data to be analyzed is small, and the identification speed can be increased. The main purpose is to provide a method for detecting the corners and distances of objects using a single lens. The main Xiangdan-charge surface-mounting device ((10)) lens is taken by the computer built-in image analysis system with two secants after capturing the image. Chasing the corner detection algorithm that compares the slope between them, which can correctly calculate the possible corners of each object boundary. Therefore, it can be applied to a single battery. The coupling device (CCD) camera to find the boundary of the object feature recognition, such as the original point positioning items and the like may use the search objects. 1274845 In order to achieve the purpose of the present invention, the use of a single-lens-detecting object to separate the corners of the garment and method, the I system includes: single-charge-consuming device ((10)) lens 'responsible for taking the machine body In the external environment, the image is transmitted back to the main control computer by wireless communication; the main control computer has an image analysis system built in, and after receiving the image of the single electric homozygous device (10), it immediately analyzes and calculates the target object. == The command to control the main body of the machine is also transmitted to the main body of the machine by the wireless communication system; the main body of the machine (such as the robot (RGbGt)), the built-in wireless receiving module is used to receive the command of the main control computer. To control the main body of the machine to perform the required work 'to control the movement of avoiding objects or working outside; the method is to directly extract the specific object wheel line from the single-charge surface-mounting device (CCD) lens, borrowing from the computer The built image analysis system uses two secants to follow each other and compare the corners of the slope between them to analyze the contour information in the image, that is, to use two secants on the rim segment. Find all the corners in the image that match the corner features, and at the same time calculate the correct data of the target, and correctly obtain the distance between the target and the lens, the relative orientation of the target and the lens, and the direction angle of the obstacle. According to the above, the design of the present invention can complete the rapid detection and identification of the target object by using a single charge coupled device (CCD) lens and using a complete image analysis system built in the connection master computer. (or obstacle H = condition, target (4) chapter obstruction) environment, target (or obstacle) azimuth, target (or obstacle) direction angle and the angle contained in the target (or obstacle) 1274845 The position of the drop is used to precisely control the main body of the machine to perform the work, and the use value of the industry is improved. [Embodiment] The device and method for detecting the corner and distance of the object using a single lens of the present invention adopt the technical means and achieve the effect. And the purpose of the best embodiment: the illustrations, component symbols are described in detail below. · First, please refer to the figure - as shown in this figure. A schematic diagram of a hardware device required for using a single lens; the device for manufacturing a corner and a distance of an object using a single-lens hardware device, comprising: a single-electric device lens i, responsible for capturing the machine body The external environment, and the image is transmitted back to the main control computer by wireless communication. 2. The main control computer 2 has an image analysis system 21 built therein, and receives the data wirelessly.
電荷搞合器件鏡頭丨之影像後,便立即分析並計算目標物之資 訊’將欲控制機器主體3之命令,亦以無線通訊系統Μ傳送至 機器主體3 ; 機器主體3 (如機器人、 成為人U〇b〇t)),内建無線接收模組31 上,為接收主控電腦2所僂逵之人人 , 4寻運之〒中,以控制機器主體3執行 控制迴避物體或對外界作工之運動。 再者’請參閱圖二所示,传主 叮不係為本發明影像分析系統之割線 斜度比較演异法之示意圖; 1274845 一本發明使用單-鏡·測物體邊角及距離之方法,係由單 ^荷轉合器件鏡頭u取出具體之物體輪廓線,藉影像分析 用、,21進行分析影像中的輪廓資訊,該影像分析系統即利 線斜度比較演算法(如圖二所示),以兩條割線於輪廓線段 上找出影像中的所有符合角落特徵的角落,並同時計算出目標 物之正確數據,來獲取目標物之辨識與鏡頭之距離、目標物與 、兄員之相對方位场礙物之方向角,且亦可透過檢測人機介 :’將目標物之三點座標、使用單一電荷搞合器件鏡頭i之設 疋參數、所計算出目標物之數據’係包括目標物與單一電荷耦 合器件鏡頭1之距離、目標物與單一電荷耗合器件鏡頭i之相 對方位及障礙物之方向角,以及由影像分析系統Η所計算出目 標物之其他角落的距離數據,俾利機器主體3得以精確且快速 完成其欲執行之工作。 亦即,上述影像分析系統21係利用割線斜度比較演算法, 進行角落的偵測,其邊緣摘測利用邊緣具有亮度或灰階度急遽 轉變的特性’並對影像微分纽處理,強化影像上緣邊界,並 設定邊界Η檻值以求出_條描繪輪靡的線段,再利用近鄰串聯 相鄰邊界’建構出所求之邊界鏈;而設定邊緣強度的門插值, 以去除多餘強度不夠的雜訊線段; §已卞取出具體之物體輪廓線後,為了進行下一步的幾何 樣本圖片比對,先分析影像中輪廓的分佈資訊,由分析出的綠 1274845 ㈣廓/ ’在其線段設定兩條割線运與巧,線段起始值由輪摩 線段之端點1始且2點必須引導在'點前,可以設d+2、 2 、印斜度叫=〇^2 ^))/(6+3-4及冠斜 所2 =(Xx2 +2)-Xx2))/2 ; x2 由戈去 ^ 之始,在母一個執行階段增量移動一 個位置,並檢杳h—所2丨县不士丄 左 彳欢一 疋否有大於門檻值ΐ;而門檻值τ可以 奴著不同的目標物而設定其偵測角落的靈敏度;After the charge is applied to the image of the lens of the device, the information of the target object is immediately analyzed and calculated. 'The command to control the main body of the machine 3 is also transmitted to the main body 3 by the wireless communication system; the main body of the machine 3 (such as a robot, becoming a human) U〇b〇t)), on the built-in wireless receiving module 31, in order to receive the host computer 2, in the middle of the search, to control the machine body 3 to perform control to avoid objects or to the outside world Work movement. Furthermore, please refer to FIG. 2, which is not a schematic diagram of the slanting degree comparison of the image analysis system of the present invention; 1274845 A method for using a single-mirror to measure the corner and distance of an object, The specific object contour is taken out by the single-transfer device lens u, and the contour information in the image is analyzed by image analysis, and the image analysis system is a gradient comparison algorithm (shown in FIG. 2). ), use two secants to find all the corners in the image that match the corner features on the contour line, and calculate the correct data of the target at the same time to obtain the distance between the target and the lens, the target and the brother. The relative azimuth field obstructs the direction angle of the object, and can also detect the human-machine interface: 'Set the three-point coordinate of the target, use a single charge to fit the parameter of the device lens i, calculate the data of the target' The distance between the target and the single charge coupled device lens 1, the relative orientation of the target and the single charge-consuming device lens i, and the direction angle of the obstacle, and the calculated image by the image analysis system Other data from the corners thereof, the machine body 3 serve to facilitate accurate and rapid completion of its work to be executed. That is, the image analysis system 21 uses the secant slope comparison algorithm to perform corner detection, and the edge smearing uses the characteristic that the edge has a sharp transition of brightness or gray scale', and the image differential processing is processed to enhance the image. Edge boundary, and set the boundary Η槛 value to find the _ strip to depict the line segment of the rim, and then use the adjacent adjacent series boundary to construct the desired boundary chain; and set the edge strength of the door interpolation to remove the excess intensity After the specific object contour has been extracted, in order to carry out the next geometric sample image comparison, the distribution information of the contour in the image is analyzed first, and the analyzed green 1274845 (four) profile / 'set two in its line segment The secant is transported and skillful. The starting value of the line segment starts from the end point 1 of the wheel line segment and the 2 points must be guided before the point. You can set d+2, 2, and the slope is called =〇^2 ^))/(6 +3-4 and the crown oblique 2 = (Xx2 + 2) - Xx2)) / 2; x2 From the beginning of the go to ^, incrementally move a position in the mother's implementation stage, and check the h-shen 2 county If you don’t have a 丄 丄 彳 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 疋 ΐ The object set with its corner detection sensitivity;
“復,請參閱圖三,係本發明割線斜度比較演算法求取角 落之結果圖; …關於本發㈣像分㈣統21以職斜度比較演算法所計 异求出之結果,在本例中,以設定門檻值Τ值為0.6,若 卜1 - W2| 2 ,則A設為角落,並設^且繼續接下去搜 個角落; 對上述割線比較演算法其流程規則可由下所示: (1) ^, =(Xx2 +3)-^(,,))/^ +3-Xi)).W2 =(ΧΧ2 +2)_Χχ2))/2> (2) IFK -^|^Tthenrecordx2 to be a corner; let =x2;g〇t〇(i). (3 ) x2 = a +1; goto (l) 由此即可以找出影像中的所有符合角落特徵的角落,且 角落不受其朝向角度之影響。 繼而,請參閱圖四所干 μ岡也 所不,此圖為使用早一鏡頭偵測物體距離 之方法的影像透視投影示意圖; 為建立-個影像近似線性投影模型,即利用透視投影的方"Repeat, please refer to Figure 3, which is the result of finding the corner of the secant slope comparison algorithm of the present invention; ...About the result of the calculation of the difference between the scores of the four points (4) In this example, the threshold value is set to 0.6, if 1 - W2| 2 , then A is set to the corner, and ^ is set and continues to search for a corner; the flow rule of the above secant comparison algorithm can be as follows Show: (1) ^, =(Xx2 +3)-^(,,))/^ +3-Xi)).W2 =(ΧΧ2 +2)_Χχ2))/2> (2) IFK -^|^ Tthenrecordx2 to be a corner; let =x2;g〇t〇(i). (3) x2 = a +1; goto (l) This will find all the corners in the image that match the corner features, and the corners are not It is affected by the angle of the orientation. Then, please refer to Figure 4 for the image of the perspective projection of the method of detecting the distance of the object using the early lens; to establish an approximate linear projection model of the image, The side that uses perspective projection
1274845 法,將成像在顯示器的物體作視覺轉換,以得到實際物體的距 離,而轉換方程式(1)、(2)、(3)可表示顯示器上物件角落的一 個座標點,其中所示L!為投影鏡片至電荷耦合器件(CCD) 的投影線、α為投影點垂直至投影原點之角度、θ為投影點水平 座標至投影原點之角度,如式(4)、(5)所示。 X = Zj COS^J COS^! y = LY sin^ z- Lx sin^ cos^ p -i」x2 + y2 (9 = tan -— a = tan (1)(2) (3) (4) (5) 實際物體座標轉換可由(6)、(7)、(8)式表示,其中z;為實 際投影距離、他;為實際物體投影之座標點,由此可求得目標 物之距離。 ” x\ - L[ cos ^ cos at y[ = L\ sin at z\ = L\ sin θί cos at 清參看圖五所示 簡化投影示意圖; (6) (7)(8) 係為本發明使用單一鏡頭所產生影像透視 田單電祷輕合器件鏡頭1之高度與目標物目標點7軸為同 等兩度時’可將三維投影化簡為二維投影。其中β、A、 為目 12 1274845 標物三邊與單一電荷耦合器件鏡頭】中心的夾角、r為目標物與單 一電荷耦合器件鏡頭1的水平夾角、目標物邊長可用(9)、(10)、 (11)、(12)式表示。 xx-x2=owc〇sr v 9 ) Z1-Z2=^ sinr (1〇) z3 — z9 = Or cosr (11) x2 = Or sin r (12) 其中0^仏為目標物之邊長,由其邊長與角度的關係可得 (13)、(14)、(15)式。 tan & : _ z2+Ow sin/ x2 +Ow cos/ (13) tan^2 = x2 (14) tan = _ z2 + Ol cos γ x2 - 〇L sin γ (15) 由上述(13)、(14)、(15)式可以得到鏡頭至目標物的距離、 目標物的方位及目標物的方向。 請參看圖六所示,係為本發明單一鏡頭偵測物體距離之方 法的檢測人機介面; 如圖中之編號4.即X1、X2、X3為本發明於影像分析系統使 用割線比較演算法在影像中所找出目標物之三點座標,如圖中 之編號5·為使用單-電荷搞合器件鏡頭i之狀參數,如圖中 之編號6.為使財發明上財法料算出目標物之數據,其包 13 1274845 括目標物與單一電荷耦合器件鏡頭丨之距離、目標物與單一電 荷耦合器件鏡頭1之相對方位及障礙物之方向角,如圖中之編 號7.為系統使用本發明上述方法所計算出目標物之其他角落的 距離數據。The 1274845 method visually converts an object imaged on the display to obtain the distance of the actual object, and the conversion equations (1), (2), and (3) represent a coordinate point in the corner of the object on the display, where L! The projection line from the projection lens to the charge coupled device (CCD), α is the angle from the projection point perpendicular to the projection origin, and θ is the angle from the horizontal coordinate of the projection point to the projection origin, as shown in equations (4) and (5). . X = Zj COS^J COS^! y = LY sin^ z- Lx sin^ cos^ p -i"x2 + y2 (9 = tan -— a = tan (1)(2) (3) (4) ( 5) The actual object coordinate transformation can be expressed by equations (6), (7), and (8), where z is the actual projection distance, and it is the coordinate point of the actual object projection, thereby obtaining the distance of the target object. x\ - L[ cos ^ cos at y[ = L\ sin at z\ = L\ sin θί cos at clear schematic diagram of the simplified projection shown in Figure 5; (6) (7)(8) is a single use of the present invention The image produced by the lens is the same as when the height of the lens 1 is equal to the target target point 7 axis. The 3D projection can be reduced to a 2D projection. Among them, β, A, and 12 1274845 The angle between the center of the object and the single charge-coupled device lens, r is the horizontal angle between the target and the single charge-coupled device lens 1, and the target length is available (9), (10), (11), (12) Xx-x2=owc〇sr v 9 ) Z1-Z2=^ sinr (1〇) z3 — z9 = Or cosr (11) x2 = Or sin r (12) where 0^仏 is the side length of the target, The equations (13), (14), and (15) are obtained from the relationship between the side length and the angle. Tan & : _ z2+Ow sin/ x2 +Ow cos/ (13) tan^2 = x2 (14) tan = _ z2 + Ol cos γ x2 - 〇L sin γ (15) From the above (13), ( 14), (15) can get the distance from the lens to the target, the orientation of the target and the direction of the target. Please refer to Figure 6 for the detection of the human-machine interface of the method for detecting the distance of a single lens of the present invention. As shown in the figure, X1, X2, and X3 are the three-point coordinates of the target object found in the image by the slash comparison algorithm in the image analysis system, as shown in the figure, the number is 5. The charge is matched with the parameter of the lens i of the device, as shown in the figure 6. In order to calculate the target data of the financial invention, the package 13 1274845 includes the distance between the target and the lens of the single charge coupled device, and the target The relative orientation with the lens 1 of the single charge coupled device and the direction angle of the obstacle, as shown in the figure 7. The distance data of the other corners of the target are calculated by the system using the above method of the present invention.
據此,當本發明使用影像分析系統以割線斜度比較演算 法、經計算求得數據結果以檢測人機介面之顯示,清楚將目標 物之三點座標、使用單一電荷耦合器件鏡頭丨之設定參數、所 計异出目標物之數據’係包括目標物與單—電荷輕合器件鏡頭上 之距離、目標物與單-電餘合器件鏡頭i之相對方位及障礙 物之方向角,以及由影像分析系統21所計算出目標物之其他角 落的距離數據,配合欲下達機器主體3作動之命令,透過無線 通訊系統32傳送於機器主體3之無線接收模組31,得以操控機 器主體3進行精確且快速地完成其欲執行之工作。 綜上所述,上列詳細說明係針對本發明之一可行實施例之 具體說明,惟該實施例並非用以限制本發明之專利範圍,凡未 脫離本發明技藝精神所為之等效實施或變更,均應包含於本案 之專利範圍中。 综上所述,本案不但在空間型態上確屬創新,並能較習用 物品增進上❹項功效,應已充分符合龍性及進步性之法定 發明專利要件,爰依法提出巾請,騎#局核准本件發明專 利申請案,以勵發明,至感德便。 14 d 1274845 【圖式簡單說明】 圖一·係為本發明使用單一鏡頭所需之硬體裝置示意圖; 圖一·係為本發明運用割線斜度比較演算法之示意圖; 圖二:係為本發明運用割線斜度比較演算法求取角落之結 果圖; 圖四·係為本發明以單一鏡頭獲取影像透視投影示意圖; 圖五·係為本發明以單一鏡頭所得影像透視簡化投影示意 圖; 圖/、·係為本發明使用單一鏡頭之物體測距之方法之檢 ’則人機介面圖。 【主要元件符號說明】 1單一電荷耦合器件鏡頭 2主控電腦 21影像分析系統 22無線通訊系統 3機器主體 31無線接收模組 15Accordingly, when the present invention uses the image analysis system to use the image slanting degree comparison algorithm and the calculated data result to detect the display of the human-machine interface, it is clear that the three-point coordinates of the target are set using a single charge-coupled device lens. The parameter, the data of the measured object is included, the distance between the target and the single-charge coupling device lens, the relative orientation of the target and the single-electrical coherent device lens i, and the direction angle of the obstacle, and The distance data of the other corners of the target object calculated by the image analysis system 21 is transmitted to the wireless receiving module 31 of the machine body 3 through the wireless communication system 32 in response to an instruction to be actuated by the machine body 3, so that the machine body 3 can be manipulated to perform precise And quickly complete the work that he wants to perform. In the above, the detailed description of the present invention is not intended to limit the scope of the invention, and is not intended to limit the scope of the invention. , should be included in the scope of the patent in this case. In summary, this case is not only innovative in terms of space type, but also can improve the effectiveness of the Shangyu item compared with the customary items. It should have fully complied with the legal requirements of the legal and inventive patents of dragons and progressives. The bureau approved the application for the invention patent, in order to invent the invention, to the sense of virtue. 14 d 1274845 [Simple description of the diagram] Figure 1 is a schematic diagram of the hardware device required for the use of a single lens in the present invention; Figure 1 is a schematic diagram of the comparison algorithm using the secant slope in the present invention; The invention uses the secant slope comparison algorithm to obtain the result map of the corner; FIG. 4 is a schematic diagram of the perspective projection of the image obtained by the single lens in the present invention; FIG. 5 is a schematic diagram of the simplified projection of the image obtained by the single lens of the present invention; The system is a human-machine interface diagram for the method of measuring the distance of an object using a single lens. [Main component symbol description] 1 Single charge coupled device lens 2 Main control computer 21 Image analysis system 22 Wireless communication system 3 Machine main body 31 Wireless receiving module 15