Image brim detects the technology is soldering the application in environmental identifying

G.

Roberts begins a system the earliest to study the brim detects [6] . Can appear every year later a lot of articles that detect about the brim, serious article is published in IEEE Trans mostly.

On Pattern Analysis And Machine Intelligence, CVGIP: Image Processing, IEEE Trans.

On Image Processing, journal Of The ACM go up. Although already had by tens of thousands the brim detects method, but have versatility and extensive adaptability without a kind of method, can take directly use on some kind of specific application, because this exploration already had algorithmic characteristic,discovery suits to solder the algorithm with the algorithm of the characteristic and new development will be used at soldering the process has certain real sense. 1.

Analysis of characteristic of type of welding line brim is in image processing theory [7] , image brim nods corresponding and may different physics meaning. The different physics sense with corresponding basis can assign the margin it is 4 kinds as follows (following graph) : Type of graph image brim (1) 1 kind of brims. If pursue 1 win the bid for the brim line of 1, 1 kind of brims are two different curved surface or cross a line planarly, the normal way that handles object appearance at this o'clock is discontinuous, in the both sides of line of 1 kind of brims, the value has the gray scale of image to get change significantly. (2) 2 kinds of brims. By material the type differs or color difference produces line of 2 kinds of brims. In pursueing on, by two kinds different material is comprised, because they are right the reflection coefficient of light is different, make the two side gray scale of 2 brims line has marked change. (3) 3 kinds of brims. Line of 3 kinds of brims is the boundary of object and setting. There are lines of two 3 kinds of brims on the cylinder in as above graph, the contour line outside be being called commonly. On dot of 3 kinds of brims, the normal direction of three-dimensional object surface is successive, because see an object from proper perspective,occurrence brim dot is when, 3 kinds of border nod the boundary that is object and setting. Because object and setting are in the respect difference such as illumination condition and material reflex coefficient is very big, be in accordingly two side of 3 kinds of brims, the gray scale of image also has marked change. The graph wins the bid the brim with 3 ′ , the boundary of since object and setting, also be the discontinuous place of normal of the surface on the object, but causing its two side gray scale to jump the reason that change is former. (4) 4 kinds of brims. 4 it is the brim that the shadow causes. As a result of part of object surface some by another object keep out, make it cannot get the illuminate of illuminant, cause the brim to nod value of two side gray scale to have bigger difference thereby. To the juncture before solder, no matter be the butt joint welding line, welding line that takes groove perhaps builds received welding line to be behaved,be two clear edges, be received to building and character can regard the jackknife that is the brim. And to soldering bath expression is an apparent brim outline, appearance of pool of corresponding Yu Rong. The environment solders in the meantime relatively complex, for instance the abhorrent sex of weldment surface (color of nick, oxidation, mark, smeary etc) , be like the interference of arc again. Aluminous to resembling such strong reflex workpiece, still put in smooth reflex, welding torch the inverted image problem that wait, make identify unusually difficult. Analyse these characteristics of welding line and bath, can see solder environmental brim includes afore-mentioned 2, 3, 4 kinds of brims. The impact of these elements must consider when choose and putting forward new margin to extract algorithm. 2.

Commonly used brim extraction algorithm and its are soldering before medium applicability is like, environmental identifying is narrated, the means that the brim collects can be divided roughly detect for gradient law (wait like Roberts functor, Prewitt functor and Sobel functor) , 2 rank derivative 0 crossing detect many law, statistic method, small measure detects, ambiguous maths method, still network of mathematical morphology, nerve, brim sheds a law to wait detect method. It the ability when so much algorithm is aimed at specific applied domain only says which are planted is better that the ability when so much algorithm is aimed at specific applied domain only says which are planted. Canny[8] was summed up 1986 mix theoretically before the achievement of practice, put forward the brim to detect Canny3 criterion: ? The fixed position with good? of Mao of Tun of Mei collect Hu still has low to onefold brim repeat answer, gave out their mathematical expression. Union solders below environmental characteristic, the brim with existing analysis extracts algorithm to soldering the environment knows other applicability. 2.

1 differential functor 2.

1.

Gradient of 1 gradient functor is corresponding at first-order derivative, corresponding gradient functor is corresponding at first-order derivative functor. To F(x of a successive function, y) , its are in (X, the gradient that Y) is in defines as follows: Gradient is a vector, its amplitude and phasic it is respectively: This algorithm needs to resemble element position to have consideration to each, such operation amounts are very large, actual middling often uses small pattern plate to use roll indigestion operation is similar, gx and Gy use a pattern plate severally. Commonly used pattern plate is Robert functor, relatively complex commonly used pattern plate has Prewitt functor, Sobel functor, Kirsch functor. And its stand or fall depends on in these a few kinds of functor the structure of noise, if be in,every dot noise is identical, so Prewitt functor is better, if stand by the noise of the brim,be 2 times along the brim, so Sobel functor is better. Such gradient functor considered extent and way two parameter. And soldering in environmental identifying, we hope real time obtains the information of welding line, hope to be able to be forecasted again at the same time below the direction with ongoing welding line. So gradient functor provided a kind of feasible means to us. But gradient functor is sensitive to noise, this to soldering complexly for the environment, can identify a brim of a lot of bogus. Because this uses gradient functor alone,extract the effect to welding line brim not ideal. Can pass do to image first flowing in order to improve a result, meet an a few brims that lean together are flowing drop, and the fixed position that can affect subtense predestined relationship. Coil with these pattern plate the edge that after accumulating, gains may be jump several to nod and crossing is not one is nodded, because this should consider two factors at the same time, the brim just should not be more than threshold values like element, and what should be more than it in the size of the gradient on gradient direction is former the latter with it, restrain what this method says to be not extreme value. 2.

1.

2Functor of Laplacian functor Laplacian is a kind 2 rank differential functor, its state the form is in digital image commonly: Among them S is with F(m, n) is the rally that central adjacent kindles, but 4 or 8 adjacent dot. Can resemble element gray scale in order to serve as the brim directly, also but serve as the brim like element. Gradient functor and Laplacian functor are more sensitive to noise. Average to adjacent territory is being used first before this is doing the brim to extract standard does flowing processing, can use gauss form at the same time 2 dimension are low connect filter to have filter wave to image, make Laplacian brim extraction again next. This is commonly used Laplacian-Gauss functor. Soldering there are a lot of noises to nod on workpiece, and the noise dot that those differential operation is right isolating has " diffuse " action, especially Laplacian functor, not only diffuse and intensity increases significantly. Detect in differential functor so noise had better keep clear of before the brim. And the welding line to the butt joint, when clearance is lesser its brim feature shows for fine linear, the meeting after differential operation broadens. Because gradient functor can detect the direction of image brim, more agree with solder environmental identifying. 2.

Brim of 2 small many measure detects method [the understanding with 9][10] is small OK and informal commutation is: Will primitive signal and small left end of function undertake comparative, beg the similar sex coefficient that gives two function, next function of will small wave is right move the distance of function of a small wave, undertake compare and be computationaled, till complete the operation of whole signal; Get the small wave coefficient below a measure so. Function of will small wave expands, repeat afore-mentioned processes, get the small wave coefficient below a series of measure. The choppy dot in image is the crucial feature when analysing image, it is the brim feature that be interested normally. The brim detects even if find on the metabolic gradient direction of wave coefficient as a child rank jumped choppy point. To detect the brim of big target structure in image is mixed detail feature, investigator advanced the idea that much measure brim detects, detect on large scale namely the big edge that gives a cause, detect on small measure target detail. Relevant theory can refer to document [9][10] . This method is one of heat in current picture processing, have better development prospect. Already document applies his solder in the processing of bath image [10] . To soldering the environment is told, this method has better adaptability, can be opposite workpiece or bath gives a cause from the search on large scale, draw interested detail next. 2.

Method of 3 maths morphology [11] maths morphology is research number the theory of feature of video shape structure and fast and collateral processing technique, it is the goal that achieves structural analysis and diagnostic extraction through video to the target configuration commutation. Mathematical morphology is research target with the configuration feature of image, its main content is to design concept of one a complete set of, commutation and algorithm, with the main feature that will describe picture and basic structure, describe the relationship between the element in image and element, part and part namely. The object in image and image feature depend on directly appearance, the purpose of mathematical morphology is appearance studies in time domain space, so morphology applies to image processing. Configuration operation is corroded mediumly, expand, , shutting is to be based on aggregate operation. In these operation structural element has very crucial effect, it adjusts the geometrical structure that image feature alternates. Operation of configuration of have the aid of can introduce image brim check to measure functor. Mathematical morphology expands mediumly and corrode operation to having very intuitionistic geometry background, the image that they can make be handled is added on certain way thick or decrease thin, the difference of former image and this two kinds of operation can use as all-around brim detects, namely or but the edge that check gives picture. In addition, morphology method is OK still undertake correction through getting used to a method to obtain image margin to place oneself, adjust dimension of structural element window stage by stage, achieve the goal that enhance ambiguous brim effectively and removes noise effect appropriately. 2.

4 inferior check of brim elephantine element calculates a standard afore-mentioned these brim check calculate a standard is undertake on class elephantine element, inferior brim elephantine element detects is to show the elephant near the brim element is decomposed, decide the brim is in accurately thereby. Inferior brim elephantine element detects go to image data map the Hilbert space that 9 parameter are formed, decide brim parameter thereby. Ghosal and Mehrotal put forward to use Zernike quadrature first (Zernike Moments ZMs) will detect inferior brim elephantine element, built ideal to the brim in their algorithm rank jump gray scale model, differ through 3 of computational image rank second ZMs, the 4 parameter map that ideal rank jumps gray scale model arrives in 3 ZMs, these 3 ZMs come to repass computation gives the brim to be in linear parameter, decide the brim truly thereby inferior coordinate of class elephantine element. Li Jinquan [12] had relatively thorough research to ZMs algorithm, the place that points out its are insufficient put forward corresponding improvement algorithmic, apply its in welding line identifying, detected brim has precision tall, refine the brim and interference rejection sex to wait for an advantage by force oneself. 3.

The directional change of welding line of conclusion great majority won't be too acuteness, it is successive linear or curve. Be in local can be being regarded inside little scope is two parallel is linear. Because this is soldering,can pass when environmental identifying search linear method to detect welding line. There already was algorithm in these in, gradient functor can detect the brim of welding line still can is opposite at the same time direction undertakes forecasting, the direction with ongoing welding line still can be forecasted while such real time image are handled, relatively agree with solder environmental identifying. But the gender differs differential functor interference rejection, to soldering complexly the environment is told, cannot take application directly, should try to improve and combine other algorithm. Check of brim of many small measure, morphology calculates a law to wait is this domain study one of heat, its characteristic agrees with to solder complexly environmental identifying, should try to study deep. A few inferior check elephantine element calculates a law to be able to obtain more accurate detecting result, it is to raise image to handle accuracy and solder the effort of the result one of direction. The brim detects methodological without number, have its specific suitable scope, when choose or developing new algorithm, must consider to solder the characteristic of itself. CNC Milling CNC Machining