JPH0553225B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to a surface defect inspection method for inspecting surface defects of an object to be inspected, such as a painted surface of an automobile body panel.

[Conventional technology]

Inspection of defects such as scratches, protrusions, lumps, dirt, etc. on the surface of objects to be inspected, such as painted surfaces of car body panels, has changed from visual inspection by inefficient and individual workers to automatic inspection using highly directional laser light. There is a tendency to shift to inspection.

As shown in FIG. 14, a surface defect inspection device using such a laser beam projects a slit light LST from a laser slit light generator 1, which is a light source, onto a surface to be inspected 2, which is the surface of an object to be inspected. hand,
The reflected light LST' is projected onto a screen 3 formed by a diffusion plate, and the image on the screen 3 is focused by a condensing lens 4 of a camera section and formed onto a line sensor 5 such as a CCD. The present applicant has previously proposed a device that detects surface defects on the surface to be inspected 2 by detecting them (Japanese Patent Application No.
-83889, patent application No. 59-83890).

According to this surface defect inspection device, for example, the 15th
As shown in the figure, if there is a defect a such as a scratch or a lump on the slit light projection position on the surface to be inspected 2, the slit light LST projected from the laser slit light generator 1
is scattered there, so that on the screen 3 that receives the specularly reflected light LST' from the surface to be inspected 2, there is a slit image cut at the location corresponding to defect a, as shown in the figure.
SL will be projected.

Therefore, the slit image on this screen 3
The 14th station is constantly photographing the position where the SL is projected.
The video output signal of the line sensor 5 in the figure is the 16th
The result is as shown in Figure a, and the envelope signal Vs shown in Figure b and the comparison signal Vr that is integrated and averaged.
(threshold level)
If a binary signal is formed that becomes high level "H" only when Vs < Vr and becomes low level "L" otherwise, as shown in Figure 2, defect a on the surface to be inspected 2 can be The defect can be detected when the level of the defect signal becomes "H", and the size of the defect can also be known from the width W of this pulse.

In this case, the line sensor 5 shown in FIG. 14 captures the projected slit image SL on the screen 3 by receiving the scattered light SC diffused by the screen 3 placed at the focal position of the condenser lens 4. Therefore, the focus of the condensing lens 4 can be fixed at all times, and the projection slit image SL can be adjusted by diffusion.
Since the area is enlarged, the detection resolution of surface defects is improved.

In order to inspect a large surface area such as a painted surface of a car body panel using such a surface defect inspection device,
The detection head of this device is attached to a robot, for example, and the robot is operated according to scanning position data taught in advance, so that the detection head is always at a constant distance from the surface to be inspected and the incident angle of the slit light is also constant. You can scan it so that

In that case, the relative positional relationship between the laser slit light generator 1, the screen 3, and the surface to be inspected 2 is adjusted to the state shown by the solid line in FIG. 17, for example, so that the reflected light from the surface to be inspected 2 of the slit light LST The slit image SL is formed within a range 3a that can be imaged by the line sensor on the screen 3 (hereinafter referred to as "detectable range").

However, if the surface to be inspected 2 is tilted (inclination angles α and β) as shown by the broken line in the figure with respect to the normal state shown by the solid line during scanning, the slit light
A slit image is formed on the screen 3 as the incident angle and reflection position of the LST on the surface to be inspected changes.
SL' deviates from the slit image SL in the normal state and tilts two-dimensionally, leaving the detectable range 3a, making inspection impossible.

Therefore, it is necessary to perform correction so as to always maintain a normal relative positional relationship with a certain degree of accuracy.

Therefore, a plurality of line sensors for position correction are arranged in the direction intersecting this slit image SL or SL', and the amount of deviation and inclination (rotation) angle of the slit image are detected from the difference in the light receiving points, that is, the intersection positions. ,
The position of the inspection device with respect to the surface to be inspected 2 is corrected based on the detection result.

By the way, when the detection head of such a surface defect inspection device is attached to a robot and the surface of an object to be inspected is scanned for surface defects, the object to be inspected is present at a preset inspection position in a predetermined posture. Otherwise, even if the robot is operated according to the scan position data taught in advance, the detection head will not be able to scan the surface of the object to be inspected while always maintaining a constant distance, making it impossible to detect surface defects. There is a fear.

Therefore, in the past, multiple video cameras were installed on the robot's hand or fixed part to detect the posture of the object by capturing the movement of the feature points of the object. The data for the detection head scan, which had been taught in advance, was corrected.

[Problem that the invention seeks to solve]

However, with such surface defect inspection methods, not only does the equipment cost for detecting the posture of the inspected object increase, but there are also restrictions on the detection locations using the video camera, and moreover, when performing surface defect inspection using a surface defect inspection device, However, there was a problem in that this video camera obstructed the robot's work and operating range.

This invention aims to solve these problems.

[Means for solving problems]

Therefore, in the present invention, prior to the start of surface defect inspection by the surface defect inspection apparatus as described above, the surface defect inspection apparatus is used to detect a reference point at a certain distance from the detection head at a plurality of preset locations on the object to be inspected. The attitude of the object to be inspected is detected by measuring the amount of deviation relative to the position, and based on the attitude detection result, the data for scanning the detection head that is taught to the robot in advance is corrected. .

[Effect]

There is no need to install a dedicated video camera etc. to detect the posture of the inspected object, and the detection head of the surface defect inspection device can also be used to detect the posture of the inspected object.
The amount of deviation from the reference position at a certain distance from the detection head is measured at multiple preset locations on the object to be inspected, the posture of the object to be inspected is detected, and the detection head is scanned as previously taught to the robot. Since the data is corrected, there is no extra equipment cost, the measurement position can be set arbitrarily, and there are no obstacles during surface defect inspection work.

〔Example〕

Embodiments of the present invention will be described below based on the drawings.

First, an outline of an apparatus for implementing the surface defect inspection method according to the present invention will be explained with reference to FIG.

Reference numeral 6 denotes a surface defect inspection device, which includes a detection head 7 having a laser slit light generator 10 and a detector 20 equipped with a screen, a condensing lens, a plurality of line sensors, etc., and drives each of the line sensors to generate a video signal. Sensor receiver 3 that reads and binarizes
0 and FIG. 16c from this sensor receiver 30.
The defect detection circuit 31 detects a defect by comparing the pulse width of the binary pulse signal with the detected defect width as shown in FIG. Inspection surface 2 which is the surface of the inspection object (hereinafter referred to as "work") W
Posture correction circuit 3 for correcting the angle etc.
It is composed of 2.

As shown in FIG. 2, for example, the detection head 7 is attached to the tip of the arm of a robot 8, and during surface defect detection work, the robot 8 operates based on scanning position data taught in advance. , the workpiece W whose detection head 7 is the surface to be inspected
12 along the surface 12 at regular intervals and angles to scan the entire surface.

Reference numeral 33 denotes a microcomputer unit (hereinafter referred to as "CPU unit") for correcting traveling data, which is composed of a sensor interface 34, a central processing unit 35, and a robot interface 36.

By this CPU unit 33, before the surface defect inspection device 6 starts inspecting surface defects,
By sequentially positioning the detection head 7 relative to a plurality of preset locations on the workpiece W and processing the detection signal from the sensor receiver 30 at each position, the workpiece can be positioned relative to a reference position at a fixed distance from the preset detection head 7. The attitude of the workpiece W is detected by measuring the amount of positional deviation, and based on the attitude detection result, correction data for correcting the detection head scanning data taught to the control device of the robot 8 in advance is generated. It is calculated and transferred to the robot 8 (including the control device).

Note that this CPU unit 33 can also be used as a microcomputer that constitutes the defect detection circuit 31 and the posture correction circuit 32.

In FIG. 3, when the workpiece W is the body of an automobile, examples of positional deviation measurement points for detecting the posture of the workpiece W are indicated by . Note that the y direction is the traveling direction of the conveyor that conveys this work W, the x direction is the conveyor width direction (vehicle width direction of the work W), and the z direction is
The direction is perpendicular to the xy plane.

Here, data for sequentially moving the detection head 7 to predetermined positions opposite to the measurement points .about. in FIG. 3 is stored in advance in the memory of the robot 8 by teaching.

Then, when the workpiece W is carried to a surface defect inspection position by a conveyor (not shown), the CPU unit 33 of FIG. 1 operates according to the flowchart shown in FIG. The posture is detected and the correction data is sequentially transferred to the robot 8, and the data for scanning the detection head during surface defect inspection, which has been previously taught to the robot 8 and stored in the memory of its control device, is corrected.

To briefly explain this operation, first, the amount of deviation ΔD ₁ in the x direction from the reference position of the workpiece W at the measurement point shown in Fig. 3 is measured (distance measurement), and then the deviation in the x direction at the measurement point is also measured. Measure the quantity ΔD ₂ .

Next, the correction data Δx in the x direction is calculated using either ΔD ₁ or ΔD ₂ , and the rotation angle Δθz of the workpiece W in the y direction about the z direction is calculated from ΔD ₁ and ΔD ₂ . .

Then, the correction data of Δx and Δθz is transferred to the robot 8, the detection head position data for the measurement point that has been taught in advance is corrected, the detection head 7 is moved to the measurement position, and the reference position of the workpiece W is Measure the amount of deviation ΔD ₃ in the y direction from . Correction data Δy in the y direction in FIG. 3 is calculated from this ΔD ₃ .

Next, each correction data of Δx, Δθz and this Δy calculated earlier is transferred to the robot 8, and the detection head position data for the vertical measurement point that has been taught in advance is corrected, and the detection head 7 is measured. z from the reference position of the workpiece W.
Measure the amount of deviation in the direction. From this ΔD ₄ , the correction data Δx in the z direction in FIG. 3 is calculated.

Then, the previously calculated Δx, Δθz, Δy and this Δz
After transferring each correction data to the robot 8 and comprehensively correcting the detection head scanning data previously taught to the robot 8, the robot 8 is made to start scanning work for surface defect inspection.

Note that the distance measurement method for measuring the positional deviation at each measurement location and the calculation method of the rotation angle Δθz will be explained later.

Next, a specific structural example of the detection head of the surface defect inspection apparatus used in the present invention will be explained with reference to FIGS. 5 to 8.

5 is a plan view of the detection head, FIG. 6 is a front view, FIG. 7 is a right side view, and FIG. 8 is a schematic diagram showing the arrangement of an optical system and a line sensor inside the detector.

The laser slit light generator 10 in the detection head 7 includes a laser oscillation tube 10a, a lens system 10b including a cylindrical lens for converting the laser beam generated by the tube into a slit light, and a lens system 10b that prevents the slit light from spreading in the longitudinal direction. It consists of a hood part 10c equipped with a Fresnel lens 10d for producing parallel slit light at its tip.

On the other hand, the detector 20 includes a condensing lens 14, a line sensor 15 for defect detection, line sensors 16 and 17 for position correction, and a beam splitter 40 shown in FIG. The detection tube 20 is integrally attached to the front surface of the section 20a.
A screen 13 made of a diffuser plate such as ground glass is attached to the tip of b.

In addition, in order to make the detector 20 smaller and to increase the reduction magnification of the slit image, the detection tube 20b is made into a bent shape, and two mirrors (plane mirrors) 18 and 19 are arranged inside.
The slit image projected on the mirrors 18 and 19
line sensor 1 by means of condensing lens 14.
5 and line sensors 16 and 17.

This detector 20 is fixed to the base end of a hawk-shaped frame 21, and the laser slit light generator 10 is supported at the tip of the frame 21 by a pivot 22, and the optical axis _O1 of the detector 20 is fixed to the base end of a hawk-shaped frame 21. The intersection angle δ of the optical axis O ₂ of the laser slit light generator 10 with respect to the laser slit light generator 10 can be adjusted.

Frame 2 of this laser slit light generator 10
1 is adjusted by a first adjustment mechanism 24 driven by a motor 23 in FIG.
The adjustment is made by rotating in the direction of arrow θx in FIG.

The frame 21 also has a bracket 25 for attaching the detection head 7 to the tip of the arm of the robot 8 as shown in FIG. 2, and a second adjustment mechanism (see FIG. 6) driven by a motor 26. 27, it is attached so as to be swingable in the direction of arrow θy.

When the frame 21 swings in the direction of the arrow θy, the laser slit light generator 10 and the detector 20 attached to the frame 21 move together, so that the optical axes O ₁ , O as shown in FIG. The inclination angle ε of the surface formed by ₂ with respect to the surface to be inspected 2 changes, and is normally adjusted so that ε=90°.

In addition, 28 in FIG. 6 is a defect marker, which is a defect that does not interfere with the laser slit light when a defect such as a bump or scratch is detected at a position on the surface to be inspected 2 that is irradiated with the laser slit light LST. This is a device for spraying a wipeable paint such as a compound in the vicinity to mark the presence of a defect, and is attached to a bracket 29 integrated with the detector 20 so as to be able to swing at a predetermined angle.

Here, the line sensors 15, 1 in the detector 20
The arrangement relationship between 6 and 17 will be explained with reference to FIG.

A line sensor 15 for defect detection has a detection line on the imaging reference line R of the slit image SL projected on the screen 13 on a sensor mounting plate 41 provided behind the beam splitter 40 and parallel to the condensing lens 14. are arranged to match.

Note that if a PCD (Plasma Coupled Device) line sensor, which has a much wider detection width than a CCD line sensor, is used as the line sensor 15 for defect detection, the relative positional relationship between the inspection device and the surface to be inspected can be slightly changed. Variations are allowed.

On the other hand, line sensors 16 and 17 for position correction
is the imaging reference of the slit image SL reflected by the half mirror surface of the beam splitter 40 and formed on the sensor mounting plate 42 which is provided at a right angle in front of the sensor mounting plate 41. The detection lines are perpendicular to the line R' at the center and are arranged parallel to each other at a predetermined interval.

As the line sensors 16 and 17 for position correction, CCD or PDA line sensors are used.

According to this surface defect detection device, the reflected light of the laser slit light LST projected onto the inspected surface 2 of the workpiece W by the laser slit light generator 10
A slit image SL formed by projecting LST' onto a screen 13 is focused by a condensing lens 14 onto a detection line of a line sensor 15 via a beam splitter 40.

Therefore, as in the case of the conventional example explained with reference to FIGS. 14 to 16, the presence or absence of defects can be detected by binarizing the video output signal read out from the line sensor 15. can.
Defect detection circuit 31 of FIG. 1 does this.

By the way, when this detection head 7 is attached to a robot to detect surface defects while scanning the surface 2 to be inspected, the laser slit light generator 10 and the detector 2
0 and the surface to be inspected 2 changes,
If the angle δ in FIG. 6 and the angle ε in FIG. It deviates from the range and cannot be detected.

In that case, the slit images intersecting the line sensors 16 and 17 are also shifted and tilted with respect to the imaging reference line R', so there is a difference between the slit image detection positions of each line sensor 16 and 17, and the average position and the detection line. The deviation amount of angles ε and δ is detected based on the difference from the center position (1024th bit when using a 2048-bit line sensor), and the first
By driving the motor 23 of the second adjustment mechanism 24 and the motor 26 of the second adjustment mechanism 27 to correct the posture, surface defects are always detected reliably. Posture correction circuit 32 of FIG. 1 does this.

Next, the principle of distance measurement for measuring the amount of deviation from the reference position at each measurement point of the workpiece W using the detection head 7 will be explained.

Distance Measurement Principle () First, the basic distance measurement principle will be explained with reference to FIG.

When performing distance measurement, for example, the angle δ in Figure 6
= 60° and the angle ε = 90° in Figure 7.
When the second adjustment mechanisms 24 and 27 are fixed and the surface to be inspected 2 of the workpiece W is correctly placed at the reference position, the laser slit light projected from the laser slit light generator 10 is reflected from the surface to be measured 2. The slit image on the screen 13 is projected by the slit light LST' onto the imaging reference line R of the defect detection line sensor 15 and the pair of position correction line sensors 1 shown in FIG.
6, 17 imaging reference line R' (line sensor 1
6 and 17)).

Then, the slit image obtained by one of the position correction line sensors 16 and 17 (line sensor 16 in FIG. 9) when actually measuring the surface 2 to be inspected of the workpiece W (shown by the thick line in FIG. 9). Based on the amount of deviation of the detected position _z2 from the center position _z1 , the amount of deviation ΔD of the surface to be inspected 2 from the reference position is determined by the following equation.

ΔD=k(z ₁ −z ₂ ) …(1) Here, if δ/2=β and the resolution of the line sensor 16 including the magnification of the condenser lens 14 is α (mm/bit), then k=α/ 2tanβ. Therefore, equation (1) becomes as follows.

ΔD=(z ₁ − z ₂ )×α/2tanβ …(2) Practical example of distance measurement When actually performing distance measurement according to the distance measurement principle () above, the reference position (teaching point) of the surface to be inspected 2 If the amount of deviation ΔD from the position relative to the line sensor 16 is large, the position where the slit image is formed by the reflective slit hole LST' will be outside the detectable range of the line sensor 16 and cannot be measured.

Therefore, after the detection head 7 has been returned to its original position, all correction mechanisms are disabled and the detection head 7 is in a fixed state.
The robot is operated to position the robot at the approach point shown in FIG. 10, and then approaches the workpiece W step by step at a constant pitch as shown by the arrows.

In FIG. 10, the first stage feed completion position is indicated by a broken line.

Then, when the position correction line sensor 16 or 17 of the detection head 7 grasps the slit image and becomes capable of distance measurement, further constant pitch feeding is stopped,
The number of feed stages at that time and the line sensor 16 or 17
The amount of deviation ΔD of the surface to be inspected 2 from the reference position is determined from the value calculated from the reading (z ₂ ) using the above-mentioned equation (2). In this way, the range of distance measurement can be greatly expanded.

Note that the approach point is set far enough away so that even if the workpiece W deviates the most in the direction toward the detection head, the two will not collide.

An example of the operation of the robot 8 and the CPU unit 33 in FIG. 1 when implementing this distance measuring method is shown in a flowchart in FIG.

Here, "limit over" means that the slit image is outside the range that can be detected by the line sensors 16 and 17, and in that case, the feed amount is increased sequentially until the number of feeds is 5 to bring the detection head closer to the workpiece W. However, if the number of times exceeds 5, an attitude error is displayed and distance measurement is stopped.

Distance Measurement Principle () Next, another distance measurement principle in which the projection angle of the slit light LST by the laser slit light generator 10 is changed will be explained with reference to FIG.

In this method, when the surface to be inspected 2 is at the reference position and the detection head is positioned at the teaching point shown in the figure, the angle of inclination of the optical axis of the laser slit light generator 10 with respect to the direction perpendicular to the surface to be inspected 2 is θ.
Then, the reflected slit light LST' from the surface to be inspected 2 is projected onto the center of the screen 13, and the image is formed through the center of the line sensor 16, and when the surface to be inspected 2 is actually measured. Assume that by adjusting the above-mentioned inclination angle to θ' (using the first adjustment mechanism 24 shown in FIGS. 5 to 7), the line sensor 16 can now detect the slit image at its center bit. .

Then, the distance from the reference position to the center of the screen 13 is H ₀ , the distance to the center of the rotation axis of the laser slit light generator 10 is H ₁ , and the distance between the perpendicular lines drawn from both points to the surface to be inspected 2 is calculated. Then, there is a relationship as follows: = tan θ'(H ₀ +H ₁ +2ΔD), and from this, the amount of deviation ΔD of the surface to be inspected 2 from the reference position can be determined by the following equation.

ΔD=/2tanθ′−(H ₀ −H ₁ )/2 …(3) Distance measurement principle () Each of the above-mentioned distance measurement principles is based on This is related to the distance measurement method when the workpiece W is shifted in parallel in the x, y, and z directions in Figure 3.
A distance measurement method for a field in which the rotation direction is also shifted and a method for calculating the rotation angle will be explained with reference to FIG.

In this example, when the workpiece W is rotated around the z-axis by an angle θ (completely different from θ in FIG. 12) with respect to the y-direction in FIG. This is a method of determining the positional deviation amounts ΔD ₁ and ΔD ₂ and the above-mentioned angle θ.

In Fig. 13, when the surface to be inspected 2 is at the reference position or a position parallel to it, the angle formed by the laser slit light and its reflected slit light is 2β (2β = δ), and the distance in the x direction from the measurement point is L. (mm), and the distance between the reference position of the surface to be inspected 2 and the line sensor 16 (in the above embodiment, the optical path is bent, but this is made into a straight line) at each measurement point.
₁ , ₂ ( ₁ = ₂ in this example).

Then, when the surface to be inspected 2 is at the reference position,
The slit image detection positions by the line sensor 16 at the measurement location are _A10 (bit) and _A20, respectively.
(bit), and similar detection positions when the surface to be inspected 2 is shifted as shown by the thick line in FIG. 13 are A ₁₁ and A _21, respectively.

In addition, if the surface to be inspected 2 only has positional deviations ΔD ₁ and ΔD 2 at the measurement location and ΔD ₂ but does not have a rotational deviation θ, A ₁₁
becomes A ₁₁ ′ and A ₂₁ becomes A ₂₁ ′.

The deviation amounts ΔD ₁ and ΔD ₂ in this case are the same as equation (2) of the distance measurement principle (), assuming that the resolution of the line sensor 16 including the magnification of the condensing lens (not shown) is α (mm/bit). becomes as follows.

D ₁ = (A ₁₁ ′−A ₁₀ )×α/2tanβ …(4) D ₂ = (A ₂₁ ′−A ₂₀ )×α/2tanβ (5) However, in reality, a rotation of angle θ has occurred. Therefore, the reflection angle of the laser slit beam LST is expanded by 2θ, and A ₁₁ ′ in equation (4) is changed to A ₁₁ in equation (5).
When A ₂₁ ′ is replaced with A ₂₁ and ΔD ₁ and ΔD ₂ are calculated,
This results in distance measurement of the surface to be inspected at the virtual position indicated by the broken line in the figure.

Therefore, suppose that ΔD ₁ and ΔD ₂ are determined by the following equations as described above, and then θ is determined from these ΔD ₁ , ΔD ₂ and L.

D ₁ = (A ₁₁ − A ₁₀ )×α/tanβ …(6) D ₂ = (A ₂₁ −A ₂₀ )×α/2tanβ …(7) θ≒tan ⁻¹ {(ΔD ₁ −ΔD ₂ )/ L}...(8) Based on the values of ΔD ₁ , ΔD ₂ , and θ, ΔD ₁ ′ and ΔD ₂ ′ are determined again using the following equations.

ΔD ₁ ′=( _{11 10} − _{11 11} ′)/2tanβ=(A ₁₁ −A
₁₀ )×α−(ΔD ₁ + ₁ ) {tan(2θ+β)−tanβ}/
2tanβ…(9) ΔD ₂ ′=(A ₂₁ A ₂₀ −A ₂₁ A ₂₁ ′)/2tanβ=(A ₂₁ −A ₂₀ )
×α−(ΔD ₂ + ₂ ) {tan (2θ+β)−tanβ}/2tan
β...(10) Here, ΔD ₁ and ΔD ₁ ′ and ΔD ₂ and ΔD ₂ ′ are compared, and the following inequality is satisfied for the external setting value ε (completely different from ε in Figure 7). Determine whether or not.

|ΔD ₁ −ΔD ₁ ′|＜ε …(11) |ΔD ₂ −ΔD ₂ ′|＜ε …(12) If these formulas (11) and (12) are not satisfied, formula (9) and
Replace ΔD ₁ and ΔD ₂ in equation (10) with ΔD ₁ ′ and ΔD ₂ ′, calculate new ΔD ₁ ′ and ΔD ₂ ′, and repeat until equations (11) and (12) are satisfied. , ΔD ₁ ′ when satisfied,
Let ΔD ₂ ' be the value of ΔD ₁ and ΔD ₂ .

Further, θ obtained by equation (8) corresponds to Δθz explained in FIG. 4.

Note that it is desirable to set a maximum allowed number of times for the above-mentioned number of repeated calculations, and to display an abnormality when this is provided.

In this example, we have described the distance measurement and rotation angle calculation method when only the rotation of the workpiece around the z-axis with respect to the y-direction occurs. By adding measurement points that are separate from each other in the x direction, accurate distance measurement and calculation of each rotation angle can be performed in the same way as in the case described above, even when there is rotation around the x and y axes. It is possible.

〔Effect of the invention〕

As explained above, according to this invention,
Before the surface defect inspection device starts inspecting surface defects, the surface defect inspection device itself is used to detect the posture of the object to be inspected, and the detection head scan is taught to the robot in advance based on the result. Since the data is corrected, it is possible to accurately inspect surface defects even if there is a positional shift when setting the object to be inspected, and there is no need to install a dedicated video camera or the like to detect the posture of the object to be inspected. Since there is no,
Equipment costs can be reduced.

Furthermore, the measurement location for detecting the posture of the workpiece can be arbitrarily set, and there is no problem in surface defect inspection work.

[Brief explanation of the drawing]

FIG. 1 is a schematic configuration diagram showing an example of an apparatus for carrying out the present invention, FIG. 2 is a side view showing an example of a detection head mounted on a robot and scanning a surface to be inspected, and FIG. An explanatory diagram showing an example of an object to be inspected and measurement points for detecting its posture, FIG. 4 is a flowchart showing an example of the operation by the CPU unit 33 of FIG. 1, and FIGS. 5 to 7 are diagrams showing the present invention. FIG. 8 is a plan view, front view, and right side view showing the specific structure of the surface defect inspection device used for the inspection, and FIG. Figure 9 is an explanatory diagram of the distance measurement principle () used in this invention, Figure 10 is an explanatory diagram of a method for expanding the range of distance measurement using the distance measurement principle (), and Figure 11 is the first diagram in that case. A flow diagram showing the operation of the robot side and the CPU unit side, and Figure 12 shows the distance measurement principle used in this invention ().
FIG. 13 is an explanatory diagram of the distance measurement principle ( ) used in this invention. FIG. 14 is an explanatory diagram of the principle of a surface defect inspection apparatus to which this invention is applied. FIG. 16 is a waveform diagram for explaining the binarization process of the video output signal by the line sensor, and FIG. 17 is an explanatory diagram for explaining the principle of posture correction of the inspection apparatus. W...Object to be inspected (work), 2... Surface to be inspected, 6...
Surface defect inspection device, 7... Detection head, 8... Robot, 10... Laser slit light generator, 13... Screen, 14... Condensing lens, 15... Line sensor for defect detection, 16, 17... Line sensor for position correction , 18, 19...mirror, 20...detector, 3
0... Sensor receiver, 31... Defect detection circuit, 32
... Attitude correction circuit, 33... Microcomputer (CPU unit) for scanning data correction, 40... Beam splitter.

Claims (1)

[Claims] 1. Projecting laser slit light onto the surface of the object to be inspected, projecting the reflected light onto a screen formed by a diffuser plate, and concentrating the slit-shaped scattered light onto the line sensor using a condensing lens. In a surface defect inspection method in which a detection head of a surface defect inspection device that detects by imaging is attached to a robot and scans the surface of the object to be inspected while keeping a constant distance, defects on the surface are inspected, Prior to the start of surface defect inspection by the surface defect inspection device, using the surface defect inspection device,
The posture of the object to be inspected is detected by measuring the amount of deviation from a reference position at a certain distance from the detection head at a plurality of preset locations on the object to be inspected, and the robot 1. A method for inspecting a surface defect, comprising correcting data for scanning the detection head that has been taught in advance.