JP3584346B2 - Inspection device - Google Patents

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an inspection device that optically and electrically detects defects such as scratches, dirt, and foreign matter on an optical disk or the like using a CCD linear image sensor or the like.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, there has been provided an inspection apparatus which receives reflected light from an object to be inspected using a CCD sensor, compares an output signal thereof with a threshold signal, and detects defects such as scratches, dirt, and foreign matter. In this inspection device, the threshold signal is formed from the output signal of the CCD sensor. This is because the output signal itself of the CCD sensor fluctuates irrespective of the defect, and it is necessary to use a threshold with a fluctuation similar to the fluctuation of the output signal.
[0003]
However, when an object to be inspected is, for example, a mini disk (MD) or the like, a mirror surface region and an utilization region having different average reflectivities are formed adjacent to each other on the object. Of these, the inspection device needs to detect a defect in the use area, but if no measures are taken, the mirror area having higher reflectivity than the use area is replaced with a defect on the inspection object. May be mistaken.
[0004]
Therefore, in the above inspection apparatus, there is a discontinuity in the reflectance between the use area, that is, the inspection area and the mirror area, and due to this, an edge (mirror edge) occurs in the output signal. Attention is paid to identifying an inspection area using the detected mirror edge, thereby preventing erroneous detection of a defect. FIG. 2 is a time chart showing signals of various parts in the inspection apparatus according to the present invention. The detection of a mirror edge in the conventional inspection apparatus will be described with reference to FIG. The output signal is shown. The output signal has mirror edges GE and NE. On the other hand, FIG. 3D shows a detection window signal for detecting one mirror edge GE from the output signal, and a detection window having a window width corresponding to 128 pixels of the CCD sensor is used to detect the inspection object. The position S set in advance from outside according to the shape or the like is formed as the right end. Then, the mirror edge GE is detected by detecting the falling edge of the output signal in the detection window. The other mirror edge NE is similarly detected using the detection window signal shown in FIG.
[0005]
[Problems to be solved by the invention]
The reason why the width of the detection window is set to 128 pixels in the above-described conventional inspection apparatus is that a mirror edge is always captured in the detection window in consideration of, for example, differences in stampers, eccentricity of a disc, clearance of a spindle core, and the like. It is to be able to be. However, since such a relatively wide detection window is used, when a defect such as a scratch exists in the mirror surface area, the edge of the output signal generated by the defect is captured in the detection window, and this is referred to as mirror edge GE, NE. There was a problem of misidentification. When such a misperception occurs, a portion of the output signal caused by light reflected from a mirror surface region having a different average reflectivity affects a threshold signal, thereby causing a threshold bell to fluctuate and erroneously detecting a defect. There was a problem that would.
[0006]
SUMMARY OF THE INVENTION The present invention has been made to solve the above-described conventional drawbacks, and an object of the present invention is to accurately detect a boundary between a mirror surface area and an inspection area on an object to be inspected, thereby detecting an erroneous defect. It is an object of the present invention to provide an inspection device capable of preventing the inspection.
[0007]
[Means for Solving the Problems]
Therefore, the inspection apparatus according to claim 1 is configured such that a mirror area M and an inspection area D having different average reflectances are adjacent to each other, and a boundary A between the two areas M and D is formed continuously. , A linear detection area L is set over both the areas M and D, the reflected light R from the detection area L is received by the sensor 8, and a change in the output signal causes the mirror area in the detection area L to change. The position of the boundary B between the M and the inspection area D is grasped, defects such as foreign matter in the inspection area D are detected, and the detection area L is sequentially moved in the direction in which the boundary B is continuous to thereby inspect the inspection object A. In the inspection apparatus configured to scan the upper side, the range where the boundary B can be located in the detection area L after the movement is estimated using the position of the boundary B grasped in the detection area L before the movement. It is characterized by:
[0008]
In the inspection apparatus of the first aspect, since the boundary B between the mirror surface area M and the inspection area D is continuously formed, the detection after the movement is performed using the position of the boundary B grasped by the immediately preceding inspection. The position of the boundary B in the area L can be estimated with relatively high accuracy. Therefore, it is possible to avoid erroneously recognizing the position of the boundary B due to a defect or the like of the mirror surface area M, and prevent erroneous detection of the defect.
[0009]
In the inspection apparatus according to the second aspect, as shown in FIG. 1, a mirror area M and an inspection area D having different average reflectances are adjacent to each other, and a boundary B between the two areas M and D is formed continuously. A scanning means 41 for setting a linear detection area L over the two areas M and D and sequentially moving the set detection area L in a direction in which the boundary B is continuous, A sensor 8 for receiving the reflected light R from the detection area L, a window signal generating means 42 for generating a detection window signal having a predetermined detection window, and detecting the mirror edges GE and NE from the edge signals existing in the detection window. And a defect detecting means 44 for detecting the position of the boundary B between the mirror area M and the inspection area D using the mirror edges GE and NE, and detecting a defect such as a foreign matter in the inspection area D. And configured with In the inspection apparatus, the window signal generating means 42 generates a detection window signal having a detection window of a predetermined first window width at a preset position immediately after the start of the inspection, while the mirror edge detecting means 43 When the specular edges GE and NE are detected, in the next inspection performed by moving the detection area L, a position smaller than the first window width is set at a position set using the detected specular edges GE and NE. A detection window signal having a detection window having a second window width is generated.
[0010]
In the inspection apparatus of the second aspect, since the boundary B between the mirror surface area M and the inspection area D is formed continuously, even if the second window width is relatively narrow, the mirror edge GE, NE can be accurately detected. Therefore, an edge of an output signal caused by a defect or the like existing in the mirror area M can be excluded from the detection window, and the boundary between the mirror area M and the inspection area D can be accurately grasped to prevent erroneous detection of a defect. It becomes possible.
[0011]
BEST MODE FOR CARRYING OUT THE INVENTION
Next, specific embodiments of the inspection apparatus of the present invention will be described in detail with reference to the drawings.
[0012]
FIG. 3 is a schematic system configuration diagram of the entire inspection apparatus. In the figure, A is an object to be inspected. The object to be inspected A is, for example, a mini-disc (MD), and a mirror surface region M having a high average reflectance is formed on the inner and outer peripheral sides thereof. I have. A utilization area, that is, an inspection area D in which it is necessary to detect a defect such as a scratch or a foreign substance, is formed between the two mirror areas M adjacent to the mirror area M. Further, in the inspection area D, the average reflectance is lower than the average reflectance of the mirror area M, so that the reflectance B is discontinuous at the boundary B between the two areas M and D. Further, the mini-disc A to be inspected is rotationally driven by a motor 1 composed of, for example, a step motor, and the rotation of the motor 1 is controlled by a motor driving unit 2.
[0013]
This inspection apparatus irradiates the surface of the mini-disc A with light, and optically and electrically detects defects such as scratches, dirt, foreign matter, and pinholes on the disc surface from the reflected light. Is used as the lamp 3. The light from the lamp 3 is totally reflected by the reflecting mirror 5 via the light projecting optical system 4 and is projected on the surface of the mini disc A. The reflected light R from the surface of the mini-disc A is totally reflected by the reflecting mirror 6 and received by the CCD sensor 8 via the light receiving optical system 7. Here, for example, a tungsten halogen lamp is used as the lamp 3, and a condenser lens system or a collimator lens system is used as the optical systems 4 and 7, for example. As the CCD sensor 8, a high-speed driving type CCD line sensor is used. Since the CCD line sensor is used as described above, the CCD sensor 8 receives the reflected light R from the linear detection area L formed on the surface of the optical disc A. The detection area L is set along both the mirror area M and the inspection area D from the inner peripheral end to the outer peripheral end of the optical disk A along the radial direction. Then, by rotating the optical disk A by a small angle by the motor 1, the detection area L is sequentially moved in the circumferential direction, and the inspection area D of the optical disk A is inspected without fail. As described above, the motor 1 and the motor driving unit 2 constitute the scanning unit 41 shown in FIG.
[0014]
Next, the function of the defect detection unit 44 constituted by the control unit 31 will be described. The output signal w1 of the CCD sensor 8 is amplified by a CCD driver 9 and input to a threshold generation circuit 13 including a function of an A / D converter 10 and a low-pass filter (LPF) at the next stage. The output signal w2 of the CCD driver 9 fluctuates up and down with respect to the reference level when a foreign substance or a pinhole is detected. The signal which fluctuates up and down is detected by using a threshold level. However, since the detection signal itself fluctuates, if the threshold level is set to an absolute constant value, there is a risk of erroneous detection. Therefore, the threshold level is also changed according to the change of the detection signal from the CCD sensor 8. The low-pass filter included in the threshold generation circuit 13 is used to generate the threshold level that fluctuates in this manner. That is, since the output of the CCD sensor 8 includes a pulse-like signal projecting up and down when a pinhole or foreign matter is detected, a threshold level that smoothly changes by using a low-pass filter to suppress this steep fluctuation is used. It is generating.
[0015]
On the other hand, by using a low-pass filter to generate the threshold level, the threshold level is delayed in time. Therefore, the output signal from the CCD sensor 8 cannot correspond to the threshold level to be compared with the output signal. Therefore, a delay circuit (digital delay line) is provided at the next stage of the A / D converter 10 to eliminate the time delay between the output signal of the CCD sensor 8 and the threshold level, and to make the two correspond exactly. The digital signal output from the delay circuit 11 is converted into an analog signal by the D / A converter 12, and the output signal w3 of the D / A converter 12 and the signal w5 from the threshold generation circuit 13 are sent to the signal processing circuit 14. I have. The circuits 10 to 14 constitute the control unit 31.
[0016]
The signal processing circuit 14 composed of a microcomputer or the like has a configuration as shown in FIG. 4 and is set by the level setting unit 15 for setting the threshold level to a predetermined value level and the level setting unit 15. Comparators 16 and 17 for comparing the threshold level with the output signal w3 of the D / A converter 12, and binarized pulses corresponding to scratches, pinholes, foreign matter, etc. of the mini disc A by obtaining the outputs of these comparators 16 and 17 And a determination unit 20 that receives the output signal w2 of the CCD driver 9 and generates a sampling interval pulse w4 in response thereto, and stores the determination result of the determination unit 20, the position of a flaw, a pinhole, and the like. It comprises an input / output interface I / O 22 for inputting and outputting data between the memory 21 and the personal computer 23. There. The binarized pulse train data of the determination result of the determination unit 20 is stored in the memory 21, and the data is arithmetically processed while being sequentially read out by the personal computer 23. The color CRT 24 displays the mapping of the defective portion almost simultaneously with the determination result. So that you can instantly know the image of the defect. The printer 26 can output the above determination result and the like. 25 is a keyboard. Note that (1) and (2) in FIG. 3 correspond to (1) and (2) in FIG.
[0017]
Next, a principle of signal processing for generating a binarized pulse for detecting a defect such as a flaw, a pinhole, or a foreign matter will be described with reference to FIGS. First, when an image is formed on the CCD sensor 8 by the optical system, video data is output by processing in the CCD driver 9. Based on the video data, the control section 31 performs signal processing in real time to form data that can be processed by the personal computer 23. Therefore, the principle from capture of video data to generation of a binarized pulse (flaw binarized pulse) will be described below. This corresponds to the function of the defect detection means 44 shown in FIG. As a specific example, a case will be described in which a minidisk (MD) is to be inspected as the inspection object A as described above.
[0018]
The SH pulse signal (source signal) in FIG. 5A is a signal indicating the scan timing for one line of the CCD sensor 8 and is a waveform at the input end of the control unit 31. In addition, the SH pulse (after correction) shown in FIG. 5 (b) has a waveform that is inevitably delayed each time various waveform processing is performed in the process of extracting various binarized pulses (flaw binarized pulses). Therefore, the SH pulse is obtained by delaying the delay time T1 corresponding to the most delay by the shift register and correcting the relationship with the binarized pulse position. FIG. 5C shows a raw video waveform w2 output from the CCD driver 9, which is a waveform at the input end of the control unit 31 as in FIG. The raw video waveform w2 between the SH pulses (source signals) is a waveform obtained from one of the linear detection areas L in one scan.
[0019]
FIG. 5D shows a sampling section pulse w4. This is a pulse that determines the section when sampling the video raw waveform w2 in FIG. 5C in order to follow the background level of the video waveform. This is formed by the determination unit 20 based on the video raw waveform w2, and the details will be described later. FIG. 5E shows a sampling video waveform w5, which is a basic waveform for generating various threshold waveforms. In the section where the sampling section pulse w4 is not output as shown in FIG. 5D, the level sampled at the sampling tail position of the previous scan is held as the hold level section until the sampling head position of the next scan. However, this is performed by the threshold generation circuit 13 described above.
[0020]
Here, the present inspection apparatus has various inspection modes according to types of scratches, pinholes, and the like on the inspection object, and generates a binarized pulse corresponding to the inspection result. And performs signal processing. In addition, various modes can be inspected for each mode, and inspection can be performed simultaneously for each mode.
[0021]
The waveform shown in FIG. 5 (f) is an AK mode waveform w3, and when the A mode binarization pulse w9 and the K mode binarization pulse w8 are generated, the inspection target inputted to the comparators 16 and 17 It is a waveform which becomes. Then, as in the case of (b) above, the delay time T1 equivalent to the most delayed threshold waveform is delayed by the digital delay line (delay circuit 11), and the video waveform is corrected in relation to the binarized pulse position. is there.
[0022]
FIG. 5G illustrates a K-mode threshold waveform w6, which is a threshold level waveform input to the comparator 16 when the K-mode binarized pulse w8 is generated. Here, the K mode threshold level is used in a range of 100% or more and less than 200% with respect to the background level of the AK mode video waveform w3. This K-mode threshold waveform w6 is generated by the level setting unit 15 of the signal processing circuit 14. The comparator 16 compares the A / K mode video waveform w3 with the K mode threshold waveform w6 to generate the above-described K mode binarized pulse w8.
[0023]
FIG. 5H shows an A-mode threshold waveform w7, which is a threshold level waveform input to the comparator 17 when the A-mode binarized pulse w9 is generated. The A-mode threshold level is used at 100% or less of the background level of the AK mode video waveform w3. The A-mode threshold level waveform w7 is generated by the level setting unit 15, and the comparator 17 compares the A / K mode video waveform w3 with the A-mode threshold waveform w7 to generate the A-mode binarization pulse w9 described above. It has become.
[0024]
FIG. 5 (i) shows a K-mode binarization pulse w8, which is an H-level pulse when the AK video waveform level becomes higher than the K-mode threshold level. If the surface of the mini-disc A has a flaw or the like, the reflectivity increases, and the K-mode binarization pulse w8 is generated in that portion. Since the mirror surface area M having a high reflectance is regarded as a flaw in the K mode, in order to prevent this, only the K mode is forcibly masked (hardware forced mask) in sections other than the sampling section, and excluded from the flaw inspection target range. ing.
[0025]
FIG. 5 (j) shows an A-mode binarization pulse w9, which becomes an H level when the AK video waveform level becomes lower than the A-mode threshold level. When a pinhole is present in the mini-disc A, the light is not reflected, so that the AK video waveform level is lowered, and this A-mode binarization pulse is generated.
[0026]
Here, the software compulsory mask shown in FIG. 5 will be described. In the video output of the CCD sensor 8, there are always dummy pixels, before and after the SH pulse, where no video waveform appears and where it does not appear reliable. Since the number of the dummy pixels differs depending on the type of CCD, the CCD type is automatically read by software, and the dummy pixel section corresponding to the type is forcibly masked. This mask area is common to the K mode and the A mode.
[0027]
The user mask refers to the following mask. That is, by inputting the inner diameter and outer diameter of the disk with the keyboard of the personal computer, the area outside the disk is automatically masked, which is called a user mask. This mask area is common to the K mode and the A mode as described above.
[0028]
Next, the function of the determination unit 20 that forms the sampling section pulse w4 based on the video raw waveform w2 will be described with reference to FIG. This corresponds to the function of the window signal generating means 42 and the mirror edge detecting means 43 shown in FIG.
[0029]
FIG. 2A shows an SH pulse (source signal), and FIG. 2B shows a video raw waveform w2. As described above, the video raw waveform w2 between the SH pulses is a waveform obtained by scanning one detection area L set on the mini disc A. Then, every time an SH pulse is generated, that is, each time one scan is completed, the mini disc A to be inspected is rotated by the motor 1 at a slight angle. Further, the boundary B where the reflectance is discontinuous is formed on the inner peripheral side and the outer peripheral side of the mini disk A as described above. The boundary B appears in the raw video waveform w2 as the outer peripheral mirror edge GE and the inner peripheral mirror edge NE. Therefore, the waveform between the outer mirror edge GE and the inner mirror edge NE is a waveform obtained from the inspection area D of the mini disc A.
[0030]
Therefore, before the start of the inspection, first, the rightmost value (S shown in FIG. 4D) of the detection window (outer peripheral designation gate) for detecting the outer peripheral mirror edge GE and the detection window for detecting the inner peripheral mirror edge NE. The rightmost value (E shown in FIG. 10E) of the (inner circumference designation gate) is input from the keyboard 25 in advance. Then, a detection window signal having a first window width corresponding to 128 pixels of the CCD sensor 8 from the rightmost end values S and E to the left, that is, an outer periphery designation gate signal shown in FIG. An inner circumference designation gate signal shown in (e) is generated. Here, the rightmost values S and E and the window width of 128 pixels of the CCD sensor 8 are determined in consideration of the size of the mini disc A, the difference of the stamper, the eccentricity, the spindle clearance of the spindle, and the like. It is.
[0031]
Next, the slope of the waveform is determined from the video raw waveform w2 set in each of the designated gates of the outer peripheral designated gate signal and the inner designated gate signal (FIG. 3 (c)). Or through a differentiating circuit, and assuming that this is the outer peripheral mirror edge GE or the inner peripheral mirror edge NE, a start edge pulse or an end edge pulse shown in FIGS. Then, the element counter values (see (h) in the figure) when the start edge pulse and the end edge pulse are generated are stored in the start memory and the end memory provided in the memory 21 ((i) and (j) in the figure). )reference). Here, the element counter is cleared to "0" by the SH pulse, and starts counting from that position. The counter values stored in the start memory and the end memory continue to be stored until the next scan, and the sampling interval pulse w4 is formed based on the counter values.
[0032]
The above is the operation of the first scan immediately after the start of the inspection. In the next scan, the detection window is not based on the rightmost values S and E input from the keyboard 25 but on the counter values stored in the start memory and the end memory. Generate a signal. That is, a start prediction gate signal ((k)) having a second window width of ± 4 pixels before and after the counter value stored in the start memory and the counter value stored in the end memory. And an end prediction gate signal ((l) in the figure) having the second window width of ± 4 pixels before and after the center. Then, these are used as detection window signals instead of the outer circumference specifying gate signal and the inner circumference specifying gate signal, so that the outer mirror edge GE and the inner mirror edge NE are detected in the same manner as described above. In the subsequent scans, the start prediction gate signal and the end prediction gate signal are generated by using the counter values stored in the start memory and the end memory in the immediately preceding scan, and the detection of the mirror edges GE and NE is sequentially performed. Do.
[0033]
In the inspection object such as the mini disk A, the boundary B between the mirror surface area M and the inspection area D is continuously formed in the circumferential direction, and the scanning is performed by moving the detection area L by a small distance in the circumferential direction. Is going. Therefore, even if there is a large amount of eccentricity in the mini disc A, the element counter value at which the mirror edges GE and NE are captured between adjacent scans hardly changes. Therefore, the second window width of the start prediction gate signal and the end prediction gate signal corresponds to the eight pixels of the CCD sensor 8, that is, the first window width of the outer periphery designation gate signal or the inner periphery designation gate signal. Even at 1/16, the specular edges GE and NE can be reliably captured in the prediction gate. Since the second window width is made very narrow in this way, even if a defect such as a scratch exists in the mirror area M, and an edge signal is generated in the video raw waveform w2, the edge signal is reduced by the narrow prediction gate. It can be easily eliminated. Therefore, an accurate sampling section pulse w4 can be formed, and erroneous detection of a defect caused by a fluctuation in the threshold level can be prevented. If any emergency occurs during the scan and the specular edges GE and NE are not detected during the prediction gate ((k) and (l) in FIG. 4), the designated gate signal ((d) in FIG. The mirror edges GE and NE are detected using e)).
[0034]
On the other hand, the operation of detecting the mirror edges GE and NE as described above is continuously performed even when the mini disc A as the inspection object is replaced. However, when the mini disc A is replaced, the light is not focused on the surface of the mini disc A, so that the video raw waveform w2 becomes a distorted waveform, so that the mirror edges GE and NE are detected at positions different from the original positions. May have consequences. If this occurs, the sampling section pulse w4 will not be accurate, and the threshold level will be affected, resulting in erroneous detection of a defect. Therefore, in this inspection apparatus, the element counter value stored in the memory by the last scan is not updated with the mirror edges GE and NE detected during the replacement of the mini disk A, and the first inspection in the next inspection of the mini disk A is performed. It is maintained until the start of scanning. With the same stamper lot, the mirror edges GE and NE of the next mini-disc A are also detected at very close positions, so that the operation using the prediction gate can be performed more smoothly.
[0035]
Although the specific embodiments of the present invention have been described above, the present invention is not limited to the above embodiments, and can be implemented with various modifications within the scope of the present invention. In the above description, a mini disk is used as an object to be inspected, but other optical disks such as a compact disk may be used. Further, the specific window width of the designated gate or the prediction gate can be appropriately changed depending on the size and variation of the inspection object.
[0036]
【The invention's effect】
In the inspection apparatus of the first aspect, since the boundary between the mirror surface area and the inspection area is continuously formed, the position of the boundary in the detection area after the movement is determined using the position of the boundary obtained by the immediately preceding inspection. The position can be estimated with relatively high accuracy. Therefore, it is possible to avoid erroneously recognizing the position of the boundary due to a defect in the mirror surface area or the like, and prevent erroneous detection of the defect.
[0037]
In the inspection apparatus according to the second aspect, since the boundary between the mirror surface region and the inspection region is formed continuously, even if the second window width is relatively narrow, the mirror surface edge can be accurately captured in the detection window. Can be. Therefore, the edge of the output signal caused by a defect or the like existing in the mirror area can be excluded from the detection window, and the boundary between the mirror area and the inspection area can be accurately grasped, thereby preventing erroneous detection of the defect. It becomes.
[Brief description of the drawings]
FIG. 1 is a schematic block diagram of an embodiment of an inspection device of the present invention.
FIG. 2 is a timing chart for explaining detection of a mirror edge in the inspection apparatus.
FIG. 3 is a schematic system configuration diagram of the entire inspection apparatus.
FIG. 4 is a block diagram of a signal processing circuit of the inspection device.
FIG. 5 is a timing chart for explaining the principle of signal processing for generating a binarized pulse in the inspection apparatus.
[Explanation of symbols]
8 CCD sensor
41 Scanning means
42 Window signal generating means
43 Mirror edge detection means
44 Defect detection means
A mini disk
B boundary
D Inspection area
L detection area
M mirror surface area
R reflected light
GE mirror edge
NE Mirror surface edge

Claims (2)

An inspection object (A) in which a mirror area (M) and an inspection area (D) having different average reflectances are adjacent to each other, and a boundary (B) between both areas (M) and (D) is continuously formed. On the other hand, a linear detection area (L) is set over both the areas (M) and (D), and the reflected light (R) from the detection area (L) is received by the sensor (8). From the change in the output signal, the position of the boundary (B) between the mirror area (M) and the inspection area (D) in the detection area (L) is grasped, and a defect such as a foreign substance in the inspection area (D) is detected. Further, in the inspection apparatus configured to scan the inspection object (A) by sequentially moving the detection area (L) in a direction in which the boundary (B) is continuous, the detection area (L) before movement is grasped. Using the position of the boundary (B) thus determined, a range in which the boundary (B) can be located in the detection area (L) after the movement. Inspection apparatus is characterized in that so as to estimate. For an inspection object in which a mirror surface area (M) and an inspection area (D) having different average reflectances are adjacent to each other, and a boundary (B) between both areas (M) and (D) is continuously formed, Scanning means (41) for setting a linear detection area (L) over both the areas (M) and (D) and sequentially moving the set detection area (L) in a direction in which the boundary (B) is continuous. ), A sensor (8) for receiving the reflected light (R) from the detection area (L), a window signal generating means (42) for generating a detection window signal having a predetermined detection window, And a mirror edge detecting means (43) for detecting a mirror edge (GE) (NE) from an edge signal existing in a mirror area (M) and an inspection area (D) using the mirror edge (GE) (NE). The position of the boundary (B) is grasped and defects such as foreign matter in the inspection area (D) are detected. In the inspection apparatus provided with a defect detection means (44) for outputting the detection signal, the window signal generation means (42) is provided with a detection window having a predetermined first window width at a preset position immediately after the start of the inspection. On the other hand, when the mirror edge detecting means (43) detects the mirror edge (GE) (NE), in the next inspection performed by moving the detection area (L), A detection window signal having a detection window having a second window width smaller than the first window width is generated at a position set using the detected mirror edge (GE) (NE). Inspection equipment.

Images