200911063 九、發明說明: 【發明所屬之技術領域】 目標 本發明涉及成像技術’特別涉及一種識別圖像中 物體的裝置和方法。 【先前技術】 取1豕孜術匕經/尹、 ,丨、丄/王 d ·吊-生活及醫病 $ ”,例如表面"K#(surfaceM_tin ⑽福哪’ SMT )。表面貼裝技術用於將電子元件 印刷電路板上以得到相應的電子產品,例如, ^BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an imaging technique', and more particularly to an apparatus and method for identifying an object in an image. [Prior Art] Take 1 豕孜 匕 / / Yin, 丨, 丄 / 王 d · hang - life and medical treatment $ ”, such as surface "K# (surfaceM_tin (10) Funao SMT). Surface Mount Technology Used to print electronic components on a circuit board to obtain corresponding electronic products, for example, ^
是藉由表面貼裴技術將北橋晶片#J 對應位置。為了將電子元件精確地定位在;:電=“ =貼裝技術即採用了成像技術來獲得電子元件:=’ =圖像中電子元件㈣別來判斷電子 藉由圖像識別確定電子元件大小、形狀 設備的電子子:=問題’如送入表面貼褒 轉表面貼ί 像不一致,則可以藉由移動、旋 和表面貼裝設備的吸嘴帝- % 與預定的圖像—致,r而;件移動或旋轉使電子元件 路板上的位置偏移。防止電子元件安裝在印刷電 傳統的識別圖像中目標物 體(例如電子元彳的方法為將獲得的目標物 對,根據與預定圖像進行逐個圖吻 移動、旋轉等別:標物體圖像的大小、形狀、 、 由於圖像的資料量通常較大, 200911063 .例如解析度為2〇48*2048的圖像’僅區分黑白單色的未壓 .縮圖像就有4M的資料量,大資料量的處理需要較強的計 鼻能力和較長的處理時間,這種傳統的圖像識別方法不利 於在工業生產、日常生活及醫療設備中的應用。 【發明内容】 u 鑒於此,有必要提供—種 體的方法。 丧速識別圖像中目標物 此外,還有必要提供—種^ ^ ^ ^ ^ ^ ^ 體的裝置。 陕速識別圖像中目標物 卿中目標物體的方法,包括如下步驟: =所述仏物體的計算掃財徑;以所述中心 _心,所述掃描半徑為半#對所述目標物體進行 知描,根據所述圓周掃描的&I 口周 -種“ 識別所述目標物體。 種識別圖像中目標物體的裝置,包括 組,用於獲得所述目標物體的中心;掃摇半^址^ 掃:半徑;掃描位址模組,用於根據;述中心: 知W径確定圓周掃描的圖元點的位址’·圖像存儲^ 70用於存儲所述圖像的資料’·掃描資料存儲扣一-於存館所述圖像存儲單元令 ' =早兀,用 較所述掃描資料在链时_ 、 車乂早兀,用於比 資料,伊Μ μ\、:早兀,、所述預定資料存儲單元中的 、4根據比較結果識別所述目標物體。 上述識別圖像中目標物體 描圖像中的目栌铷胂万沄/、裝置,稭由圓周掃 "的目軚物體,根據圓周掃描的結果 、 不而要對圖像進订逐個圖元的比對,可快速識別 200911063 ,圖像中目標物體。 【實施方式】 =所示’其為一較佳實施方式的圖像示意圖。圖像 4方ΐ更標物體12和背景14。為簡化描述使本 式更易於理解,目標物肋的形狀為三角形,杏狹 目標物體12也可以為矩行、 田然 形。 订或其他規則、不規則的圖 、、本貫施方式的圖像識別方法藉由獲得目標物如的中 二一再以射心為圓心作圓周掃描,根據在圓周上取得的The north bridge wafer #J corresponds to the position by surface mount technology. In order to accurately position the electronic components;: electricity = " = placement technology that uses imaging technology to obtain electronic components: = ' = electronic components in the image (four) do not judge the electronic image size to determine the size of the electronic component, The electrons of the shape device: = problem 'If the surface is attached to the surface, the surface is е, if it is inconsistent, the nozzle can be moved by the moving, rotating and surface mounting device - % with the predetermined image, r The movement or rotation of the member shifts the position on the electronic component board. The electronic component is prevented from being mounted on the target object in the conventional identification image of the printed circuit (for example, the method of the electronic element is the target pair to be obtained, according to the predetermined map Like moving one by one, moving, rotating, etc.: the size, shape, and size of the image of the target object are usually large, 200911063. For example, an image with a resolution of 2〇48*2048 is only distinguished from black and white. The color is not compressed. The image has 4M data volume. The processing of large data volume requires strong nasal capacity and long processing time. This traditional image recognition method is not conducive to industrial production and daily life. and Application in medical equipment. [Invention] In view of this, it is necessary to provide a method of seeding. It is necessary to provide a device of ^ ^ ^ ^ ^ ^ ^ in addition to the target in the image. A method for identifying a target object in a target object in the image of the image, comprising the steps of: = calculating a sweeping path of the object; and performing the target radius with the center radius Knowing, according to the circumference of the scan & I port - "identify the target object. The device for identifying the target object in the image, including the group, for obtaining the center of the target object; sweeping half ^ Address ^ Sweep: Radius; Scanning Address Module for use in; according to the center: Knowing the W path to determine the address of the circle point of the circle scan '· Image storage ^ 70 for storing the data of the image '· Scan data storage deduction - the image storage unit of the depository store is '= early, with the scan data in the chain _, the car 乂 early, for the ratio data, IΜ μ\,: early 兀, in the predetermined data storage unit, 4 identifies the location based on the comparison result Target object. In the above-mentioned identification image, the target object in the image of the target object, and the device, the target object of the straw sweeped by the circle, according to the result of the circumferential scan, does not have to order the image. The alignment of the primitives can quickly identify 200911063, the target object in the image. [Embodiment] = shown as a schematic image of a preferred embodiment. Image 4 squared object 12 and background 14 In order to simplify the description, the formula is more easily understood, the shape of the target rib is a triangle, and the apricot target object 12 can also be a rectangular line or a field shape. Order or other rules, irregular patterns, and patterns of the present embodiment The image recognition method is obtained by circumferentially scanning the center of the target by obtaining the target, such as the middle two, according to the circumference.
圖元點的集合盘預定咨斗立、隹,一 L /、預疋貝枓進订比較來識別目標物體。 目標物體12的中心122的獲得方法如 圖像10並記錄一行中备棚同-机^ 、仃怀才田 母们圖兀J的值,例如,將圖元為目 t物體12的記為h將圖元為背景14的記為0。如圖2所示, ,、為圖1中弟-掃描線31掃描的值與圖元的橫坐標即义座 票的關係圖’由圖2可知’圖元點如果為目標物體12,則值 為如f為背景14 ’則值為〇°類似地,如圖3及圖 4所不’第一、第二掃插線32、33可以獲得圖像10的相應行 點的值’進而可以判斷該圖元點是目標物體12還是 背景14。如果圖像在縱坐標方向有2048個圖元,則有2048 條掃描線’根據掃描線在圖像1〇中的位置,可以確定掃描 一 凡铩逐仃的線掃描過程中,對值為1的圖 凡點計數,例如掃描完圖像1G後值為1的圖70點總數為Ν, Ρ目払物體12的圖元點總數為Ν,則中心122的X座標為值 為1的圖tl點的X座標之和除以Ν ; γ座標為值為1的圖元點 200911063 .的Y座標之和除以N。目標物體12的中心122在座標確定 .後,即確定了中心122在圖像存儲中的位址。 獲得目標物體12的中心122後,計算目標物體12的掃插 半徑R。如圖5所示,掃描半徑R為以中心122為圓心的掃插 圓40的半役。知描圓40用於在掃描圓4〇的圓周上掃描,以 獲得圖像10在掃描圓40的圓周上的圖元點的值。掃描半押 R使得掃描圓40的面積與目標物體12滿足預定關係,從而 掃描半徑R可以由以下公式確定·· ^=λ[ν7κ ; 其中,Ν為目標物體12的圖元點總數,κ為預定常數, 例如 π、3.5、0.9π等。The collection point of the element point is scheduled to be confessed, 隹, a L /, pre- 疋 枓 枓 枓 枓 枓 枓 枓 枓 枓 识别 识别 识别 识别 识别 识别The method of obtaining the center 122 of the target object 12 is as shown in the image 10 and recording the value of the map in the row of the same machine, and for example, the symbol of the object 12 is h. The primitive is marked as 0 for background 14. As shown in FIG. 2, the relationship between the value scanned by the scan line 31 in FIG. 1 and the abscissa of the primitive, that is, the sense seat ticket, is shown in FIG. 2. If the pixel point is the target object 12, the value is For example, if f is the background 14', the value is 〇°. Similarly, as shown in FIG. 3 and FIG. 4, the first and second sweep lines 32 and 33 can obtain the value of the corresponding row point of the image 10, and then can be judged. Whether the primitive point is the target object 12 or the background 14. If the image has 2048 primitives in the ordinate direction, then there are 2048 scan lines 'according to the position of the scan line in the image 1〇, it can be determined that during the line scan of the scan, the value is 1 The graph of the point count, for example, the total number of points in FIG. 70 after the image 1G is 1 is Ν, and the total number of primitive points of the object 12 is Ν, then the X coordinate of the center 122 is a value of 1 The sum of the X coordinates of the point is divided by Ν; the γ coordinate is the sum of the Y coordinates of the primitive point 200911063 with a value of 1 divided by N. The center 122 of the target object 12 is determined at coordinates. After that, the address of the center 122 in the image storage is determined. After the center 122 of the target object 12 is obtained, the sweep radius R of the target object 12 is calculated. As shown in Fig. 5, the scanning radius R is a semi-dumb of the sweeping circle 40 centered on the center 122. The wipe circle 40 is used to scan over the circumference of the scan circle 4 to obtain the value of the feature point of the image 10 on the circumference of the scan circle 40. The scanning half-b is such that the area of the scanning circle 40 satisfies a predetermined relationship with the target object 12, so that the scanning radius R can be determined by the following formula: ·^=λ[ν7κ; where Ν is the total number of primitive points of the target object 12, κ is Predetermined constants, such as π, 3.5, 0.9π, and the like.
確定中心122即確定掃描圓40的圓心和掃描半徑R 後,在掃描圓40的圓周上掃描,以獲得圖像1〇在掃描圓扣 的圓周上的圖元點的值。圓周上的圖元點數量可以依據識 別精度確定,例如,將圖元點數量確定為2〇48,則每次掃 描對應的圓心角為2π/2048。根據中心i22的位址和掃描半 徑R,每次掃描圓周上的一個圖元點,並將掃描的圖元點 的值保存。 ^根據掃描一周後的圖元點的值,可以得到圖6所示的波 形圖。即在掃描過程中,對應的圓心角θ每次增加2兀/2〇48。 在θ為〇時,掃描點圖元位於背景14上,因此圖元點的值為 〇;在逐^增加圓心角0後,當掃描圓4〇與三角形目標物體 才父日才,則圖元點的值為i。根據圖6所示的波形圖,亦 即根據掃描所得圖元點的值的集合,可以獲得目標物體12 200911063 、的形狀。 • 如圖7所示,圖像7〇包括目標物體72與背景74。同樣 地,根據上述線掃描的方法,可以獲得目標物體72圖元點 總數Μ及中心722,再依據圖元點總數M及中心722可以獲 得掃描圓80。從而可以獲得掃描圓8〇的圓周上的圖元點的 值,掃描圓80的圓周上的圖元點的值的波形圖如圖8所示。 對比圖6和圖8可以看出,不同的目標物體12和72可以 獲得不同的波形圖,亦即可以獲得不同的圖元點的值的集 合’根據該圖元點的值的集合可以識別目標物體12和的 形狀。 如圖9所示’圖像90包括目標物體92與背景94。相對於 圖像10,目標物體92與目標物體12的形狀、大小相同,但 是旋轉了一個角度。同樣地,藉由掃描後獲得的波形圖如 圖10所示。對比圖6和圖1〇可以看出,圖1〇的波形圖可以藉 由平移圖6的波形獲得,即圖6和圖1〇的波形圖僅存在相位 上的差別。也就是說,相同形狀的目標物體,其旋轉的角 度可以藉由波形的相位差識別出來。 如圖11所示,圖像11〇包括目標物體112與背景114。相 對於圖像10,目標物體112與目標物體12的形狀、旋轉角度 相同,但是大小有差異。同樣地,藉由掃描後獲得的波形 圖如圖12所不。對比圖6和圖12可以看出,圖12的波形圖與 圖6的波形完全相同。但是目標物體12與目標物體ιΐ2大小 差異會‘致圖元點總數的差異,從而導致預定常數κ相同 的f月況下掃指半徑的差異。也就是說,才目同形狀的目標物 200911063 • ίϊ別出、ί t的差異可以藉由圖元點總數或掃描半徑的差異 隹人可、J l圓周知描所得的波形’即掃描的圖元點的 二目標物體的形狀;圓周掃描所得的波形的相 轉角产二圖元點的集合的偏移可以確定目標物體的旋 轉角度,目標物體的圖元 目標物體的大小。容易理解Jf’㈣描半徑可以霉定 易解,如果圖5中的目標物體12的位 但!’則會導致目標物體12的中心、122產生移動, 位、掃描獲得的波形圖、波形圖的相位都不會 別。 ” 2位置的移動不會影響圖像的識 …進—步地’為了更精確識別目標物體的形狀,可以採 =個_“標物體進行掃描。如圖一 二目上::體132與背景134,藉由設置兩個掃描圓 同,^ Γ圖14和圖15所示的波形。根據波形的^ 同攸而可^別圖7所示的矩形與圖13所示的多邊形。類 2如果而要識別的目標物體形狀比較複雜,可以藉由 k擇不同的預定常數κ獲得不同的掃 4 同的波形來識職雜的目標物體。 &而獲仔不 下二圖:6:'上述識別目標物體的方法可以歸納為如 下二,f先,步驟S1602’獲得圖像中目標物體的中心。 目“物體的中心可以藉由對目g % — 坐標分別求平均數1橫坐標=與縱 物體的Μ在位置。例如_所:::== 11 200911063 .描,將圖元點的值為丨的確定為目標物體圖元點,並統計目 標物體圖元點座標位址,根據統計結果求得目標物體的中 〇 步驟S1604,計算掃描半徑。掃描半徑的計算公式為· ’其中R為所述掃描半#,N為所述目標物體圖元 點的總數,K為預定常數。 …步驟S祕’以目標物體的—心為圓心,掃描半 半控對目標物體進行圓周掃描。圓周掃描過程中,圓周^ 各圖元點的位址A可以藉由如下方式計管. A=A〇 + dx + dy*W ; 其中,A0為中心的位址;^為掃描圖元點的位址 Π址的橫_量;彻描圖元點的位址與中心:址 =偏移! ; W為圖像的寬度,例如测個圖元。由於 二域圖己元^ 7址順序存儲的,因此圖像的縱向增 " ,、立址需要增加1個圖像寬度的位址。 體。將掃描過程中周掃描的結果,識別目標物 *對,根據比對結果判與預定資料進行 點的結果盥二& / 目軚物體。例如,提取的各圖元 為:角^㊉目標物體最為接近,縣目標物體識別 據:二I:可以根據中心位址確定三角形的位置,根 像a &物體圖元點 1 與預定資料的差異,二° “定目標物體的大小;根據 的旋轉角度。以 圖10與圖6的相位差判定目標物體 明參閱圖17,其為識別圖像中目標物體的裝置示意 12 200911063 n別裝置,包括中心位址模組5iq、掃 • 52〇、掃描位址模組伽、圖像存儲單元54〇、掃描資 單元550、預定資料存儲單元56〇、 、子儲 58〇。 匕季乂早兀570及计數器 中心位址模組5_於獲得目標物體的中心 Γ510對目標物體圖元點的橫坐標與縱坐標分別求平二 數,以棱坐標與縱坐標的平均數作為目標物體的中心所在 :,位置。具體計算方法已經就圖❻行過詳細描述,不再費述。 掃描半徑模組520用於計算掃描半徑。掃描半徑模电 別计异掃描半控的計算公式為:r=v^,其中㈣掃描半 徑:N為目標物體圖元點的總數,κ為預定常數。具體 方法如前所述。 τ 掃描位址模組530用於根據中心和掃描半徑確定圓周 掃描的圖元點的位址。掃描位址模組53〇包括偏差計算單元 532及位址計算單元534。偏差計算單元532與計數器划相 連,計數器580用於根據一個圓周需要掃描的圖元點數產生 =進位數字,例如,一個圓周需要掃描的圖元點數為2〇48 吟,則產生π位的二進位數字,即211=2〇48。偏差計算單 元532用於根據計數器58〇產生的二進位數字計算該二=位After determining the center 122, i.e., determining the center of the scan circle 40 and the scan radius R, scanning is performed on the circumference of the scan circle 40 to obtain the value of the image point on the circumference of the scan circle. The number of primitive points on the circumference can be determined according to the recognition accuracy. For example, if the number of primitive points is determined to be 2〇48, the corresponding central angle of each scan is 2π/2048. According to the address of the center i22 and the scanning radius R, one pixel point on the circumference is scanned each time, and the value of the scanned element point is saved. ^ According to the value of the primitive point after one week of scanning, the waveform diagram shown in Fig. 6 can be obtained. That is, during the scanning process, the corresponding central angle θ is increased by 2兀/2〇48 each time. When θ is 〇, the scanning point primitive is located on the background 14, so the value of the primitive point is 〇; after increasing the central angle 0 by ^, when the scanning circle 4〇 and the triangular target object are only the parent day, the primitive The value of the point is i. According to the waveform diagram shown in Fig. 6, that is, according to the set of values of the scanned primitive points, the shape of the target object 12 200911063 can be obtained. • As shown in FIG. 7, the image 7〇 includes the target object 72 and the background 74. Similarly, according to the above-described line scanning method, the total number of primitive points of the target object 72 and the center 722 can be obtained, and the scanning circle 80 can be obtained according to the total number M of the primitive points and the center 722. Thereby, the value of the primitive point on the circumference of the scanning circle 8 可以获得 can be obtained, and the waveform diagram of the value of the primitive point on the circumference of the scanning circle 80 is as shown in Fig. 8. Comparing Fig. 6 and Fig. 8, it can be seen that different target objects 12 and 72 can obtain different waveform patterns, that is, a set of values of different primitive points can be obtained. 'The target can be identified according to the set of values of the primitive points. The shape of the object 12 and. As shown in Fig. 9, the image 90 includes a target object 92 and a background 94. With respect to the image 10, the target object 92 has the same shape and size as the target object 12, but is rotated by an angle. Similarly, the waveform obtained by scanning is as shown in Fig. 10. Comparing Fig. 6 with Fig. 1A, it can be seen that the waveform diagram of Fig. 1A can be obtained by shifting the waveform of Fig. 6, that is, the waveform diagrams of Fig. 6 and Fig. 1 have only a phase difference. That is to say, the angle of rotation of the target object of the same shape can be recognized by the phase difference of the waveform. As shown in FIG. 11, the image 11 includes a target object 112 and a background 114. With respect to the image 10, the shape and rotation angle of the target object 112 and the target object 12 are the same, but the sizes are different. Similarly, the waveform obtained by scanning is as shown in Fig. 12. As can be seen by comparing Fig. 6 and Fig. 12, the waveform diagram of Fig. 12 is identical to the waveform of Fig. 6. However, the difference in size between the target object 12 and the target object ιΐ2 will 'cause the difference in the total number of primitive points, resulting in a difference in the radius of the swipe relative to the same f-month of the predetermined constant κ. That is to say, the target of the same shape 200911063 • ϊ ϊ ϊ 、 、 ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί ί The shape of the two target objects of the element; the phase angle of the waveform obtained by the circumferential scan produces an offset of the set of points of the two primitives to determine the rotation angle of the target object, and the size of the target object of the target object. It is easy to understand that the Jf' (4) trajectory radius can be easily solved. If the position of the target object 12 in Fig. 5 but!' will cause the center of the target object 12, 122 to move, the bit, the waveform obtained by scanning, and the waveform diagram The phase is no different. The movement of the 2 position does not affect the recognition of the image. In order to more accurately identify the shape of the target object, the target object can be scanned. As shown in Fig. 1 and 2: body 132 and background 134, by setting two scanning circles, the waveforms shown in Fig. 14 and Fig. 15 are obtained. The rectangle shown in Fig. 7 and the polygon shown in Fig. 13 can be distinguished from the same waveform. Class 2 If the shape of the target object to be identified is relatively complicated, it is possible to obtain a different target waveform by using different predetermined constants κ to obtain different target waveforms. And the above picture is taken: 6: 'The above method of identifying the target object can be summarized as follows. First, step S1602' obtains the center of the target object in the image. "The center of the object can be obtained by averaging the number of abscissas = the coordinates of the vertical object and the position of the vertical object. For example, _:::== 11 200911063. The value of the primitive point is丨 is determined as the target object primitive point, and the target object primitive point coordinate address is counted, and the target object is obtained in the middle step S1604 according to the statistical result, and the scanning radius is calculated. The calculation formula of the scanning radius is · 'where R is the The scan half#, N is the total number of the target object primitive points, K is a predetermined constant. ... Step S secret 'takes the center of the target object as the center, and scans the half control to scan the target object circumferentially. , circle ^ The address A of each primitive point can be calculated by the following method: A=A〇+ dx + dy*W ; where A0 is the center address; ^ is the address of the scanned primitive point The horizontal _ quantity; the address and center of the trace element point: address = offset!; W is the width of the image, for example, the measured primitive. Since the two domain maps are stored in order, the image The vertical increase ", the address needs to increase the address of one image width. The result of the mid-week scan in the process, identifying the target* pair, and determining the result of the point based on the comparison result, the second & / target object. For example, the extracted primitives are: angle ^ ten target object is closest County target object identification data: II I: The position of the triangle can be determined according to the central address, the root image a & object pixel point 1 and the predetermined data difference, two ° "the size of the target object; according to the rotation angle. The phase difference between FIG. 10 and FIG. 6 is used to determine the target object. Referring to FIG. 17, which is a device for identifying the target object in the image, the device includes a central address module 5iq, a scan address, a scan address, and a scan address. The module gamma, the image storage unit 54A, the scan resource unit 550, the predetermined material storage unit 56〇, and the sub-store 58〇.匕 乂 乂 兀 兀 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器 计数器The number is the center of the target object: position. The specific calculation method has been described in detail in the figure, and will not be described. The scan radius module 520 is used to calculate the scan radius. The calculation formula of the scanning radius mode differential counting half control is: r=v^, where (4) scanning radius: N is the total number of target object point points, and κ is a predetermined constant. The specific method is as described above. The τ scan address module 530 is configured to determine the address of the primitive point of the circumferential scan based on the center and the scan radius. The scan address module 53A includes a deviation calculation unit 532 and an address calculation unit 534. The deviation calculation unit 532 is connected to the counter, and the counter 580 is configured to generate a = carry number according to the number of primitive points that need to be scanned on one circumference. For example, if the number of primitive points to be scanned on one circumference is 2〇48 吟, the π-bit is generated. The binary digit is 211=2〇48. The deviation calculating unit 532 is configured to calculate the two= bits according to the binary digits generated by the counter 58〇
數字對應的掃描圖元點的位址與中心位址的橫向偏移量^ 及縱向偏移量dy。 X 偏差計算單元532包括查找表5322、第一偏差計算單元 5324、第二偏差計算單元5326及映射單元“π。 查找表5322與計數器580相連,用於存儲與計數器58〇 13 200911063 二進位數字對應的正弦值與佘弦值。由於正弦值* ί值的計算較為複雜’藉由存儲正弦值與余弦值在查找 53=Π以提高目標物體的識別速度。為了減㈣ 入、谷里’可以僅存儲坐標系統中第一象限的正弦值鱼 =弦值,對於η位的二進位數字來說,其中前2位二進位數 子可以用於表示四個不同的象限,後9位用於對應在第—象 限的正弦值與余弦值。 第-偏是計异單元5324用於計算第一象限中在橫坐標 量,即掃描半徑與余弦值的乘積R*c〇se;第二偏差 片""單凡5326用於計算第—象限中在縱坐標的偏移量,即 2半徑與!'弦值的乘積R*sine。映射單元53則於根據 I偏^十异單元5324與第二偏差計算單元5326的計算結 奐算關係如下表2斤示: 橫坐標的 偏移量.d:The horizontal offset of the address of the scanned primitive point corresponding to the number and the central address ^ and the longitudinal offset dy. The X deviation calculation unit 532 includes a lookup table 5322, a first deviation calculation unit 5324, a second deviation calculation unit 5326, and a mapping unit "π. The lookup table 5322 is connected to the counter 580 for storage corresponding to the counter 58〇13 200911063 binary digits. The sine value and the chord value. Because the calculation of the sine value * ί value is more complicated 'by storing the sine and cosine values in the search for 53 = Π to improve the recognition speed of the target object. In order to reduce (four) into, valley 'can only The sine value of the first quadrant in the coordinate system is stored as a chord value. For the n-bit binary digits, the first two digits can be used to represent four different quadrants, and the last nine digits are used to correspond to The sine and cosine values of the first quadrant. The first-bias is the calculation unit 5324 for calculating the amount of the abscissa in the first quadrant, that is, the product of the scan radius and the cosine value R*c〇se; the second deviation piece""Single 5326 is used to calculate the offset in the first quadrant on the ordinate, that is, the product of the radius of 2 and the !' chord R*sine. The mapping unit 53 is based on the I-differentiation unit 5324 and the second The calculation result of the deviation calculation unit 5326 2 kg of the relational table below shows: the abscissa offset 2.d:
縱坐標的 偏移量dv R*sin0 R*cos0 R*sin0 R*cos6 位址計算單元534用純據中心的位址以及掃描位^ ^中心位址的偏移量計算所述圓周掃描的位址。位址計篇 早/ 534計算圓周掃描的位址的計算公式為: 〇 X + dy W,其中,A〇為中心的位址;dx為掃描圖元點 的=址與中心位址的橫向偏移量,即掃描半徑與圓心角( 的示弦值的乘積;dy為掃描圖元點的位址與中心、位址的縱 14 200911063 像:ϊ:,與圓心角㈣正弦值的乘積’· W為圖 了以猎由讀取圖像存儲單元540獲得。 中圖==元5,於存儲圖像咖,例如存娜 中對單元550用於存儲所述圖像存儲單元_ 的資料。Τ田的位址的貧料’例如存儲圖5中圓周40掃描 預(貞料存料元56⑽於存儲預定資料。例 角形、矩形等目標物體的資料。 省― 比較早兀570用於比較掃描資料存儲單元55〇 ^儲單元細中的資料,《比較結果識別所述目、^ =識別圖像中目標物體的方法與裝置,藉由圓周掃 :圖像中的目標物體,根據圓周掃描的結果,識別所述目 ’不需要對圖像進行逐個圖元的比對,可快速識別 圖像中目標物體。採用查找表5322保存三角函數值,可以 快速查找三角函數值而不需計算,進—步加快了圖像的識 別速度。藉由設置不同的掃描半徑進行多次掃描,可以識 別複雜的目標物體。 綜上所述’本發明符合發明專利要件,爰依法提出專 利申凊。惟,以上所述者僅為本發明之較佳實施方式,舉 凡熟悉本案技藝之人士,在援依本案創作精神所作之等效 修飾或變化,皆應包含於以下之申請專利範圍内。 15 200911063 .【圖式簡單說明] 圖 圖1中包括 .條掃Li為—較佳實施方式的圖“意 圖’2為圖1中第—條掃描線的掃福站果 圖°為圖1中第二條掃描線的掃插:果。 圖4為圖!中第三條掃描線 。。 圖5為第-較佳實施方式的圖像 圖6為圖5的圓周掃描結果示意圖:知插示意圖。 二8 ^ I 7數、施方式的圖像圓周掃插亍立g 圖8為圖7的圓周掃描結果示音圖 疮不思圖。 圖9為第三較佳實施方式的圖像周 圖1〇她的圓周掃描結果示意圖“— 圖Π為第四較佳實施方式的圖像 圖调圖U的圓周掃描結果示意圖物苗示意圖。 式的圖像圓周掃福示意圖。 3 14為圖13的弟—圓周掃描結果示意圖。 圖15為圖13的第二圓周掃描結果示意圖。 法流Ξ圖 16。為一較佳實施方式的識別圖像中目標物體的a 目標物體的裝 固17為車乂佳實施方式的識別圖像中 置示意圖。 【主要元件符號說明】 圖像 目標物體 10 、 70 、 90 、 110 、 13〇 12 、 72 、 92 、 112 、 132 16 200911063 ,背景 弟·一掃描線 第三掃描線 掃描圓 中心 14 、 74 、 94 、 114 、 134 31 第二掃描線 32 33 40 、 80 ' 140 > 150 122 ' 722 識別裝置 500 掃描半徑模組 520 偏差計算單元 532 圖像存儲單元 540 預定資料存儲單元560 計數器 580 第一偏差計算單元5324 映射單元 5328 中心位址模組 510 掃描位址模組 530 位址計算單元 534 掃描資料存儲單元 550 比較單元 570 查找表 5322 第二偏差計算單元 5326 識別目標物體的方法流程步驟S1602〜S1608 17The offset of the ordinate dv R*sin0 R*cos0 R*sin0 R*cos6 The address calculation unit 534 calculates the bit of the circumferential scan by the address of the pure data center and the offset of the scan bit ^^ center address. site. The calculation formula of the address of the circumferential scan is calculated as: 〇X + dy W, where A〇 is the center address; dx is the lateral offset of the scan site point and the central address The amount of shift, that is, the product of the scan radius and the chord of the central angle; dy is the address of the scanned primitive point and the center, the vertical of the address 14 200911063 Image: ϊ:, the product of the sine value of the central angle (four)'· W is obtained by the hunted by the read image storage unit 540. The middle image == element 5, for storing the image coffee, for example, the storage unit 550 is used to store the data of the image storage unit _. The poor material of the address of the field is, for example, stored in the circle 40 of FIG. 5 to scan the pre-processing data (the storage material element 56 (10) for storing the predetermined data. The data of the target object such as the angular shape, the rectangular shape, etc.. The storage unit 55 stores the data in the unit detail, the comparison result identifies the target, and the method and device for identifying the target object in the image are scanned by the circumference: the target object in the image, according to the result of the circumferential scan , identifying the target 'does not need to compare the images one by one, Quickly identify the target object in the image. Use the lookup table 5322 to save the trigonometric function value, you can quickly find the trigonometric function value without calculation, and further accelerate the recognition speed of the image. By setting different scanning radii for multiple scans In view of the above, the invention meets the requirements of the invention patent, and the patent application is filed according to law. However, the above description is only a preferred embodiment of the present invention, and those who are familiar with the skill of the present invention, Equivalent modifications or changes made in the spirit of the creation of the case shall be included in the scope of the following patent application. 15 200911063 . [Simple description of the diagram] Figure 1 includes: a sweep of Li is the preferred embodiment The figure "Intent" is the sweeping station of the first scanning line in Figure 1. The sweep is the second scanning line in Figure 1. Figure 4 is the third scanning line in Figure! 5 is an image of the first preferred embodiment. FIG. 6 is a schematic diagram of the circumferential scanning result of FIG. 5: a schematic diagram of the insertion. 2 8 ^ I 7 number, the image of the application mode is circumferentially inserted, and the g is shown in FIG. Circular scan result Figure 9 is a schematic view of the circumference of the image of the third preferred embodiment. The schematic diagram of the results of the circumferential scan of the image map U of the fourth preferred embodiment is a schematic diagram of the circumferential scan result of the image map U of the fourth preferred embodiment. Fig. 15 is a schematic diagram of the result of the circumferential scan of Fig. 13. Fig. 15 is a schematic diagram of the result of the second circumferential scan of Fig. 13. Flowchart Fig. 16 is a preferred embodiment The fixing 17 of the target object of the target object in the recognition image is a schematic diagram of the identification image of the embodiment of the vehicle. [Main component symbol description] Image target objects 10, 70, 90, 110, 13〇12 , 72 , 92 , 112 , 132 16 200911063 , background brother · one scan line third scan line scan circle center 14 , 74 , 94 , 114 , 134 31 second scan line 32 33 40 , 80 ' 140 > 150 122 ' 722 recognition device 500 scan radius module 520 deviation calculation unit 532 image storage unit 540 predetermined data storage unit 560 counter 580 first deviation calculation unit 5324 mapping unit 5328 central address module 510 scan position The method of process steps 530 module address scan data memory unit 534 calculating unit 550 lookup table 5322 of the second deviation calculating unit 5326 to identify the target object comparison unit 570 S1602~S1608 17