CN105607218A - Image auto-focusing method measurement data transmission device and method based on fuzzy entropy - Google Patents
- ️Wed May 25 2016
Info
-
Publication number
- CN105607218A CN105607218A CN201510522454.XA CN201510522454A CN105607218A CN 105607218 A CN105607218 A CN 105607218A CN 201510522454 A CN201510522454 A CN 201510522454A CN 105607218 A CN105607218 A CN 105607218A Authority
- CN
- China Prior art keywords
- image
- focus
- fuzzy entropy
- lens
- fuzzy Prior art date
- 2015-08-24 Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Image Analysis (AREA)
- Automatic Focus Adjustment (AREA)
Abstract
本发明公开了一种基于模糊熵的影像自动对焦方法，引入模糊熵的概念对图像信息进行表征，定义图像的边缘特征矩阵，并基于像素的测度定义图像对焦评价函数，从而确定图像测量系统的精确对焦位置。对于复杂多变的工业测量现场，由于光照变化、振动等原因而包含大量噪声影响信号的图像，利用模糊熵能够更好的表达图像的灰度梯度信息。本发明对焦效率高，对焦准确，可广泛地应用于工业影像测量系统。
The invention discloses an image autofocus method based on fuzzy entropy, which introduces the concept of fuzzy entropy to characterize image information, defines the edge feature matrix of the image, and defines the image focus evaluation function based on the measurement of pixels, so as to determine the image measurement system. Precise focus position. For complex and changeable industrial measurement sites, images containing a large amount of noise affecting signals due to illumination changes, vibrations, etc., using fuzzy entropy can better express the gray gradient information of the image. The invention has high focusing efficiency and accurate focusing, and can be widely used in industrial image measuring systems.
Description
Technical field
The present invention relates to image processes and Autofocus Technology field, particularly a kind of image based on fuzzy entropy side of focusing automaticallyMethod.
Background technology
Along with the development of computer technology and Image Information Processing technology, the non-contact measurement based on image technique is fast because of itThe feature such as prompt, convenient, intelligent, is widely applied to every field. The basis that non-contact image is measured is to obtain figure clearlyPicture, focusing is the significant process that picture system obtains picture rich in detail automatically, is the key technology of NI Vision Builder for Automated Inspection.
Automatically focusing is by selecting suitable focusing evaluation function to evaluate gathered image, according to evaluation result, and shouldBy the focusing of searching algorithm searching image, then drive focus adjusting mechanism to make CCD quick and precisely arrive focal position.
Desirable Atomatic focusing method requires focusing evaluation function to have the features such as unimodality, acuteness and uniqueness. But, byVaried in focusing scene, illumination condition comes and go, and the evaluation of estimate function curve that makes to focus is not dull under many circumstancesLevel and smooth unimodal curve, but present multiple peak values, thus make focus search easily be absorbed in local peaking. Particularly many in complexityThe industrial radiographic measurement scene becoming, focus process may be absorbed in for a long time repeatedly vibrates back and forth and causes focusing unsuccessfully.
Summary of the invention
The object of invention is to overcome the deficiency of prior art, provides a kind of image Atomatic focusing method based on fuzzy entropy to survey numberReportedly conveying device and method, focusing efficiency is high, and focusing accurately, can be widely used in industrial image measuring system.
The technical solution adopted for the present invention to solve the technical problems is: a kind of image Atomatic focusing method based on fuzzy entropy is providedSurvey data transmission device, comprising: drive camera lens to travel through whole region of search with particular step size, in the preliminary examination position of camera lens and everyPosition acquisition one sub-picture of one step, preserves its corresponding lens location to each obtained sub-picture, by fuzzy entropy to oftenOne sub-picture information characterizes, the edge feature matrix of definition image, and define image focusing evaluation letter based on estimating of pixelNumber, thereby the accurate focusing position of definite measuring system of picture.
Preferably, described to the method for each sub-picture ambiguity in definition entropy for: establish obtained m × n dimension image and there is NGray level, ambiguity in definition collection A, the gray level that its domain is image, and the gray value of each pixel in image is normalizedProcess, the membership function of ambiguity in definition collection A is:
Wherein, k is the grey scale pixel value after normalized;
On fuzzy set A, ambiguity in definition entropy is:
EA(μA(f(i,j)))＝-(μA(f(i,j)))log(μA(f(i,j)))，
Wherein, in the time of f (i, j)=k, membership function maximum and fuzzy entropy minimum, and fuzzy entropy has symmetry at f (i, j)=kProperty.
Preferably, the method for the edge feature matrix of described definition image is: the gray value with each pixel in image buildsM × n ties up matrix M, and in matrix M, centered by capture vegetarian refreshments (i, j), size is a window of l × l, and wherein l is strangeNumber; On this window, define pixel (i, j) based on fuzzy entropy estimate for:
m A ( i , j ) 1 l × l Σ m = - ( l - 1 ) / 2 ( l - 1 ) / 2 Σ n = - ( l - 1 ) / 2 ( l - 1 ) / 2 E A ( μ A ( f ( i + m , j + n ) ) ) ,
Make f (i, j)=k, in the time that pixel (i, j) is marginal point, in window, the grey value difference of each point is large, mAThe value of (i, j)Also larger; Pixels all in image is calculated respectively and estimates mA(i, j), the edge feature matrix of design of graphics pictureM[mA(i,j)]m×n。
Preferably, described definition image focusing evaluation function is:
F＝∑(i,j)∈RectmA(i, j), wherein Rect is focusing window.
Preferably, calculate the maximum of the evaluation of estimate of each sub-picture obtaining, its defined lens location is camera lens pairBurnt position.
Preferably, drive lens moving to the lens focusing position obtaining, focusing finishes.
Preferably, described two input signal f (i, j) are Zone Full or the regional areas in obtained image.
Preferably, a kind of image Atomatic focusing method based on fuzzy entropy is surveyed data transmission device, comprises the steps:
A1, automatically focusing operation starts, and the camera lens leg speed of advancing is set to 0;
A2, obtain and preserve the collection image of camera lens current location;
A3, the camera lens leg speed i that advances adds 1, calculates lens location p according to selected step-length si=i × s, preserves lens location pi。
A4, intercept the focusing window of m × n in picture centre;
A5, m × n dimension image has N gray scale collection, ambiguity in definition collection A, and the gray scale collection that its domain is image, and to imageIn the gray value of each pixel be normalized, the membership function of ambiguity in definition collection A is:Wherein, k is the grey scale pixel value after normalized, 0.5≤μA(f (i, j))≤1; FixedJustice fuzzy entropy is: EA(μA(f(i,j)))＝-(μA(f(i,j)))log(μA(f(i,j)))；
A6, build m × n tie up matrix M with the gray value of each pixel in image, centered by capture vegetarian refreshments (i, j), size isA window of l × l, wherein l is odd number; On this window calculating pixel point (i, j) based on fuzzy entropy estimate for:
m A ( i , j ) 1 l × l Σ m = - ( l - 1 ) / 2 ( l - 1 ) / 2 Σ n = - ( l - 1 ) / 2 ( l - 1 ) / 2 E A ( μ A ( f ( i + m , j + n ) ) ) ,Pixels all in image is calculated and estimated respectivelymA(i, j), the edge feature matrix M [m of design of graphics pictureA(i,j)]m×n；
A7, definition focusing evaluation function F=∑∑i,j)∈RectmA(i, j), wherein Rect is focusing window, evaluates letter according to focusingThe evaluation of estimate Q of number computed imagei；
A8, judge whether to travel through whole hunting zone according to current lens location; If reached hunting zoneLarge value, enters steps A 9, otherwise returns to steps A 2;
A9, from gathered picture appraisal value array, find out maximum QmAnd corresponding lens location pm；
A10, calculate current camera lens position pnWith the maximum corresponding lens location p of picture appraisal valuemDistanced＝pm-pn；
A11, d is fed back to drive control part, drive camera lens to focusing position;
A12, focusing finish, so memory cell empties.
The invention has the beneficial effects as follows:
1) in the time image being focused to evaluation, the estimating for calculating pixel point by fuzzy entropy. Image is clear, imageIn the grey value difference of pixel larger, estimate greatlyr thereby calculate, can well reflect the readability of image, and anti-Making an uproar property is strong, to shooting environmental strong adaptability. Therefore obtain the evaluation of estimate of image according to the fuzzy entropy of image each point, have very highFocusing accuracy.
2) adopt motor fixed step size to travel through the searching algorithm of focusing window, motor only need move from left to right according to fixed step sizeOne time, do not need to move back and forth, therefore do not need to arrange motor and postpone, the search time of effectively having saved motor.
Below in conjunction with drawings and Examples, the present invention is described in further detail; But a kind of image based on fuzzy entropy of the present inventionAtomatic focusing method is surveyed data transmission device and method is not limited to embodiment.
Brief description of the drawings
Fig. 1 is the flow chart of the inventive method;
Fig. 2 (a) is the fuzzy brake block boss picture of focusing;
Fig. 2 (b) is the brake block boss picture clearly of focusing;
Fig. 3 is the contrast of Atomatic focusing method of the present invention and classical focusing method;
Fig. 4 is the contrast of Atomatic focusing method of the present invention under best enlargement ratio different illumination conditions;
Fig. 5 (a) crosses Atomatic focusing method of the present invention under weak condition and focusing side based on Breene operator at illumination conditionThe contrast of method;
Fig. 5 (b) crosses Atomatic focusing method of the present invention under strong condition and focusing side based on Breene operator at illumination conditionThe contrast of method.
Detailed description of the invention
Embodiment 1
Shown in Figure 1, a kind of image Atomatic focusing method based on fuzzy entropy of the present invention is surveyed data transmission device, comprising:Drive camera lens to travel through whole region of search with particular step size, in the preliminary examination position of camera lens and the secondary figure of the position acquisition of each step onePicture, preserves its corresponding lens location to each obtained sub-picture, by fuzzy entropy, each sub-picture information characterized,The edge feature matrix of definition image, and define image focusing evaluation function based on estimating of pixel, thus determine image measurement systemThe accurate focusing position of system.
Further, described to the method for each sub-picture ambiguity in definition entropy for: establish obtained m × n dimension image and there is NIndividual gray level, ambiguity in definition collection A, the gray level that its domain is image, and the gray value of each pixel in image is carried out to normalizingChange and process, the membership function of ambiguity in definition collection A is:
Wherein, k is the grey scale pixel value after normalized;
On fuzzy set A, ambiguity in definition entropy is:
EA(μA(f(i,j)))＝-(μA(f(i,j)))log(μA(f(i,j)))，
Wherein, in the time of f (i, j)=k, membership function maximum and fuzzy entropy minimum, and fuzzy entropy has symmetry at f (i, j)=kProperty.
Further, the method for the edge feature matrix of described definition image is: with the gray value structure of each pixel in imageBuild m × n and tie up matrix M, in matrix M, centered by capture vegetarian refreshments (i, j), size is a window of l × l, and wherein l isOdd number; On this window, define pixel (i, j) based on fuzzy entropy estimate for:
m A ( i , j ) 1 l × l Σ m = - ( l - 1 ) / 2 ( l - 1 ) / 2 Σ n = - ( l - 1 ) / 2 ( l - 1 ) / 2 E A ( μ A ( f ( i + m , j + n ) ) ) ,
Make f (i, j)=k, in the time that pixel (i, j) is marginal point, in window, the grey value difference of each point is large, mAThe value of (i, j)Also larger; Pixels all in image is calculated respectively and estimates mA(i, j), the edge feature matrix of design of graphics pictureM[mA(i,j)]m×n。
Further, described definition image focusing evaluation function is:
F＝Σ(i,j)∈RectmA(i, j), wherein Rect is focusing window.
Further, calculate the maximum of the evaluation of estimate of each sub-picture obtaining, its defined lens location is mirrorFocusing position.
Further, drive lens moving to the lens focusing position obtaining, focusing finishes.
Further, described two input signal f (i, j) are Zone Full or the regional areas in obtained image.
Further, a kind of image Atomatic focusing method based on fuzzy entropy is surveyed data transmission device, comprises the steps:
A1, automatically focusing operation starts, and the camera lens leg speed of advancing is set to 0;
A2, obtain and preserve the collection image of camera lens current location;
A3, the camera lens leg speed i that advances adds 1, calculates lens location p according to selected step-length si=i × s, preserves lens location pi。
A4, intercept the focusing window of m × n in picture centre;
A5, m × n dimension image has N gray scale collection, ambiguity in definition collection A, and the gray scale collection that its domain is image, and to imageIn the gray value of each pixel be normalized, the membership function of ambiguity in definition collection A is:Wherein, k is the grey scale pixel value after normalized, 0.5≤μA(f (i, j))≤1; FixedJustice fuzzy entropy is: EA(μA(f(i,j)))＝-(μA(f(i,j)))log(μA(f(i,j)))；
A6, build m × n tie up matrix M with the gray value of each pixel in image, centered by capture vegetarian refreshments (i, j), size isA window of l × l, wherein l is odd number; On this window calculating pixel point (i, j) based on fuzzy entropy estimate for:
m A ( i , j ) 1 l × l Σ m = - ( l - 1 ) / 2 ( l - 1 ) / 2 Σ n = - ( l - 1 ) / 2 ( l - 1 ) / 2 E A ( μ A ( f ( i + m , j + n ) ) ) ,Pixels all in image is calculated and estimated respectivelymA(i, j), the edge feature matrix M [m of design of graphics pictureA(i,j)]m×n；
A7, definition focusing evaluation function F=Σ(i,j)∈RectmA(i, j), wherein Rect is focusing window, evaluates letter according to focusingThe evaluation of estimate Q of number computed imagei；
A8, judge whether to travel through whole hunting zone according to current lens location; If reached hunting zoneLarge value, enters steps A 9, otherwise returns to steps A 2;
A9, from gathered picture appraisal value array, find out maximum QmAnd corresponding lens location pm；
A10, calculate current camera lens position pnWith the maximum corresponding lens location p of picture appraisal valuemDistanced＝pm-pn；
A11, d is fed back to drive control part, drive camera lens to focusing position;
A12, focusing finish, so memory cell empties.
Embodiment 2
In focus process, the step-lengths such as camera lens are advanced, and the step pitch of often advancing gathers a sub-picture. The present embodiment has gathered 17The image of secondary " brake block boss ", and be of a size of 320 × 320 in the center of every width image intercepting focusing window. Image when focusingQuality is again to fuzzy from fuzzy to clear, and the evaluation of estimate of focusing accordingly also should present the Changing Pattern reducing again from small to large. Figure2 (a) have provided two different width images of focusing situation with Fig. 2 (b), be respectively fuzzy and the picture of brake block boss collection clearly.Fig. 3 provided for image sequence shown in Fig. 2 based on the classical focusing of Atomatic focusing method of the present invention and some evaluation methodsRelatively, the focusing evaluation algorithms of employing comprises: Roberts operator, Brenner operator, Laplacian operator, TenengradOperator. Fig. 3 is the comparison curves after evaluation result normalization. Fig. 3 shows, in compared focusing evaluation method, only hasFocusing method of the present invention and Brenner operator present unimodality, and focusing method of the present invention presents more precipitous peak featureHigher sensitiveness. Roberts operator, Laplacian operator and Tenengrad operator all present multiple peak points, are difficult toAlign burnt position and make correct evaluation.
The best enlargement ratio of test platform that the present embodiment adopts is 3, and optimum illumination intensity is with this understanding 35. Fig. 4Provided under best enlargement ratio, adopt different illumination conditions based on three kinds of focusing evaluation function curves of the present invention. These are three years oldPlant illumination condition respectively: intensity of illumination is got respectively 10 (on the weak side), 35 (ideals), 80 (excessively strong). Fig. 4 shows, lighting conditionCross strong or cross a little less than all can make focusing evaluation function curve floating, and curve slightly becomes smooth, acuteness weakens to some extent. But curveStill there is good unimodality, the feature of unbiasedness, show the good adaptability to environmental change.
Fig. 5 (a) and Fig. 5 (b) have provided respectively the Atomatic focusing method of the present invention under different illumination conditions and have calculated based on BreeneThe comparison of the Atomatic focusing method of son. Fig. 5 (a) for enlargement ratio be 3, illumination condition is the comparison under 10 (on the weak side) condition; Figure5 (b) for enlargement ratio be 3, illumination condition is the comparison under 80 (excessively strong) condition. As seen from the figure, Brenner operator is in light intensityUnder the environment of undesirable (cross by force or excessively weak), its curve floats, smooth, and there is local peaking's point. Of the present invention automaticAlthough focusing method is also subject to the impact of environment, show the adaptive capacity stronger to environmental change.
In sum, the focusing evaluation function based on fuzzy entropy can be evaluated image definition effectively, thereby determines figureThe accurate focusing position of picture. Meanwhile, there is in actual applications stronger antijamming capability, can adapt to non-contact image and surveyComplex environment in amount process changes, and effectively improves the unfailing performance of autofocus system.
Above-described embodiment is only used for further illustrating a kind of image Atomatic focusing method based on fuzzy entropy of the present invention and surveys data biographyConveying device and method, but the present invention is not limited to embodiment, and every foundation technical spirit of the present invention is done above embodimentAny simple modification, equivalent variations and modification, all fall in the protection domain of technical solution of the present invention.
Claims (8)
1.一种基于模糊熵的影像自动对焦方法测数据传输装置，其特征在于，包括：驱动镜头以特定步长遍历整个搜索区域，在镜头的初试位置以及每一步的位置获取一副图像，对所得到的每一副图像保存其对应的镜头位置，通过模糊熵对每一副图像信息进行表征，定义图像的边缘特征矩阵，并基于像素的测度定义图像对焦评价函数，从而确定图像测量系统的精确对焦位置。1. An image auto-focus method measurement data transmission device based on fuzzy entropy, it is characterized in that, comprising: drive camera lens to traverse whole search area with specific step-length, acquire a pair of image at the initial test position of camera lens and the position of every step, to Each obtained image saves its corresponding lens position, characterizes each image information through fuzzy entropy, defines the edge feature matrix of the image, and defines the image focus evaluation function based on the pixel measurement, so as to determine the image measurement system. Precise focus position. 2.根据权利要求1所述的一种基于模糊熵的影像自动对焦方法测数据传输装置，其特征在于：所述的对每一副图像定义模糊熵的方法为：设所获取的m×n维图像具有N个灰度级，定义模糊集A，其论域为图像的灰度级，并对图像中每个像素的灰度值进行归一化处理，定义模糊集A的隶属度函数为：2. A device for image auto-focusing data transmission based on fuzzy entropy according to claim 1, characterized in that: the method for defining fuzzy entropy for each image is as follows: assume that the acquired m×n A three-dimensional image has N gray levels, define a fuzzy set A, its domain is the gray level of the image, and normalize the gray value of each pixel in the image, define the membership function of the fuzzy set A as : 其中，k为归一化处理后的像素灰度值； Among them, k is the pixel gray value after normalization processing; 在模糊集A上定义模糊熵为：Define fuzzy entropy on fuzzy set A as: EA(μA(f(i,j)))＝-(μA(f(i,j)))log(μA(f(i,j)))，E A (μ A (f(i,j)))=-(μ A (f(i,j)))log(μ A (f(i,j))), 其中，当f(i,j)＝k时，隶属度函数最大而模糊熵最小，且模糊熵在f(i,j)＝k具有对称性。Wherein, when f(i,j)=k, the membership function is the largest and the fuzzy entropy is the smallest, and the fuzzy entropy has symmetry at f(i,j)=k. 3.根据权利要求2所述的一种基于模糊熵的影像自动对焦方法测数据传输装置，其特征在于：所述定义图像的边缘特征矩阵的方法为：以图像中每一个像素的灰度值构建m×n维矩阵M，在矩阵M中，取像素点(i,j)为中心，大小为l×l的一个窗口，其中l为奇数；在该窗口上定义像素点(i,j)的基于模糊熵的测度为：3. The data transmission device based on a fuzzy entropy-based image autofocus method according to claim 2, wherein the method for defining the edge feature matrix of the image is: the gray value of each pixel in the image Construct an m×n-dimensional matrix M. In the matrix M, take the pixel point (i,j) as the center and a window with a size of l×l, where l is an odd number; define the pixel point (i,j) on this window The fuzzy entropy-based measure of is: mm AA (( ii ,, jj )) 11 ll ×× ll ΣΣ mm == -- (( ll -- 11 )) // 22 (( ll -- 11 )) // 22 ΣΣ nno == -- (( ll -- 11 )) // 22 (( ll -- 11 )) // 22 EE. AA (( μμ AA (( ff (( ii ++ mm ,, jj ++ nno )) )) )) ,, 令f(i,j)＝k，当像素点(i,j)为边缘点时，窗口内各点的灰度值差异大，则mA(i,j)的值也较大；对图像中所有的像素点分别计算测度mA(i,j)，构建图像的边缘特征矩阵M[mA(i,j)]m×n。Let f(i,j)=k, when the pixel point (i,j) is an edge point, the gray value difference of each point in the window is large, then the value of m A (i,j) is also large; for the image Calculate the measure m A (i,j) for all the pixels in respectively, and construct the edge feature matrix M[m A (i,j)] m×n of the image. 4.根据权利要求3所述的一种基于模糊熵的影像自动对焦方法测数据传输装置，其特征在于：所述定义图像对焦评价函数为：4. A kind of image auto-focus method measurement data transmission device based on fuzzy entropy according to claim 3, is characterized in that: described defined image focus evaluation function is: F＝Σ(i,j)∈RectmA(i,j)，其中Rect为对焦窗口。F=Σ (i,j)∈Rect m A (i,j), where Rect is the focus window. 5.根据权利要求4所述的一种基于模糊熵的影像自动对焦方法测数据传输装置，其特征在于：计算所获取的每一副图像的评价值的最大值，其所定义的镜头位置即为镜头对焦位置。5. A kind of image auto-focus method measuring data transmission device based on fuzzy entropy according to claim 4, is characterized in that: calculate the maximum value of the evaluation value of each secondary image that is acquired, and the lens position defined by it is Focus position for the lens. 6.根据权利要求5所述的一种基于模糊熵的影像自动对焦方法测数据传输装置，其特征在于：驱动镜头移动至所获得的镜头对焦位置，对焦结束。6 . The data transmission device based on the fuzzy entropy-based image auto-focus method according to claim 5 , wherein the lens is driven to move to the acquired lens focus position, and the focus ends. 7.根据权利要求2所述的一种基于模糊熵的影像自动对焦方法测数据传输装置，其特征在于：所述二位输入信号f(i,j)是所获取图像中的全部区域或者局部区域。7. A fuzzy entropy-based image auto-focus method measurement data transmission device according to claim 2, characterized in that: the two-bit input signal f(i, j) is all or part of the acquired image area. 8.根据权利要求1所述的一种基于模糊熵的影像自动对焦方法测数据传输装置，其特征在于：包括如下步骤：8. A kind of image auto-focus method measurement data transmission device based on fuzzy entropy according to claim 1, it is characterized in that: comprise the steps: A1、自动对焦操作开始，镜头行进步速设置为0；A1. The autofocus operation starts, and the lens travel speed is set to 0; A2、获取并保存镜头当前位置的采集图像；A2. Obtain and save the captured image of the current position of the lens; A3、镜头行进步速i加1，根据所选定的步长s计算镜头位置pi＝i×s，保存镜头位置pi。A3. Add 1 to the lens travel speed i, calculate the lens position p i =i×s according to the selected step size s, and save the lens position p i . A4、在图像中心截取m×n的对焦窗口；A4. Intercept the focus window of m×n in the center of the image; A5、m×n维图像具有N个灰度集，定义模糊集A，其论域为图像的灰度集，并对图像中每个像素的灰度值进行归一化处理，定义模糊集A的隶属度函数为：其中，k为归一化处理后的像素灰度值，0.5≤μA(f(i,j))≤1；定义模糊熵为：EA(μA(f(i,j)))＝-(μA(f(i,j)))log(μA(f(i,j)))；A5. An m×n dimensional image has N grayscale sets. Define a fuzzy set A. The domain of discussion is the grayscale set of the image. Normalize the grayscale value of each pixel in the image to define a fuzzy set A. The membership function of is: Among them, k is the pixel gray value after normalization processing, 0.5≤μA (f(i,j))≤1; define fuzzy entropy as: E A ( μA (f(i,j)))= -( μA (f(i,j)))log( μA (f(i,j))); A6、以图像中每一个像素的灰度值构建m×n维矩阵M，取像素点(i,j)为中心，大小为l×l的一个窗口，其中l为奇数；在该窗口上计算像素点(i,j)的基于模糊熵的测度为： m A ( i , j ) 1 l × l Σ m = - ( l - 1 ) / 2 ( l - 1 ) / 2 Σ n = - ( l - 1 ) / 2 ( l - 1 ) / 2 E A ( μ A ( f ( i + m , j + n ) ) ) , 对图像中所有的像素点分别计算测度mA(i,j)，构建图像的边缘特征矩阵M[mA(i,j)]m×n；A6. Construct an m×n-dimensional matrix M with the gray value of each pixel in the image, take the pixel point (i, j) as the center, and a window with a size of l×l, where l is an odd number; calculate on this window The measure based on fuzzy entropy of pixel point (i, j) is: m A ( i , j ) 1 l × l Σ m = - ( l - 1 ) / 2 ( l - 1 ) / 2 Σ no = - ( l - 1 ) / 2 ( l - 1 ) / 2 E. A ( μ A ( f ( i + m , j + no ) ) ) , Calculate the measure m A (i, j) for all pixels in the image, and construct the edge feature matrix M[m A (i, j)] m×n of the image; A7、定义对焦评价函数F＝Σ(i,j)∈RectmA(i,j)，其中Rect为对焦窗口，根据对焦评价函数计算图像的评价值Qi；A7. Define the focus evaluation function F=Σ (i,j)∈Rect m A (i,j), wherein Rect is the focus window, and calculate the evaluation value Q i of the image according to the focus evaluation function; A8、根据当前镜头位置判断是否已经遍历整个搜索范围；如果已经达到搜索范围的最大值，则进入步骤A9，否则返回步骤A2；A8. Determine whether the entire search range has been traversed according to the current lens position; if the maximum value of the search range has been reached, proceed to step A9, otherwise return to step A2; A9、从所采集的图像评价值数组中找出最大值Qm及其对应的镜头位置pm；A9. Find the maximum value Q m and the corresponding lens position p m from the collected image evaluation value array; A10、计算当前镜头所在位置pn与图像评价值最大所对应镜头位置pm的距离d＝pm-pn；A10. Calculate the distance d=p m −p n between the current lens position p n and the lens position p m corresponding to the maximum image evaluation value; A11、将d反馈给驱动控制部分，驱动镜头至对焦位置；A11. Feedback d to the drive control part to drive the lens to the focus position; A12、对焦结束，所以存储单元清空。A12. The focus is finished, so the storage unit is cleared.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510522454.XA CN105607218A (en) | 2015-08-24 | 2015-08-24 | Image auto-focusing method measurement data transmission device and method based on fuzzy entropy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510522454.XA CN105607218A (en) | 2015-08-24 | 2015-08-24 | Image auto-focusing method measurement data transmission device and method based on fuzzy entropy |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105607218A true CN105607218A (en) | 2016-05-25 |
Family
ID=55987283
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510522454.XA Pending CN105607218A (en) | 2015-08-24 | 2015-08-24 | Image auto-focusing method measurement data transmission device and method based on fuzzy entropy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105607218A (en) |
Cited By (4)
* Cited by examiner, † Cited by third party| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108154510A (en) * | 2018-01-17 | 2018-06-12 | 深圳市亿图视觉自动化技术有限公司 | Method for detecting surface defects of products, device and computer readable storage medium |
| CN109752831A (en) * | 2017-11-02 | 2019-05-14 | 山东颐泽天泰医疗科技有限公司 | A kind of microscope auto focusing method, system and device that focal time is controllable |
| CN110121669A (en) * | 2016-12-30 | 2019-08-13 | 欧洲非线性光谱实验室(Lens) | System and method for measuring the focus state of optical instrument |
| CN114845042A (en) * | 2022-03-14 | 2022-08-02 | 南京大学 | Camera automatic focusing method based on image information entropy |
-
2015
- 2015-08-24 CN CN201510522454.XA patent/CN105607218A/en active Pending
Non-Patent Citations (2)
* Cited by examiner, † Cited by third party| Title |
|---|
| S.AL-SHARHAN ETC: "Fuzzy Entropy: a Brief Survey", 《2001 IEEE INTERNATIONAL FUZZY SYSTEMS CONFERENCE》 * |
| 孙首雷: "高像素CMOS摄像头模组检测技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (6)
* Cited by examiner, † Cited by third party| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110121669A (en) * | 2016-12-30 | 2019-08-13 | 欧洲非线性光谱实验室(Lens) | System and method for measuring the focus state of optical instrument |
| CN109752831A (en) * | 2017-11-02 | 2019-05-14 | 山东颐泽天泰医疗科技有限公司 | A kind of microscope auto focusing method, system and device that focal time is controllable |
| CN109752831B (en) * | 2017-11-02 | 2022-01-14 | 山东颐泽天泰医疗科技有限公司 | Automatic focusing method, system and device for microscope with controllable focusing time |
| CN108154510A (en) * | 2018-01-17 | 2018-06-12 | 深圳市亿图视觉自动化技术有限公司 | Method for detecting surface defects of products, device and computer readable storage medium |
| CN114845042A (en) * | 2022-03-14 | 2022-08-02 | 南京大学 | Camera automatic focusing method based on image information entropy |
| CN114845042B (en) * | 2022-03-14 | 2023-03-10 | 南京大学 | Camera automatic focusing method based on image information entropy |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7697749B2 (en) | 2010-04-13 | Stereo image processing device |
| CN103985108B (en) | 2017-02-15 | Method for multi-focus image fusion through boundary detection and multi-scale morphology definition measurement |
| CN111083365B (en) | 2021-01-15 | Method and device for rapidly detecting optimal focal plane position |
| CN107369159B (en) | 2020-04-24 | Threshold segmentation method based on multi-factor two-dimensional gray level histogram |
| CN103793918A (en) | 2014-05-14 | Image definition detecting method and device |
| CN104637064A (en) | 2015-05-20 | Defocus blurred image definition detection method based on edge intensity weight |
| US20200057886A1 (en) | 2020-02-20 | Gesture recognition method and apparatus, electronic device, and computer-readable storage medium |
| US20040217257A1 (en) | 2004-11-04 | Scene-based method for determining focus |
| CN102113017A (en) | 2011-06-29 | System and method to generate depth data using edge detection |
| CN102404602A (en) | 2012-04-04 | Camera definition detection method based on definition test card |
| CN106488122A (en) | 2017-03-08 | A kind of dynamic auto focusing algorithm based on improved sobel method |
| CN105607218A (en) | 2016-05-25 | Image auto-focusing method measurement data transmission device and method based on fuzzy entropy |
| CN101533513A (en) | 2009-09-16 | Picture falsification testing method based on double-vision geometry |
| CN103049909A (en) | 2013-04-17 | Exposure method taking license plate as focus |
| Yousef et al. | 2014 | Shoreline extraction from light detection and ranging digital elevation model data and aerial images |
| CN103096117B (en) | 2015-07-29 | Video noise detection method and device |
| CN111444913B (en) | 2023-08-22 | License plate real-time detection method based on edge guiding sparse attention mechanism |
| CN106878617B (en) | 2019-05-31 | A kind of focusing method and system |
| CN106027911A (en) | 2016-10-12 | In-orbit focusing method of earth observation satellite-borne visible light transmission camera |
| CN106981065A (en) | 2017-07-25 | A kind of image Absolute Central Moment innovatory algorithm based on exposure compensating |
| Xue et al. | 2021 | Detection of Various Types of Metal Surface Defects Based on Image Processing. |
| CN113538545A (en) | 2021-10-22 | A monocular depth estimation method based on electro-hydraulic adjustable focus lens and corresponding camera and storage medium |
| Love et al. | 2006 | An empirical study of block matching techniques for the detection of moving objects |
| CN110896469B (en) | 2022-05-27 | Resolution testing method for three-shot photography and application thereof |
| CN105141839B (en) | 2018-01-26 | A high-definition image acquisition method based on aperture time control |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 2016-05-25 | C06 | Publication | |
| 2016-05-25 | PB01 | Publication | |
| 2016-06-22 | C10 | Entry into substantive examination | |
| 2016-06-22 | SE01 | Entry into force of request for substantive examination | |
| 2019-10-15 | RJ01 | Rejection of invention patent application after publication |
Application publication date: 20160525 |
| 2019-10-15 | RJ01 | Rejection of invention patent application after publication |