JPH08159980A - Object surface recognition processing system for external appearance inspecting apparatus - Google Patents

Abstract

PURPOSE: To suitably, accurately display the state of a surface to be inspected by altering the conversion conditions to be the reference of describing the state according to information given corresponding to the properties of an object to be inspected. CONSTITUTION: A state describing unit 70 describes the state to be inspected from a plurality of video signals 150 of a memory 50 based on conversion references of each element. For example, when a soldered part is inspected, a reference conversion element 71 is converted a threshold value being the reference to describe the state when a category or properties to be inspected is varied. A specific ratio information detecting element 72 described different pixel unit or area to be distinguished from the ratio relation of video information level by illuminating conditions. An abnormal transition state detecting element 73 corrects a state code for distinguishing the malfunction when the transition of an angle gradient of a fillet surface is abnormal. A specific distribution state detecting element 74 describes to display the malfunction corresponding to the concentration degree of the level or the area of luminance, and non-definite area correcting element 75 describes the area where the state to be inspected is not definite in a range to be interpolated by estimating.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention uses a plurality of sets of illuminators having different irradiation characteristics for an inspection object, for example, irradiation angles, and a plurality of pieces of image information obtained by sequentially switching the irradiation light on the object. The appearance inspection device of the type that recognizes and determines the appearance state of the inspection target, for example, the mounting position of the electronic component on the printed wiring board or the appearance shape and appearance state of the object having a glossy surface such as a soldered joint. The present invention relates to a recognition processing method in an inspection apparatus that makes a determination, and more specifically, to improvement of a state description method of an inspection target surface based on a plurality of video information.

[0002]

2. Description of the Related Art There are various fields of application of appearance inspection machines, and there are various inspection methods. An example of this is a visual inspection machine for soldering joints of electronic components on a printed circuit board, which has been recently increased in density. In other words, in the process of mounting printed wiring boards used in electronic devices, the quality of the "soldered" part that joins the electronic components mounted on the boards affects product reliability, which improves work efficiency and work quality. Automation of visual inspection of soldering for the purpose of securing
Mechanization is becoming common.

Various types of automatic appearance inspection apparatuses for fine target parts typified by solder joints have already been developed. As an effective means therefor, an inspection method using multistage illumination is used. Japanese Patent Application Laid-Open No. 4-30154
I applied for a patent as 9. The outline will be described below.

FIG. 2 shows a basic system configuration of a multi-stage lighting type solder inspection device. The illumination is performed by the illuminator 20 having a multi-stage structure (in the figure, two stages are annular) having different light projection angles, and the television camera 30 receives the reflected light from the inspection target 10. A plurality of sets of video signals 130 obtained by switching the brightness distribution pattern of the target surface.
Stored in the image storage device 50. The code generation unit 70 generates a code distribution pattern indicating the angle distribution of the target surface based on the stored contents of the plurality of sets, and the recognition processing unit 90 uses this to recognize the three-dimensional shape of the target. Also, a pass / fail judgment is executed.

FIG. 3 shows a series of processing steps in the case where the inspection object is the IC lead soldering. The screen of the TV camera contains multiple leads of ICs,
An inspection window is sequentially set for each of them, and the quality of soldering is judged for each. The setting position of the inspection window is usually set in a form that matches the soldering lands, based on the teaching information performed in advance. A code signal for describing the angle of the target surface is generated by using a plurality of luminance information for each pixel in this inspection window by the method described later, and the three-dimensional shape of the target is grasped from this code distribution pattern. Then, the quality of the target is judged.

In FIG. 2, 60 is an illumination switching device, 40 is a movement table for sequentially moving the object,
Further, reference numeral 80 is a drive device thereof. The recognition processing unit 90, which also has a function as a control unit, sequentially moves the 40 through the table control signal 180 to a position suitable for inspecting each of a plurality of inspection targets, as necessary, and outputs an illumination control signal. 1
The illumination is switched via 60, and the processing operation of FIG. 3 is executed sequentially.

FIG. 4 is a side view showing more concretely the relationship between the optical system and the inspection object in the multi-stage illumination type solder inspection apparatus. The illuminator is an annular illuminator composed of four stages of 20-, 20-, 20-, and 20-, and they are arranged on concentric circles when viewed from above in the drawing. A plurality of television cameras 30 are often used in order to enable inspection according to the target structure, but the vertically arranged 30-a is basically used.

FIG. 5 shows the relationship between the projected light from each illuminator and the solder fillet of the IC lead to be inspected. In FIG. 5, 11 is the cross-sectional shape of the IC lead, and 12 is
Shows a cross section of a "solder fillet" that electrically joins the lead and the land 8 on the printed wiring board 9. Also, each arrow, 21-, 21-, 21-
, And 21- are illuminators 20- and 20-, respectively.
The optical paths of reflected light from-, 20-, and 20- through the object are shown.

Since the illuminators have different projection angles, the reflected light having the highest brightness is made to be received by the television camera at the upper part of the drawing at the different angle portions of the solder fillet to be inspected. That is, when each illuminator is given a slight surface illumination characteristic, the reflected light from each illuminator to the TV camera shows the maximum brightness at a specific fillet angle (hereinafter, this is called the reflection center angle of the illuminator). The brightness decreases as the target surface shifts from this angle.

The characteristic of the solid line in FIG. 6 shows the relationship between the target surface angle and the brightness of the reflected light. The reflection center angle of each illuminator is determined by the structure of the optical system, but in the case of the figure, examples are shown in which they are designed at 5 °, 15, 25 °, and 35 °, respectively.

By utilizing the reflection characteristics of each illuminator as described above, the luminance distribution pattern on the target surface and the angle code distribution pattern based on the luminance distribution pattern are formed by the method shown in FIGS. To do. That is, when the solder fillet 12 shown in FIG. 7A is to be inspected, the inspection window 100 (indicated by a broken line) based on the land 8 is first set.

Next, by switching the illuminator, as shown in FIG.
The video signal shown in is obtained in the inspection window area. S1, S3, S5, and S7 in FIG. 8 are video signal patterns of the television camera corresponding to the illumination of the illuminators 20-, 20-, 20-, and 20-, respectively. The corresponding luminance distribution is shown. In the figure
Indicates that the luminance level is high, and ca indicates that the luminance level is low. It should be noted that one unit 101 of the mesh portion in FIGS. 7 and 8 corresponds to one pixel in the image of the television camera, and its size is determined by the configuration of the optical system.

Each image area in FIG. 8 corresponds to the inspection window 100 in FIG. 7, but the angle code generation processing using the above four types of images is executed for each pixel in this window. When a four-stage illuminator is used, the code is basically generated in 8 to 10 stages with reference to the reflection center angle of each illuminator.

That is, the illuminators 20-2 shown in FIG.
In the vicinity of the reflection center angle of 0-, 20-, or 20-, a region in which only the level of the video signal associated with any illumination is extremely high, for example, the threshold value: St2 in FIG.
For the above areas, code 1, 3, 5, or 7
Based on the luminance ratio of each image for those intermediate angles, and code 0 when only the image signal level of 20-is obtained, 2 on the contrary
If only the video signal level of 0- is obtained,
When all video signal levels are low, for example, when the threshold value: St3 in FIG. 6 or less, an uncertain code 9 is assigned.

There are various methods for determining the code based on the luminance ratio, but one example is shown in FIG. 9 using the relationship between the video signal S5 obtained by the illumination 20- and the video signal S7 obtained by the illumination 20- as a model. Shown in. In FIG. 9, reference numeral 111 is a generation area of the uncertain code 9.

FIG. 10 shows a concrete example of the code pattern generated based on such calculation. FIG. 10A shows an example in which a good fillet is formed.
It shows a pattern distribution in which the angle code gradually increases from the end of the (land) 100 to the tip of the lead 11, and it can be determined that the shape of the solder fillet is good.
Further, the code array on the land center line is cumulatively calculated from the end portion of the inspection land to the tip position of the lead to calculate the fillet cross-sectional shape shown in 13 of FIG. 10, that is, the height and length of the fillet. be able to.

On the other hand, FIG. 10B shows a lead floating defect in which the lead is lifted and the fillet on the land is not joined to the lead. In the code distribution pattern in this case, the code showing a high fillet gradient is distributed around the inspection window 100, and the code showing a flat angle is concentrated in the center of the window. From this, it can be determined that the fillet cross-sectional shape is in the state of 13 in FIG. 10 and that the solder and the lead are not joined to each other.

As described above, the multi-stage illumination type visual inspection apparatus is
Based on a plurality of target image patterns obtained by using illuminators with different projection angles, an angle code pattern distribution that describes the three-dimensional shape of the target surface is generated to accurately determine the shape of the target. Have the ability to perform.

However, since the recent increase in the density of printed circuit boards is remarkable, the number of cases in which the target state cannot be accurately recognized by such a conventional code generation system is increasing. The following are mentioned as such specific examples.

1) Change in soldering material: With the progress of double-sided mounting of a printed circuit board, a soldering material which melts at a lower temperature than before has been widely used. These solder materials often have scattering characteristics rather than specular characteristics as compared with conventional solder materials.

2) Trend of no-cleaning of substrates: For reasons of environmental protection, no-cleaning or use of a cleaning solvent having a weaker cleaning power than in the past has been generalized. Under such conditions, flux often remains on the surface of the solder, which is the object of inspection, but the flux coating generally lowers the level of reflected brightness on the solder surface.

3) Miniaturization of parts: Miniaturization of parts means miniaturization of solder joint lands and increase of influence of relative position error of inspection window with respect to lands.

4) Proximity mounting of components: The proximity of mounting components due to high-density mounting is caused by a light-shielding phenomenon in which the illumination light to the inspection target is blocked by other components or reflected light from other components. This causes an increase in the secondary reflection phenomenon that illuminates the inspection target. When the inspection target is in the above state, the state of the target cannot be accurately described with the conventional code generation method, and thus the inspection result may be erroneous, and means for improving it is required. It had been.

[0024]

SUMMARY OF THE INVENTION The present invention is a method of describing a target surface capable of coping with various disturbances and fluctuations in the above-described inspection target, in other words, a method of appropriately expressing the state of the target surface. -The purpose is to provide a code generation method. Therefore, the problems of the present invention when the mounting state of the electronic component and the soldering quality judgment are modeled are as follows.

1) To propose a target surface description method that makes it easy to distinguish the solder surface from other areas.

2) For the solder surface, regardless of its reflection characteristics, or even when the target surface is affected by a disturbance such as a light-shielding phenomenon or a secondary reflection phenomenon, the angle distribution can be determined by identifying it. Propose a target surface description method that can be grasped more accurately.

3) Propose a method for enabling state description of this portion based on ambient conditions or other information when it is difficult to describe the state locally on the surface to be inspected.

[0028]

In order to solve the above-mentioned problems, the present invention irradiates a plurality of sets of illuminators having different irradiation conditions, for example, irradiation angles and hues, to an inspection object, and the irradiation light of these irradiation objects is passed through the inspection object. The plurality of pieces of image information obtained by receiving the reflected light by the image pickup device set at a predetermined position are stored, and the minimum pixel unit of the image or each specific image unit having a fixed relationship with the minimum pixel unit is stored. A method of describing the state of the target portion corresponding to the pixel unit based on a certain conversion condition from the stored information and recognizing the state of the inspection target, for example, its position or shape from the distribution state of the description result. The state description system used in the appearance inspection apparatus according to claim 1, wherein the conversion condition can be changed by information given according to the property of the inspection object.

FIG. 1 shows the present invention more specifically. In FIG. 1, a plurality of video signals 15 of the storage device 50
The state description unit 70 for describing a target state from 0 based on a predetermined conversion criterion includes a reference conversion element: 71, a singular ratio information detection (correction) element: 72, an abnormal transition state detection (correction) element: 73, and a specification. The distribution state detection (correction) element: 74 and the non-determined area correction element: 75.

[0030]

[Operation] The operation contents of each of these functions are as follows.

In the example of the solder joint, the reference conversion element 71 has a brightness level which varies depending on the solder material, the cleaning / non-cleaning conditions of the substrate, and the like even at the same angle portion of the fillet. different. Depending on the inspection purpose, such a change itself may be detected, but when accurately describing the angle of the target surface, the conversion standard or conversion algorithm from each video information to the description code is changed. There must be. The conversion standard changing element has such a function.

Specifically, when the target category or property changes, the difference is given in the form of some detection information or external designation information to change the state description reference conversion condition or conversion method. This makes it possible to accurately grasp the state of the target surface. For example, an increase in scattered components on the surface to be inspected due to the solder material or a decrease in the target reflection level due to no cleaning commonly occur in each illumination image, so it is detected by comparing the overall average levels of multiple image signals. It is possible to Therefore, when such a condition is detected, the description condition such as the threshold value of the conversion reference for the state description may be changed. If the target material change is known in advance, the detection function may be omitted.

A more important function of the reference conversion element is to have a function of generating uncertain information when the conversion into the state description information, which is the conversion object, is impossible. The impossible condition referred to here is not only that the video information level is low and thus cannot be converted, but it is logically impossible to convert the target state description information from the relationship of a plurality of information levels. Including cases. That is, from the relationship between the angle of the solder fillet and each image information shown in FIG. 6, a possible image level relationship only in the case of solder is derived, and the logical relationship necessary for the above function is set based on this. The state is described by the following elements for the state uncertain portion obtained in this way.

Singular ratio information detection element: 72 is, for example, I
Taking inspection of C part as an example, even if it is the same flat part,
On the land surface to which the solder is attached, the land surface to which the solder is not attached and the copper foil is exposed, or the lead surface, the image information level ratio relationship has a unique value. This is because the target material composition and the surface treatment state are different.

Conversely, when the image information obtained by each illumination is different from the original ratio relationship of the solder shown in FIG. 6, the state of this portion can be distinguished from the solder surface. Therefore, after such a distinction is made, a state description indicating whether the pixel unit or region different from solder is a lead or a land may be made from the above-mentioned unique level ratio relationship.

As described above, the unique ratio information detecting element has a function of discriminating the original inspection object from other regions on the state description and further subdividing the states of the other regions as necessary.
Including the contents of different categories in the state description in this way is an extremely effective means in object recognition.

The abnormal transition state detection element 73 is, for example, in a solder joint portion of electronic parts densely mounted,
A secondary reflection phenomenon from another component may occur on the inclined surface on which the solder fillet is formed. In such a case, a plurality of image relationships in the area are accompanied by a condition inconsistent with the original image level relationship that should correspond to the gradient of the solder surface. When the occurrence of such a condition is not detected by the above-mentioned specific ratio information detection function, the abnormal transition state may be detected.

That is, since the fillet shape of the solder joint is determined by the surface tension and gravity at the melting stage, the slope change of the surface has a certain continuity. Therefore, there is a certain condition for the transition of the angular gradient of the adjacent regions, and it is possible to determine that a transition state that is not this condition is abnormal. In such a case, a state code indicating that it is an abnormal region is added, and if necessary, it is finally determined from other information whether it is secondary reflection or another state, and depending on the case, Correct the status code to distinguish these abnormalities.

Specific distribution state detection element: 74 is, for example, I
When the C lead is lifted, the image of the TV camera of the inspection machine is defocused in this area. A feature of a defocused image is that the variation in the luminance level is small, that is, the luminance signal in the area is almost constant.

Therefore, the frequency distribution of the brightness signal is checked for the entire surface to be inspected or for each area, and the area where the brightness level is substantially constant is checked, and an abnormality is detected according to the level or area-wise concentration. Make the description shown.

Non-determined area correction element: 75 position + interpolation In the case of a non-determined area where the target state cannot be confirmed even by the above-mentioned elements, state description is performed within the range possible by this correction element. One of the means is interpolation, and when the uncertain area is an extremely small area, interpolation is performed by a smoothing method such as averaging based on the surrounding state description contents. This method is effective when a part of the solder fillet is locally contaminated by flux or the like.

The second means is a method of estimating using other information. For example, in IC inspection or the like, when information such as the relative position of a lead or land is already detected, the state of the uncertain area is described by referring to this information. For example, inside the land near the lead,
It can be estimated that all the regions where the luminance signal level is low are the solder portions where the fillet angle gradient is large. On the contrary, if it is outside the land, it can be estimated that it is a flat resist portion.
Such a state description based on estimation is extremely effective when the inspection target range is limited.

The principle of the present invention has been described above with reference to FIG. 1. However, some of the operations of the above elements can be omitted depending on the type of object and the purpose of inspection and the order of execution. It need not be as in FIG.

In addition, although the description result to the target state may be accompanied by noise due to various disturbances, it is normal to perform statistical processing based on the degree of distribution concentration and occurrence frequency, smoothing processing by smoothing or filtering, etc. The state description accuracy may be improved by applying the method used for the recognition processing.

By using the principle of the present invention described above, for example, the quality of the solder joint corresponding to various mounting conditions can be judged more accurately.

The concept of the present invention can also be used as a general means in various inspections other than soldering of printed circuit boards.

[0047]

EXAMPLES The contents of the present invention will be described in more detail below based on examples in which the mounting state of electronic components and the quality of soldering are used as models.

First, an embodiment of the conversion standard changing element 71 will be described with reference to FIGS. 11 to 14. FIG. 11 shows the relationship between the angle of the solder surface having a relatively strong scattering reflection characteristic and the image brightness level. It can be seen that, as compared with the solder having the strong specular reflection characteristics shown in FIG. 6, the brightness level is relatively lowered, and at the same time, the degree of overlap of images by each illumination is increased.

A slight change in the solder material can be dealt with by using the same conversion algorithm and changing at least the conversion coefficient in the conversion from the image information to the state description, in this case, the angle code. However, if the target characteristic changes to this extent, the conversion algorithm itself must be changed. That is, if the conversion method described with reference to FIG. 9 is used for the target characteristic of FIG. 11, the angle code generation result may be unstable. As described above, a conversion method based on which illumination has the highest image level is necessary for a characteristic having a large number of overlapping portions, instead of a threshold reference, and a specific example thereof is shown in FIG. FIG. 13 shows C of FIG.
The section, that is, a region where S5 is higher than other signals is taken as an example.

Therefore, when the inspection machine handles two types of objects shown in FIGS. 6 and 11, the system shown in FIG. 12 is used as the code conversion function. 12, the conversion algorithm of FIG. 9 is used for the conversion 1 portion, and the conversion algorithm of FIG. 13 is used for the conversion 2 portion.

On the other hand, in order to determine whether the object is specular reflection or scattered reflection, it is general to specify from the outside for each work lot of the inspection machine. In contrast,
As a method for automatically making the judgment, it is considered that information sampling is performed from the board positioning marker 16 installed on the printed board 9 shown in FIG. 14 or the soldering land 8 which is a part of the inspection object. . That is, in the first stage of executing the inspection of each component in the board, the luminance distribution of each image in the sampling information is examined, and the target characteristic is determined from the level distribution. In this case, the object to be sampled is a constant gradient part, for example, a flat part,
Alternatively, if the changing portion from the flat portion to the steep slope portion is set, the determination accuracy can be further improved.

Next, the specific ratio information detection (correction) element: 72
An embodiment will be described with reference to FIGS.
FIG. 15 is a luminance level state diagram of each stage illumination image in which solder having a specular reflection characteristic is used as an example. That is, in FIG.
For example, a high brightness level equal to or higher than the threshold value St2 is a, a middle brightness level equal to or higher than the threshold value St3 and lower than St2 is b, St3.
When the following low brightness level is c, the brightness level state of the solder surface is 1 from mode A to J in FIG.
Only 0 states can exist. Therefore, the other brightness level states are regarded as areas other than the solder fillet, and area separation is performed.

Except when the information can be inspected only by information other than the separated areas, the separated areas are given meanings and further converted into a plurality of kinds of state description contents. An example of the method is the state separation based on the unique luminance ratio shown in FIG.

FIG. 16 shows an example of the luminance ratio relationship of S1, S3, S5, and S7 with respect to an example of the background portion other than the solder fillet that is usually found in the mounted printed circuit board. If the non-solder area is detected,
State classification is performed based on this ratio relationship that is measured and stored in advance, and an example of separation of the lead area is shown in FIG. As a method of such state classification, a look-up table method and a conversion function setting method can be considered.

FIGS. 18 and 19 show an example of state separation based on such a unique luminance ratio, using an IC as a model. FIG.
8 is each illumination image S1, S3, S5 in the inspection window,
And the brightness levels of S7. From this distribution, the state description based on the state diagram of FIG. 15 yields FIG. 19 (1). In this figure, the numerical values are A in the state diagram of FIG.
The pixel portions corresponding to the modes A to J are represented in the form of 0 to 9, and X and Y indicate the case not corresponding to the modes A to J. Of these, the Y area is determined to be in a state corresponding to the lead 11 from the characteristics of FIG. 16, and the X portion is locally generated, so that the result is interpolated by the surrounding codes. Finally, the state code distribution shown in FIG. 19B is obtained, and the existence of a good quality solder fillet without lead bending can be grasped.

Next, abnormal transition state detection (correction) element: 73
An embodiment will be described with reference to FIGS. 20 to 23.

FIG. 20 is a state transition diagram of the solder fillet surface based on the luminance distribution shown in FIG. 6 and the luminance level state shown in FIG. 15. When the target surface is the solder joint area including the flat portion, The code transition in the area is limited to the condition of FIG. However, repeated transitions in the same mode are omitted. Also, the solid line in the figure represents a normal transition, and the broken line represents a transition with a low occurrence frequency. The chain line indicates a transition whose continuity is limited to a certain range.

On the contrary, the state transition occurring on the target surface is shown in FIG.
In all other cases, it must be regarded as an abnormal state. Such an abnormal state, that is, a non-linear state change, implies the occurrence of the above-described secondary reflection phenomenon in addition to the boundary position of a foreign object such as a land end and a lead end.

The detection of the secondary reflection state among them is performed by the method shown in FIG. 21, for example. That is, since the secondary reflection phenomenon usually occurs in the central portion of the solder fillet having a relatively high gradient and is particularly based on the illumination 20-, the mode B, C, or D in FIG. Start under the conditions that occur above,
After checking the presence or absence of a state transition abnormality in the secondary direction from the outer peripheral area of these mode areas according to the contents of FIG. 20,
It is determined whether or not it is in an abnormal state.

22 to 23 show state description examples of the secondary reflection generating fillet by such a method.

From the luminance distribution of FIG. 22, first, FIG.
The description state is obtained. The parts indicated by the codes from A to J are the codes corresponding to the solder in FIG.
Y indicates a lead, as in FIG. On the other hand, the boundary positions x1 and x of the areas B and C in the center of the fillet
When the state transition of each column from 2y1 and y2 in the vertical and horizontal directions is examined, it is determined that there is an abnormality because the transition to each mode via the uncertain portion J is not included in FIG. )
State code Z. The area indicated by Z may be handled in accordance with the linked inspection algorithm.

Next, the specific distribution state detection (correction) element: 74
24 will be described with reference to FIG.

The specific distribution state means a peculiarity of the brightness distribution such that the brightness distribution of each pixel unit in the target area is concentrated or limited to a range set by a predetermined threshold value. .

FIG. 24 shows only one example of extracting the state where the brightness is extremely concentrated. Depending on the purpose, the range from the central value of the distribution to the predetermined range or the predetermined number of pixels is detected. Methods and methods of calculating the degree of dispersion can be used.

In the case of FIG. 24, the pixel having the highest brightness level in the target area is described by a specific status code, and the target status is determined from the distribution status. One of the more specific purposes of this method is, for example, detection of lead floating of an IC. That is, since the brightness of the lead that has floated and caused defocusing shows a concentrated distribution, the detection target area is set near the lead, or the position where the position is extracted imagewise by the method already described. If the pixel distribution is set in the negative area and the distribution of the pixels in the state corresponding to the brightness level concentrated distribution characteristic in this range is checked, it is possible to detect the lead floating based on the positional concentration degree and the occurrence ratio.

Another example of detecting the specific distribution state is a state description in which the position and shape of the IC package can be easily recognized. For example, consider a case where there is an IC package in the inspection window as shown in FIG. 25A and it is necessary to recognize its exact position and partial shape.

Under normal inspection conditions, the package direction in the window is given as information, so the specific distribution state detection area 102 is set at a position corresponding to this, and the luminance distribution is examined. As the lighting image to use,
Select an appropriate one for package material detection. Package
Since the upper surface 10-a of the image shows a luminance distribution concentrated in a certain range except for the character printing portion, the upper and lower thresholds of the luminance level representing the package are determined on the basis of this luminance distribution. By binarizing the window with this threshold value, a state code distribution in which only the upper surface of the package is closed up is obtained as shown by the broken line in FIG. Based on this, it becomes possible to recognize the exact position of the object and the shape of the missing state. Furthermore, based on this detection result,
By setting the specific distribution state detection region: 102 in the package taper portion as shown in (3), the state code distribution obtained by closing up only the taper portion: 10-b becomes (4 ), And using this state description, IC
It is possible to grasp the position and shape as a whole.

Further, an embodiment of the non-determined area correction element: 75 will be described with reference to FIGS.

FIG. 26 shows the image level distribution when the solder fillet slope is relatively steep in the IC. In such a case, the part where the gradient is particularly large is shown in FIG.
In the uncertain mode J in the state diagram of FIG.
(The state description result of 1 is obtained. In this case, since the lead portion is given by the state code Y and the uncertain portion is close to this, it is estimated that this is a steep gradient portion. .
By combining these results and converting each pixel into an angle code, FIG. 27B is obtained.

[0070]

As described above, the embodiments of the target surface state description method according to the present invention have been described. As can be seen from the above, according to the present invention, conventional inspection such as soldering inspection of a high-density mounting board is difficult. The inspection of the existing object can be performed with higher accuracy, and the effect seen industrially is great. In each of the embodiments, the mounting state of the electronic components on the printed circuit board and the quality judgment of soldering are used as a model, and the description is given in the form of application to the appearance inspection, but the scope and application of the principle of the present invention The subject is not limited to these.

Further, although the state description is described in the form of a code, this is for the purpose of helping understanding of the explanation, and inside the actual processing device, it may be different depending on the convenience of the processing device. The expression method may be used. further,
A variety of processing forms are conceivable, such as when various state descriptions are made in a complex form for the same pixel unit.

[Brief description of drawings]

FIG. 1 is a configuration block diagram of a target surface state description element according to the present invention.

FIG. 2 is a basic configuration block diagram of a conventional inspection device.

FIG. 3 is a flowchart of basic operation contents of a conventional inspection device.

FIG. 4 is an external view of a basic configuration of an optical system of a conventional inspection device.

FIG. 5 is an explanatory diagram of an operation principle of an optical system of a conventional inspection device.

FIG. 6 is a diagram showing a relationship between an inspection target angle and a detected image level in the multi-stage illumination inspection machine.

FIG. 7 is a stereoscopic view of a state of each pixel in a multi-stage lighting inspection machine.

FIG. 8: Inspection window and image level distribution diagram in multi-stage lighting inspection machine

FIG. 9 is a diagram showing a conversion method from a multi-stage illumination image level to an angle code.

FIG. 10 is a diagram showing the relationship between the inspection object shape and the angle code distribution.

FIG. 11 is another relationship diagram between the inspection target angle and the multi-stage illumination image level.

FIG. 12 is a state description system diagram of an embodiment of the present invention.

FIG. 13 is a diagram of another conversion method from multi-stage illumination image level to angle code.

FIG. 14 is an explanatory diagram of state description assisting means for explaining an embodiment of the present invention.

[Fig. 15] State diagram of multi-stage illumination video level

FIG. 16 is an explanatory diagram showing an example of a target having a peculiar multi-stage illumination video level relationship.

FIG. 17 is a flowchart of a state description method for explaining an embodiment of the present invention.

FIG. 18 is a multi-stage illumination image distribution map of an inspection target for explaining an embodiment of the present invention.

FIG. 19 is an explanatory diagram of a state description result explaining an embodiment of the present invention.

FIG. 20 is an explanatory diagram showing a state transition of a multi-stage illumination image level for explaining an embodiment of the present invention.

FIG. 21 is a flowchart of a state description method for explaining an embodiment of the present invention.

FIG. 22 is a multi-stage illumination image distribution diagram of an inspection target for explaining an embodiment of the present invention.

FIG. 23 is an explanatory diagram of the above state description result for explaining an embodiment of the present invention.

FIG. 24 is a state description flowchart illustrating an embodiment of the present invention.

FIG. 25 is an explanatory diagram of a state description result example for explaining the embodiment of the present invention.

FIG. 26 is a multi-stage illumination image distribution diagram of an inspection target for explaining an embodiment of the present invention.

FIG. 27 is an explanatory diagram of a state description result explaining an embodiment of the present invention.

[Explanation of symbols]

9 printed circuit board, 10 parts to be inspected, 11 IC lead, 12 solder fillet, 20 lighting device, 30
TV camera, 40 moving table of parts to be inspected, 5
0 storage device, 60 illumination switching device, 70 state description unit, 80 moving table control device, 90 processing unit such as recognition, 100 inspection window, 101 pixel unit, 10
2 detection areas, 103 detection areas

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Fumihiko Nagura 32 Miyuki-cho, Kodaira-shi, Tokyo Inside the Koganei factory, Hitachi Electronics Co., Ltd.

Claims (4)

[Claims] 1. An irradiation condition for an inspection target, for example, a plurality of sets of illuminators having different irradiation angles and hues are irradiated, and reflected light of these irradiation lights through the inspection target is received by an imaging device set at a predetermined position. By storing a plurality of video information obtained by the above, a conversion element having a predetermined conversion condition from the stored information is stored in the minimum pixel unit of the video or for each specific image unit having a fixed relationship with the minimum pixel unit. The state description method used in the appearance inspection apparatus in which the state of the target portion corresponding to the pixel unit is described via the state of the inspection object and the state of the inspection object is recognized from the distribution state of the description information in the inspection object range. The target surface recognition processing method of the appearance inspection apparatus, wherein the description information is composed of a plurality of types of categories having different meanings. 2. The description according to claim 1, wherein among the description information of a plurality of types of categories having different meanings, one or a plurality of types of main categories that directly correspond to the inspection purpose.
After dividing the pixel unit belonging to the main category and the pixel unit other than that by the conversion element for generating the description information, the description information belonging to one or a plurality of types of sub-categories is classified by other conversion elements for the latter. A target surface recognition processing method for an appearance inspection device characterized by generating. 3. The appearance inspection according to claim 2, wherein the conversion element for converting into the description information belonging to the sub-category is performed based on a specific ratio of a plurality of video information set in the storage device. Target surface recognition processing method of the device. 4. The state description content according to claim 1 or 2, wherein the state description contents of a part of the state description information pattern can be changed based on the positional relationship of the state description information generated in the state description information pattern. The target surface recognition processing method of the appearance inspection device characterized by the above.