1374252 九、發明說明： 【發明所屬之技術領域】 尤指一種專門檢測鋸條尺 本發明係為一種檢測設備 寸的機具’以及其檢測方法。 【先前技術】 在處處講求品質、另&方丄 、 支產力的今天，高生產， 品質已是成形工廠中重要 生產力及取佳 的提昇，就必須正確地t据 心 ’口。質 目雨並無專門針對鑛條相 / 統，因而無法提供録條加工J =寸…化檢測系 般以離線式（off-line)之 數據而— 尺十的方半 # θ 動、目視）或機械檢測鋸停 尺寸的方法’其量測誤差 4游俅 所防在丨,. #大。另外對產品檢測速度亦有 所限制，因此無法對產品 又方有 【發明内容】 T丰確及有效率的檢測。 有鑑於現有以人工戋 _ ¥ 4者機械檢測鋸條尺寸的方法且右 誤差大、效率差之缺失 丁，方/去具有 知月之主要目的在於接供一括 可以自動量測鋸條尺寸， 、，、種 〇 ^ Μ n 並知供鋸條加工機對鋸條加工之 口口貝W別及補償依據，以 ^ 旦'待.目丨曰壯sa 冋。精度與生產效率的錯條 影像測置裝置及方法。 Τ ％你 基於上述目·的，汰我 尺M 5的主要技術手段在於提供鋸條 尺寸參數之影像測量裝置，包括： 一控制農置，Α且古且 為具有影像處理運算能力且且有人機杵 制介面的裝置； ’人拭匕 一固疋平台，句也 u 匕括—基座、間隔設於基座頂面一背光 5 1374252 板與-移動座，背光板的板面是朝向移動座； 一夾具，係設於基座上並且位於背 間，於夾具上可夾持待測的鋸條； H、移動座之 -影像操取裝置，係設於移動座上 向與鋸條表面垂直，啰马# 砧取衫像的方 主夏0亥衫像擷取裝置與控制获罢+ ω、土 接，且接受控制裝置之控制對待一、私1化 it ^ H .., .’條進行影像擷取， 正肘祕取結果傳送至控制裝置； /、中。亥控制裝置依據所掏取的影像4胃 之尺寸參數。 ㈣旳〜像汁异該待測鋸條 二=:尺寸參數之影像測量方法，其步驟包括： 像；…的影像：以-數位攝影機拍攝待測之鋸條影 將擷取的影像經由灰階運算，轉換 執行灰階化轉換 為灰階影像； 以形成僅：換：將灰階影像進行二元化影像轉換， $、或白之二元化影像； ，矸鋸條影像邊緣之座標； ΐ測及運笪Μ /<r • 、條尺寸參數··以所搜尋的座標進行運瞀 鑛條之鑛齒的尺寸參數。 進-運- 貫施上 為， 义筏術手段以後，本發明可獲得的具體效益 1 . 日月目 '、備了自動定位量測或手動定位量測，量測 鋸條之齒矩、去4b 』 里叫 功能。 齒根深、齒弧曲率半徑、間隙角等多項檢測 6 /4252 2·本發明之輪送裝置可延伸至鋸條加工機的生產線， 以進行線上檢測，以提升生產效率。 3 本發明利用影像分析原理，可獲得準確、穩定的檢 測結果。 【實施方式】 凊參閱第一圖所示，為鋸條（1 0)的局部放大示意_， 鑛條（1 〇)為具有二側邊的扁平狀刀具，其中一側邊為平直 的邊’彖線’另一侧邊設有多數個連續且間隔設置的尖錐狀 鋸齒（11)，其尖錐端為齒尖（110)，二相鄰的鋸齒（11)的連 接處形成一圓弧形的齒根（12)，齒弧的二端分別與所鄰接 的齒尖（1 1〇)以一短切線（14)、一長切線（15)相切，本發明 即是為了檢測鑛齒（1彳）尖銳間隙角（0 )〔即與平直的邊緣 線平行的水平線（16)與長切線（15)的夾角）、二鋸齒之 間的齒距（Ρ )、齒根深（h )〔即齒尖（11 0)至齒根（1 2)的距 離〕、以及齒根（12)的齒弧曲率半徑（7 )。 本發明鋸條尺寸參數之影像測量裝置有二種類型：— 為第二圖所示的線上檢測系统；二為第三圖所示的非線上 «、!系統八中，第二圖所示可線上檢測的影像測量裝置 係包含-控制裝置（2〇)、—固定平台（3G)、_夾具（4〇)以及 -影像擷取裝置(50)’纟中，所述的控制裝置(2〇)為具有 影像處理運算能力且具有人機控制介面的裝置，其可以為 一部個人電腦或為-控制器，可執行-鑛條尺寸參數之事 像測量方法，本較佳實例之控制裝置（20)係為一個人電腦知 包括相互電性連接的—主機、—影示器以及—輸入單元， 與鋸條加工機連结，复 ° ”餘檢測方式均相同。1374252 IX. Description of the invention: [Technical field to which the invention pertains] In particular, a special instrument for detecting a saw blade is an apparatus for detecting an inch of a device and a detection method thereof. [Prior Art] In today's place where quality, other & production, productivity, high production, quality is an important productivity and improvement in the forming factory, it must be correctly. The color rain is not specific to the metallographic phase system, so it is impossible to provide the recording processing J = inch...the detection system is like off-line data - the ruler of the ruler #θθ, visual) Or the method of mechanically detecting the size of the sawing stop's measurement error 4 is prevented by 俅,. #大. In addition, there is a limit to the speed of product inspection, so it is impossible to have a product. [Inventive content] T is accurate and efficient. In view of the existing method of mechanically detecting the size of the saw blade by manual 戋 _ ¥ 4, and the lack of right error and poor efficiency, the main purpose of the square/going to know the moon is to automatically measure the size of the saw blade, ,,, Kind of 〇^ Μ n and know the edge of the saw blade processing machine for the processing of the saw blade and the basis for compensation, to ^ ' 待 丨曰 丨曰 丨曰 丨曰 sa 冋. A method of measuring the accuracy and production efficiency of the image measuring device and method. Τ % Based on the above objectives, the main technical means of removing the M 5 is to provide an image measuring device for the size parameter of the saw blade, including: a control farm, and an ancient image processing capability and man-machine Device for making interface; 'Personal wipes a solid platform, sentence also u 匕 — 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座 基座a clamp is disposed on the base and located between the backs, and the saw blade to be tested is clamped on the clamp; H. The movable seat-image manipulation device is disposed on the movable seat to be perpendicular to the surface of the saw blade, and the horse is mounted. # 取 取 像 像 夏 夏 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 亥 + + + + + + + + + + + ω ω ω ω ω ω ω ω ω Take, the elbow secret results are transmitted to the control device; /, medium. The Hai control device is based on the size parameter of the captured image 4 stomach. (4) 旳 像 像 像 像 待 待 = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = Conversion execution grayscale conversion to grayscale image; to form only: change: convert the grayscale image into binary image, $, or white binary image; , the coordinates of the edge of the saw blade image;笪Μ /<r • , strip size parameter ·· The size parameter of the ore tooth of the mining rod with the coordinates searched. In the future, the specific benefits that can be obtained by the invention are as follows: 1. The sun and the moon are equipped with automatic positioning measurement or manual positioning measurement, measuring the tooth moment of the saw blade, and going to 4b 』 is called function. Multiple detection of root depth, tooth radius of curvature, clearance angle, etc. 6 / 4252 2. The wheeling device of the present invention can be extended to the production line of a saw blade processing machine for on-line inspection to improve production efficiency. 3 The present invention utilizes the principle of image analysis to obtain accurate and stable detection results. [Embodiment] 凊 Refer to the first figure, which is a partial enlarged view of the saw blade (10). The mineral bar (1 〇) is a flat tool with two sides, one side of which is a flat edge. The other side of the twist line is provided with a plurality of continuous and spaced-pointed tapered serrations (11), the tip end of which is a tooth tip (110), and the joint of two adjacent serrations (11) forms an arc. The tooth root (12), the two ends of the tooth arc are respectively tangent to the adjacent tooth tip (1 1〇) by a short tangent line (14) and a long tangent line (15), and the present invention is for detecting the mineral tooth. (1彳) sharp clearance angle (0) [ie the angle between the horizontal line (16) parallel to the straight edge line and the long tangent line (15), the pitch between the two serrations (Ρ), the root depth (h) [ie the distance from the tip of the tooth (11 0) to the root (1 2)] and the radius of curvature of the tooth (7) of the root (12). There are two types of image measuring devices for the size parameter of the saw blade of the present invention: - the on-line detection system shown in the second figure; the second is the non-line on the third figure, the system is in the eighth, and the second picture is on the line. The detected image measuring device is a control device (2〇), a fixed platform (3G), a _clamp (4〇), and an image capturing device (50) 纟, the control device (2〇) The device having the image processing computing capability and having the human-machine control interface can be a personal computer or a controller, and can execute the image measuring method of the size parameter of the mineral bar. The control device of the preferred embodiment (20) It is a personal computer that includes the main body, the video recorder, and the input unit, which are electrically connected to each other, and are connected to the saw blade processing machine.
吏用時，蔣小A 以進行檢测分二二的Π1二持於夾具(4。)上^ 出不良品，另_方面i 測結果-方面可以檢 置（20)的軟體 机 匕外·可在控制裝 自動判°又疋鋸條角度、長度之範圍，藉由希 升=斷待測物件之優劣，而使產品自動進行八Γ 升產品的品質、良率。 曰轫進仃分類，以提 〇月參閱第四151你_ _ 的步驟包括： ’該鑛條尺寸參數之影像測量方法 取和'鑛條的影像：和制_罢,9 n、丄 取得所拍摄/主 '制裝置（2〇)由影像擷取裝置（50) 所拍攝待測鋸條的影像； ； 執仃灰階化轉換：將擷取的影 為灰階影像； 田次丨白運异，轉換 執行二元化轉換：將灰階影像 形成僅具有里布白夕. 又丄過一兀化處理， 式可以為Γ或^二70化影像，其中，二元化轉換的方 值），岔日^光在原始衫像中設定灰階值々（閥 )依照母個像素的灰階值大小（大於$ f於公、 轉換為愛έ斗、 A』（大方、或小於々），將像素 、為”、、色或白色，灰階值 > 々，則 值為255), “ u Η象素轉換為白色（灰階 以取得Λ 則像素轉換為黑色（灰階值為〇)， 取传㈣測鑛條的外輪廊影像，本較佳實施例之該闕值 知以Otsu演算法獲得； 料錯條影像邊緣之座標：搜尋雜條影像邊緣之座標 〔如=圖所示之二齒尖⑽)位置，即所標示座標3、$ 里測及運异錯條尺寸參數··以所搜尋的座標a、座標f 丁運"'鎇條之齒距（P )、齒根深（h )、齒弧曲率半徑（Γ )、 間隙角（0、笪# ' 、 數據，並藉由線性化、平均化作誤差修正， 以降低量測之誤差。 倘為線上檢測時，將所得到的各項檢測項目參數之數 锯鋸條加工機，以修正加工數據，再藉此數值建立 到u二質判斷及控管之具體定量分析依據；倘為非線上檢 測，則直接#么。 為。〇貝判斷及控管之具體定量分析依據，以 及修正加工之參考。 本發明『鑛條尺寸參數之影像檢測方法』是以 戶^制裝f Η的軟體「鑛條尺寸參數之影像檢測系統」 用C^n 4 S ^ ^使用高階程式語言軟體撰寫，其方式利 Camera影像擷取裝置擷取影像以幾何量測原理 1位影像處理法作鑛條之檢測。先由糊取 =，再將擷取到的影像作灰階運算，將影像轉換為灰階 值而、隹再^用演算法（〇tSU S method)，找出最佳閥 影像，行單一閥值運算’以將影像轉換為二元(黑白)數位 卜’再搜尋鑛條影像邊緣之座標來運算鑛條之齒距、齒 率半# '間隙角等數據，並藉由線性化、平 :乍δ吳差修正’其軟體程式介紹如下： 1.單一閥值運算 影像^階影像要轉為二元（黑白）數位影像，可以先在原始 小於5又疋火階值々，依照每個像素的灰階值大小（大於或 ;)，將像素轉換為黑色或白色： 火Ρ白值則像素轉換為白色（灰階值為255) 灰階值^ ’則像素轉換為黑色（灰階值為ο) 2· Otsu >臾异法（〇tsu，s method)When it is used, Jiang Xiao A will perform the test on the second and second sides of the clamp (4.) on the fixture (4.) on the defective product, and the other _ aspect i test results - the software can be installed (20) In the control device, the angle and length of the saw blade can be automatically judged, and the quality and yield of the product can be automatically carried out by using the advantage of the object to be tested. The classification of the 仃 仃 , 以 参阅 参阅 参阅 参阅 第四 第四 第四 第四 第四 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 参阅 该 该 该 该 该 该 该 该 该 该 该 该 该 该 该 该 该Shooting/main device (2〇) image of the saw blade to be tested by the image capturing device (50); ; Performing grayscale conversion: taking the captured image as a grayscale image; The conversion performs a binary transformation: the grayscale image is formed only with the lining white eve. After the 兀 兀 processing, the formula can be Γ or ^ 270 images, where the binary conversion is converted,岔日^光Set the grayscale value 阀(valve) in the original shirt image according to the grayscale value of the parent pixel (greater than $f in the public, converted to love έ, A) (generous, or less than 々), Pixel, ",, or color, grayscale value> 々, the value is 255), "u Η pixel is converted to white (gray order to obtain Λ then pixel is converted to black (gray value is 〇), The image of the outer porch of the test strip is taken (4), and the 阙 value of the preferred embodiment is obtained by the Otsu algorithm; : Search for the coordinates of the edge of the miscellaneous image (such as the two-toothed tip (10) shown in the figure), that is, the coordinates of the marked coordinates 3, $, and the error of the wrong dimension. · With the coordinates a, coordinates f "Foot" (P), root depth (h), tooth radius of curvature (Γ), clearance angle (0, 笪# ', data, and error correction by linearization and averaging, In order to reduce the error of the measurement. In the case of online inspection, the number of the test parameters obtained by the saw blade sawing machine is used to correct the processing data, and then the value is established to determine the specificity of the quality control and control. Analysis basis; if it is non-line detection, it is directly #. For the specific quantitative analysis basis of mussel judgment and control, and the reference of correction processing. The invention "image detection method of size parameter of mine" is based on household ^ The software for manufacturing f Η "image detection system for size parameters of mine bars" is written with C^n 4 S ^ ^ using high-level programming language software. The method is to capture images from the camera image capture device. Treatment method for the detection of mineral bars. =, then the captured image is gray-scaled, the image is converted to grayscale value, and then the algorithm (〇tSU S method) is used to find the best valve image, and a single threshold operation is performed. Convert the image into binary (black and white) digits and then search for the coordinates of the edge of the image to calculate the pitch of the ore, the half of the tooth rate, the gap angle, etc., and linearize and flatten: 乍δ吴The amendment 'the software program is introduced as follows: 1. The single threshold operation image ^ order image should be converted to binary (black and white) digital image, which can be first in the original less than 5 and the 疋 fire order value 々 according to the gray scale value of each pixel Size (greater than or;), converts pixels to black or white: Fire white value converts pixels to white (grayscale value is 255) Grayscale value ^ ' then pixels are converted to black (grayscale value is ο) 2· Otsu > 臾 method (〇tsu, s method)
將景{像值方圖視為一種機率分布 Pi'ntIN ;_像素灰階值為i·的機率 •灰階值為/像素數目 ":像素總數 若閱a值設為Λ ,則定義 w(k) = YJpi i~0 u^= Σλ 其中1為1灰階數（1-1則為最大灰階數卜按照定義， Hk) + u(k) = J^Pi = l /'=0 像 。十斤出（々)與’差異最大時的λ值可以先定義影 平均為 L~\Consider the scene {image value graph as a probability distribution Pi'ntIN; _ pixel gray scale value is the probability of i · gray scale value / number of pixels ": total number of pixels if the value of a is set to Λ, then define w (k) = YJpi i~0 u^= Σλ where 1 is the order of 1 gray scale (1-1 is the maximum gray scale number according to definition, Hk) + u(k) = J^Pi = l /'=0 image. The value of λ when the difference between the ten kilograms (々) and the 'the difference is the largest can be defined firstly, the average value is L~\
Ur = Σ;Λ· 找出下列函數值最大的々值 [uMk)~un^Ur = Σ;Λ· Find out the maximum value of the following function values [uMk]~un^
Hk)u{k) 3 _鋸條之齒距量測 利用單一閥值運算與〇tsu演算法料辦a 虛理，脸！ ，貞开法對鋸條影像做影像 ’衫像轉換為二元（黑白）數位影像， 像輪廓邊緣之座標a與座標f(如第 ^條影 條之齒距。 囫之-齒太)來運算鋸 錄條之齒根深量測 12 ^/4252 利用單~閱值運算與0tsu演算 處理，將影像轉換為…（Έ 子鋸條影像做影像 m .. ^ 导換為—兀（黑白）數位影像，$ 家 像邊緣之齒弧頂點座桿再禮尋鋸條影 之齒根深。 與座標a(如第―圖）來運算鑛條 5·鋸條之齒弧曲率半徑量測 利用單一閥值運算與〇tsu 處理，將影像轉換為二元（黑白）數錯條影像做影像 像邊緣之座標，如第一圖之 _ ，再搜尋鑛條影 。長切線（15)的相切處,、以 I·占c齒弧 至Η 3上的—點b，且Γ卧 至d點⑽#，w e i b 且C點 乃，C點座;p為 . 而b點座標為JC,、Hk)u{k) 3 _ saw blade pitch measurement Using a single threshold operation and 〇tsu algorithm to do a virtual, face! The splitting method is used to convert the image of the saw blade image into a binary (black and white) digital image, such as the coordinate a of the contour edge and the coordinate f (such as the pitch of the shadow strip of the ^^. The tooth root depth measurement of the saw bar is 12 ^/4252. Using the single-read value calculation and 0tsu calculation processing, the image is converted into... (the image of the saw blade image is m.. ^ is converted to - 兀 (black and white) digital image, $ Home edge of the tooth arc vertex seat bar and then look for the tooth root depth of the saw blade. With the coordinate a (such as the first figure) to calculate the radius 5 of the saw blade, the radius of curvature of the saw blade using a single threshold operation and 〇tsu Processing, convert the image into a binary (black and white) number of staggered images to make the coordinates of the edge of the image, such as the first image, and then search for the shadow of the mine. The tangent of the long tangent (15), with I· c tooth arc to — 3 - point b, and Γ lying to d point (10) #, weib and C point is, point C; p is. and point b is JC,
座铩為A、凡，d點座樟A 三點的座P分別將b、c、d «j 度 ί不；t,、乃、λ-2、 η、 3 &代入以下所示之圓方程式 )(2)、（3)，二個聯立方程式 = p . 求付二個未知係數C、D、 ’：求得的係數C、D'E帶人式⑷可求得齒弧曲率半徑。 ® ^Cx+Dy+E=〇 的形式，如式(1)、（2)、（3)。The seat P is A, where, the d point D is the third point of the seat P, respectively, b, c, d «j degrees ί; t, 乃, λ-2, η, 3 & into the circle shown below Equation) (2), (3), two simultaneous equations = p. Two unknown coefficients C, D, ': the obtained coefficients C, D'E with human (4) can obtain the radius of curvature of the tooth . ® ^Cx+Dy+E=〇 In the form of equations (1), (2), (3).
Cx,+Dyl+£ = -(X]2+yi2) ................ ⑴Cx, +Dyl+£ = -(X]2+yi2) ................ (1)
Cx2+Dy2+E = -{x\+yl) ...................... (2) ^3 + ^3+^ = -(-^3+^2) ................ ............ χ2+/+Οχ+ϋ)/+Ε=0=>(χ+^)2+()/+^)2=£!+〇^_Ε 齒弧曲率半徑；；；lc^~~Έ~2 V 4 +~4' — Ε.......................(4) 13 6. 錯條之間隙角量測 線性化，j·丨τ 取小平方法： 取得同—Λ + Α J 長切線（15)上多 對文的線性迴歸線為 ."占的座標（y,)，·..，&"，凡），設定 y y - b + αχ 使得 平方法求出迴歸係數 ~(b + αχ.))2 私.餐t卜我們可解出迴歸係數分别為Cx2+Dy2+E = -{x\+yl) ...................... (2) ^3 + ^3+^ = -(-^ 3+^2) ............................ χ2+/+Οχ+ϋ)/+Ε=0=>(χ +^)2+()/+^)2=£!+〇^_Ε Radius of curvature of the tooth;;;lc^~~Έ~2 V 4 +~4' — Ε......... ..............(4) 13 6. The gap angle measurement of the wrong strip is linearized, j·丨τ takes the small flat method: Get the same -Λ + Α J long tangent (15) The linear regression line of many pairs of texts is ."occupies the coordinates (y,), ·..,&", where), sets yy - b + αχ to make the regression method to find the regression coefficient ~ (b + αχ. )) 2 private. meal t we can solve the regression coefficient respectively
1門$ ΐ I即為第—圖中的Μ線段[即長切線(15)】的斜率 兵間P系角0 = tarr'( ό)。 【圖式簡單說明】 第圖係本發明所欲檢測之錄條的局部放大示音圖。 第二圖係本發明線上檢測型的鋸條尺寸參數之影像測 量裝置。 第二圖係本發明非線上檢測型的鋸條尺寸參數 測量裝置。 /衫 第四圖係本發明鋸條尺寸參數之影像 / β的流私 34 1374252 【主要元件符號說明】 (10)鑛條 (1 1)鋸齒 (110)齒尖 (1 2)齒根 (14)短切線 (1 5)長切線 (16)水平線 (20)控制裝置 (30)(30 A )固定平台 (31)基座 (32)輸送裝置 (320)承載單元 (3 3)移動座 (331)調整槽 (34)背光板 (40)夾具 (50)影像擷取裝置 (51)數位攝影機 (52)鏡頭 (60)鋸條加工機 (P )齒距 (h )齒根深 (r )曲率半徑 (0 )間隙角。1 door $ ΐ I is the slope of the Μ line segment [ie long tangent line (15)] in the first figure. The P line angle between soldiers is 0 = tarr' ( ό). BRIEF DESCRIPTION OF THE DRAWINGS The figure is a partially enlarged schematic view of a recorded strip to be detected by the present invention. The second figure is an image measuring device for the size of the saw blade of the present invention. The second figure is a saw blade size parameter measuring device of the present invention. / The fourth figure of the shirt is the image of the size parameter of the saw blade of the present invention / β flow private 34 1374252 [Description of the main components] (10) Mineral strip (1 1) sawtooth (110) tooth tip (1 2) tooth root (14) Chopped line (15) Long tangent (16) Horizontal line (20) Control device (30) (30 A) Fixed platform (31) Base (32) Conveying device (320) Carrying unit (3 3) Moving seat (331) Adjustment slot (34) Backlight (40) Fixture (50) Image capture device (51) Digital camera (52) Lens (60) Saw blade processing machine (P) Pitch (h) Root depth (r) Curvature radius (0 ) clearance angle.
1515