JP4334429B2 - Optical measuring method and optical measuring apparatus - Google Patents

Description

  The present invention relates to an optical measurement method and an optical measurement apparatus. More specifically, the present invention relates to an optical measurement method and an optical measurement apparatus suitable for an inspection object in which noise due to its surface characteristics exists and the noise distribution is not constant among inspection objects such as a glass substrate.

  As one of the optical measurement methods, for example, there is a method using a predetermined threshold as described in JP-A-2002-8444. By using this method, it is possible to realize a foreign matter detection device that detects foreign matter such as dust attached to a glass substrate used for a liquid crystal panel or the like.

  FIG. 6 is a diagram showing the principle of the foreign object detection device 5 using the conventional optical measurement method, and FIG. 7 is a diagram showing the relationship between the light intensity signal S1 of scattered light from the foreign object IB shown in FIG. The upper part of FIG. 7 shows that there are four foreign substances including foreign substances IBa and IBb on the glass substrate GL, and the lower graph of FIG. 7 represents the light intensity signal S1 in the line L0 in the upper part of FIG. As shown in FIG. 6, the foreign object detection device 5 receives the scattered light scattered on the glass substrate GL among the light emitted by the illumination means 51 that irradiates the glass substrate GL to be inspected and the illumination means 51. Then, it has a light receiving means 53 that can be finally converted into a two-dimensional scattered light intensity signal S1, and a comparison means 63 that compares the scattered light intensity signal S1 with a predetermined threshold T0.

  According to the foreign object detection device 5, when the foreign object IB adheres on the glass substrate GL, the light irradiated by the illumination means 51 is scattered by the foreign object IB and becomes scattered light. A part of the scattered light is received by the light receiving means 53, converted into a light intensity signal S1 of the scattered light, and output. The light intensity signal S1 of the scattered light is A / D converted by the A / D converter 61, and then compared with a predetermined threshold value T0 by the comparison unit 63, and is larger than the threshold value T0 as shown in FIG. Only in this case, the foreign substance signal SB is output. This process is performed on all the two-dimensional surfaces of the glass substrate GL. Thereby, it can be seen that the foreign matter IB is attached to the position on the glass substrate GL corresponding to the pixel of the foreign matter signal SB. Also, the size of the foreign matter IB can be known based on the total intensity of the foreign matter signal SB.

JP 2002-8444 A

  By the way, on the surface of the glass substrate GL to be inspected, there are generally small scratches, protrusions, and other surface roughness of about 1/2 to 1/10 or less than the size of the foreign matter IB to be detected. All of these surface properties can cause noise. This type of noise varies in both location and time without being constant. In the presence of such fluctuating noise, the above-described foreign matter detection device 5 has a problem that the detection accuracy is not good.

  For example, when the noise SN due to the surface characteristics is small and constant as shown in FIG. 7A, the light intensity signals SBa and SBb of the scattered light from the two foreign substances IBa and IBb are the light intensity signals S1 (= SB + SN) and a predetermined threshold T0 can be detected satisfactorily. However, as shown in FIG. 7B, when the noise SN has a non-constant distribution, it is possible to detect the left foreign object IBa by setting a predetermined threshold T0, but the right SN When the foreign matter IB adheres to a position where a relatively small noise SN exists like the foreign matter IBb, the light intensity signal S1 (= SBb + SN) of the scattered light does not exceed the threshold value T0, so the foreign matter IBb Cannot be detected.

  Furthermore, since this noise distribution varies depending on the location in the surface of the glass substrate GL to be inspected, and also varies for each glass substrate GL, it is difficult to set the predetermined threshold T0 as a common value in the first place. Many glass substrates for flat panel displays are coated with a metal film or the like. Particularly in such a case, surface roughness or the like causes noise, and the detection accuracy is significantly deteriorated. Therefore, there is a demand for a foreign object detection device that can achieve good detection accuracy regardless of the surface characteristics of the inspection object.

  The present invention has been made in view of the above-described problems, and an object thereof is to provide an optical measurement method and an optical measurement apparatus that can avoid deterioration in measurement accuracy caused by noise due to surface characteristics of an inspection object.

In order to solve the above-mentioned problem, the invention of claim 1 is an optical measurement method for optically measuring the surface state IB of the inspection object GL, as well shown in FIGS. The inspection object GL is irradiated with light, the light obtained from the inspection object GL is received and finally converted into a two-dimensional light intensity signal S1, and the surface state IB is determined from the light intensity signal S1. Only the attention signal SB is extracted by separating the attention signal SB, which is a signal necessary for measuring the signal, and the noise signal SN, which is an unnecessary signal for measuring the surface state, and the extracted attention signal SB and the predetermined signal The threshold value T1 is compared. Further, as well shown in FIGS. 2 (B) and 2 (C), the pixel of interest 40, which is a pixel in which the signal of interest SB exists, is extracted, and the pixel of interest is extracted from the extracted light intensity signal of the pixel of interest 40. The attention signal SB is extracted by removing the noise signal SN at 40. Further, as well shown in FIGS. 3 and 4, the first parameter P1 and the second parameter P2 are stored in advance in the storage unit 31, and the difference processing is sequentially performed on the light intensity signal S1 of the adjacent pixels, Based on the comparison between the difference value between adjacent pixels and the first parameter P1, the pixel at one end 42 of the target pixel 40 is recognized, and then the difference value between adjacent pixels and the second parameter P2 The position of the target pixel 40 is recognized by recognizing the pixel at the other end 44 of the target pixel 40 based on the size comparison, and the noise signal SN is calculated from the light intensity signals at both ends 42 and 44 of the recognized target pixel 40. The interpolated noise signal SN is subtracted from the light intensity signal of each pixel of interest 40. After recognizing the one end 42 of the target pixel, if the other end 44 of the target pixel 40 cannot be recognized, the position separated from the one end 42 by a predetermined pixel width is forcibly defined as the other end 44. Or it is possible to carry out the process of considering that the one end portion 42 is not recognized from the beginning.

The invention of claim 2 is an optical measuring apparatus 1 for optically measuring the surface state IB of the inspection object GL, and includes an illumination means 11 for irradiating the inspection object GL with light, and the inspection object GL. The light receiving means 13 which can receive the obtained light and finally convert it into a two-dimensional light intensity signal, and the attention signal which is a signal necessary for measuring the surface state IB from the light intensity signal S1. Attention signal extraction means 30 for extracting only attention signal SB by separating SB and noise signal SN, which is an unnecessary signal when measuring surface state IB, extracted attention signal SB and predetermined threshold value T1 The attention signal extraction means 30 includes a target pixel extraction unit 32 that extracts a target pixel 40 that is a pixel in which the target signal SB exists, and the extracted target pixel. A noise subtracting unit 33 that extracts the attention signal SB by removing the noise signal SN at the attention pixel 40 from the 0 light intensity signal S1, and the attention signal extraction means 30 includes the first parameter P1 and the second parameter P1. A storage unit 31 that stores parameter P2 in advance, and the pixel-of-interest extraction unit 34 sequentially performs a difference process on the light intensity signal S1 of the adjacent pixels obtained by the light receiving means 13, and performs a difference process between the adjacent pixels. Based on the comparison between the difference value of the first pixel P1 and the first parameter P1, the pixel at the one end portion 42 of the target pixel SB is recognized, and then the comparison between the difference value between adjacent pixels and the second parameter P2 is performed. The position of the target pixel 40 is recognized by recognizing the pixel at the other end 44 of the target pixel 40, and the noise signal SN is obtained from the light intensity signals at both ends 42 and 43 of the recognized target pixel 40. And while, the noise subtraction unit 33 is formed by subtracting the noise signal SN of interpolated from the light intensity signal S1 of each pixel of interest 40. The pixel-of-interest extraction unit 32 recognizes the one end portion 42 of the pixel of interest 40 and then moves away from the one end portion 42 by a predetermined pixel width when the other end portion 44 of the pixel-of-interest 40 cannot be recognized. Can be forcibly set to the other end 44 or the one end 42 is regarded as not recognized from the beginning.

  According to the present invention, an optical measurement method and an optical measurement apparatus capable of avoiding deterioration in measurement accuracy caused by noise due to surface characteristics of an inspection object are provided.

  FIG. 1 is a principle diagram of a foreign object detection device 1 according to the present invention, FIG. 2 is a diagram for explaining the processing in the arithmetic processing unit 20, FIG. 3 is a diagram showing a light intensity signal S1 including a foreign object signal SB, and FIG. FIG. 5 is a diagram for explaining the processing in the foreign object pixel extraction unit 32, and FIG. 5 is a diagram showing the relationship between the light intensity signal S1 and the substantial threshold value T1 ′.

  As shown in FIG. 1, the foreign object detection apparatus 1 includes an illuminating unit 11, an optical system 12, a light receiving unit 13, an arithmetic processing unit 20, and a display unit 14. The illuminating means 11 is arranged so as to irradiate a laser beam obliquely from above toward the glass substrate GL to be inspected. The optical system 12 includes a lens that is disposed between the glass substrate GL and the light receiving unit 13 and guides scattered light from the surface of the glass substrate GL to the light receiving surface of the light receiving unit 13. The light receiving means 13 is an array sensor camera that can receive this scattered light and finally convert it into a two-dimensional light intensity signal S1.

  The arithmetic processing device 20 includes an A / D conversion unit 21, a foreign object signal extraction unit 30, and a comparison unit 23. Of course, the A / D converter 21 may be disposed in the light receiving means 13. That is, the light receiving means 13 may be a so-called digital sensor that outputs a digital signal. The foreign object signal extraction unit 30 includes a storage unit 31, a foreign object pixel extraction unit 32, and a noise subtraction unit 33. The storage unit 31 stores a first parameter P1 and a second parameter P2 in advance. Parameters ΔL, ΔR and W may also be stored. The function of each parameter will be described later. The foreign object pixel extraction unit 32 detects the presence of the foreign object signal SB from the light intensity signal S1, and as shown in FIG. 2B, the pixel in which the foreign object signal SB exists or its region (hereinafter referred to as a foreign object pixel or a foreign object region). ) 40 is extracted. The foreign substance signal SB corresponds to the attention signal of the present invention. The foreign pixel extraction unit 32 corresponds to the target pixel extraction unit of the present invention.

  As shown in FIGS. 2B and 2C, the noise subtracting unit 33 extracts only the foreign matter signal SB by removing the noise signal SN in the foreign matter pixel 40 from the extracted light intensity signal S1 of the foreign matter pixel 40. . As shown in FIG. 2D, the comparison unit 23 compares the value of the foreign substance signal SB extracted by the foreign substance signal extraction unit 30 with a predetermined threshold value T1, and outputs it only when it is larger than the threshold value T1. The display means 14 is a display device having a display screen such as a CRT or a liquid crystal screen, and displays information on the foreign matter IB on the display screen based on the foreign matter signal SB output from the arithmetic processing unit 20.

  Next, the foreign object detection function of the foreign object detection apparatus 1 configured as described above will be described in detail. In FIG. 1, the illumination means 11 irradiates the glass substrate GL. When the foreign substance IB adheres to the glass substrate GL, the irradiation light hits the foreign substance IB and is scattered. A part of the scattered light is collected through the optical system 12 and guided to the light receiving means 13. The light receiving means 13 converts a part of the scattered light into a light intensity signal S1. The light intensity signal S1 includes a noise signal SN and a foreign matter signal SB.

  Here, the surface of the glass substrate GL to be inspected has minute scratches, protrusions, and other surface roughnesses of about 1/2 to 1/10 or less than the size of the foreign matter IB to be detected. Generally present, any of these surface properties can cause noise. The noise signal SN is a signal based on such surface characteristics of the glass substrate GL. The foreign matter signal SB is a light intensity signal of scattered light scattered by the foreign matter IB and is on the noise signal SN.

  The analog light intensity signal S <b> 1 captured by the arithmetic processing unit 20 is converted into a digital signal by the A / D converter 21. The foreign object pixel extraction unit 32 recognizes both end portions of the foreign object signal SB and extracts a region between them as the foreign object pixel 40. Specifically, for example, with respect to the light intensity signal S1 as shown in FIG. 3, differential processing is sequentially performed between adjacent pixels to obtain the relationship of the difference value ΔS1 as shown in FIG. As shown in FIG. 4, when there is a pixel 41 having a difference value having a larger ΔS1 than the first parameter P1 as a rising signal, the address of the pixel is set as a left end candidate of the foreign object IB. When a pixel having a difference value larger than the first parameter P1 as a rising signal continues in several adjacent pixels, the address of the rightmost pixel is set as a left end candidate.

  The address obtained by subtracting the parameter ΔL, for example, 1 to 5 pixels from the left end candidate is set as the left end 42 of the foreign object IB. By subtracting only the parameter ΔL, the end of the foreign object IB can be recognized with higher accuracy. Subsequently, when there is a pixel 43 having a difference value smaller than the second parameter P2 as a falling signal, the address of the pixel 43 is set as a right end candidate of the foreign object IB. When pixels having a difference value smaller than the second parameter P2 as falling signals continue with several adjacent pixels, the leftmost pixel is set as a left end candidate.

  The address obtained by adding the parameter ΔR, for example, 1 to 5 pixels from the right end candidate is set as the right end 44 of the foreign object IB. By adding only the parameter ΔR, the end of the foreign object can be recognized with higher accuracy. Of course, there are various variations in the recognition of the left edge candidate. For example, when a pixel having a difference value larger than the first parameter P1 as a rising signal continues in several adjacent pixels, it may be the leftmost pixel or the median value thereof. The same applies to the right edge candidate.

  Here, the first parameter P1 is preferably a value approximately equal to the amplitude of the high frequency noise. The second parameter P2 is preferably the same value as or smaller than the first parameter P1, and more preferably about ½ of the first parameter P1. The first and second parameters P1 and P2 may be set according to the high frequency noise to be removed, for example, the roughness of the film surface. Both of the parameters ΔL and ΔR preferably have an average distance between the ends of the target pixel of the foreign object IB to be detected, that is, about 1/3 of the foreign object region 40. For example, P1 = 10, P2 = −5, ΔL = 3, and ΔR = −3 are set. In this way, the left end portion 42 and the right end portion 44 of the foreign object signal SB are recognized, and an area sandwiched between them is extracted as the foreign object area 40.

  Next, processing to be performed is a process for removing the noise signal SN in the foreign substance region 40 from the light intensity signal S1 of the foreign substance region 40. The signal strength at the left end portion 42 is set as the noise signal at the left end portion 42, and similarly, the signal strength at the right end portion 44 is set as the noise signal at the right end portion 44, and linear interpolation is performed between them as shown in FIG. Thus, the noise signal SN in the foreign substance region 40 can be obtained, and a value obtained by subtracting the value of the noise signal SN from the light intensity signal S1 in the foreign substance region 40 can be obtained as the signal intensity of the foreign substance signal SB.

  If a signal that can be regarded as the right end portion 44 of the foreign object IB is not found after the certain address is regarded as the left end portion 42 of the foreign object IB, only a predetermined pixel width W (not shown) is forced from the left end portion 42. It is possible to select whether it is considered that the right end portion of the foreign object is on the right or whether the foreign object IB is not present from the beginning. By doing so, it is possible to avoid a state where the left end portion 42 is found but the right end portion 44 is not found. Alternatively, it is possible to avoid erroneously detecting the right end portion 44 at an unexpected position. The predetermined pixel width W is preferably a distance between end portions of an average target pixel of the foreign object IB to be detected, that is, an average value of the foreign object region 40.

  Further, after the left end portion 42 is considered, the processing when the left end candidate comes again before the right end portion candidate is either ignored for the new left end candidate or the new left end candidate is selected according to the purpose. It can be selected whether to consider it as the left end. Furthermore, the processing when two foreign object areas overlap due to the presence of foreign objects close to each other can be divided into two foreign object areas depending on the purpose, or two foreign object areas can be connected to one. It is possible to select processing to be performed. As a result, foreign object detection can be performed more accurately.

  The comparison unit 23 compares the value of the foreign matter signal SB extracted by the processing of the foreign matter signal extraction unit 30 with a predetermined threshold value T1, and outputs only when the value is larger than the threshold value T1. If this threshold value T1 is reduced, a smaller foreign object IB can be detected, and if it is increased, only a larger foreign object can be detected. In other words, it is a value that determines the sensitivity of detection. By performing the above processing on the entire surface of the glass substrate GL, the foreign substance signal SB corresponding to the threshold value T1 is extracted.

  In this way, the comparison unit 23 compares the value of the foreign object signal SB from which the noise signal SN is removed with a threshold value T1 having a predetermined magnitude. This substantially compares the light intensity signal S1 with the threshold value T1 'following the noise signal SN as shown in FIG. Therefore, even if the glass substrate GL has a metal film coated on the surface, it is possible to detect a foreign object with high accuracy without being affected by noise signal unevenness based on the roughness of the metal film.

  The display unit 14 displays the extracted foreign substance signal SB on the display screen as a two-dimensional image indicating the adhesion position on the glass substrate GL. Of course, the size of the foreign matter IB can be displayed, and the number distribution and the adhesion position distribution for each size can also be displayed. Furthermore, it is also possible to display the noise distribution on the contrary. That is, it is possible to know the roughness distribution of the glass substrate GL.

  In the above embodiment, the whole or part of the foreign object detection device 1, the configuration, the structure, the number, and the like, and the content and order of processing in the arithmetic processing device 20 can be changed as appropriate according to the gist of the present invention. In the above arithmetic processing unit 20, the above-described processing such as difference processing is performed on adjacent pixels, but it is also possible to perform the above-described processing on pixels that are not adjacent but one pixel or two pixels. is there. Therefore, the “adjacent pixel” in the present invention includes the meaning of a neighboring pixel. It is also possible to use more complicated low-pass filtering instead of the difference processing. Further, the present invention is not limited to the foreign matter detection device, and can be applied to an optical measurement device for other purposes. For example, the present invention can be applied to a defect inspection device, a protrusion height measuring device, a three-dimensional shape measuring device, a film thickness measuring device, a roughness measuring device, a pattern recognition device, a width measuring device, and the like.

It is a principle figure of the foreign material detection apparatus which concerns on this invention. It is a figure for demonstrating the process in an arithmetic processing part. It is a figure which shows the light intensity signal containing a foreign material signal. It is a figure for demonstrating the process in a foreign material pixel extraction part. It is a figure which shows the relationship between a light intensity signal and a threshold value substantially. It is a principle figure of the foreign material detection apparatus using the conventional optical measuring method. It is a figure which shows the relationship between the light intensity signal of the scattered light by the foreign material shown in FIG. 6, and a threshold value.

Explanation of symbols

1 Foreign object detection device (optical measurement device)
DESCRIPTION OF SYMBOLS 11 Illumination means 13 Light reception means 23 Comparison part (comparison means)
30 Foreign object signal extraction unit (attention signal extraction means)
31 Storage Unit 32 Foreign Object Pixel Extraction Unit (Target Pixel Extraction Unit)
33 Noise subtractor 42 Left end (one end)
44 Right end (the other end)
GL glass substrate (subject to inspection)
IB Foreign matter (surface condition)
S1 Light intensity signal SB Foreign object signal (attention signal)
SN noise signal T1 threshold 40 foreign object pixel, foreign object region (target pixel)
P1 First parameter P2 Second parameter

Claims (2)

An optical measurement method for optically measuring a surface state of an inspection object, irradiating the inspection object with light, receiving light obtained from the inspection object, and finally a two-dimensional light intensity The signal of interest is converted into a signal, and the signal of interest is separated from the light intensity signal by separating the signal of interest, which is a signal necessary for measuring the surface state, and a noise signal, which is an unnecessary signal when measuring the surface state. In the optical measurement method, wherein only the extracted signal of interest is compared with a predetermined threshold value,
An optical measurement method for extracting a target pixel, which is a pixel in which the target signal exists, and extracting the target signal by removing a noise signal in the target pixel from the extracted light intensity signal of the target pixel,
The first parameter and the second parameter are stored in advance in the storage unit, the difference processing is sequentially performed on the light intensity signals of the adjacent pixels, and based on the magnitude comparison between the difference value between the adjacent pixels and the first parameter. The pixel at the one end of the target pixel is recognized, and then the pixel at the other end of the target pixel is recognized by comparing the difference value between adjacent pixels with the second parameter. Recognizing the position, interpolating a noise signal from the light intensity signal at both ends of the recognized pixel of interest, and subtracting the interpolated noise signal from the light intensity signal of each pixel of interest,
After recognizing one end of the pixel of interest, if the other end of the pixel of interest cannot be recognized, a position separated by a predetermined pixel width from the one end is forced to be the other end, or An optical measurement method capable of performing either of the processes for determining that the one end portion is not recognized from the beginning. An optical measuring device for optically measuring the surface state of an inspection object, and an illumination means for irradiating the inspection object with light, and receiving light obtained from the inspection object and finally two-dimensionally A light receiving means capable of converting into a light intensity signal, a signal of interest which is a signal necessary for measuring the surface state, and a noise signal which is an unnecessary signal for measuring the surface state, among the light intensity signals. An optical measurement device comprising: attention signal extraction means for extracting only the attention signal by separating the signal; and comparison means for comparing the extracted attention signal with a predetermined threshold,
The attention signal extracting unit extracts a target pixel that extracts a target pixel that is a pixel in which the target signal exists, and removes a noise signal in the target pixel from the extracted light intensity signal of the target pixel. An optical measuring device having a noise subtracting unit for extracting
The attention signal extraction unit includes a storage unit that stores a first parameter and a second parameter in advance, and the target pixel extraction unit sequentially performs a difference process on the light intensity signals of adjacent pixels obtained by the light receiving unit. And recognizing a pixel at one end of the pixel of interest based on a comparison between the difference value between the adjacent pixels and the first parameter, and then, between the difference value between the adjacent pixels and the second parameter. Recognizing the position of the target pixel by recognizing the pixel at the other end of the target pixel based on the size comparison, interpolating a noise signal from the light intensity signals at both ends of the recognized target pixel, the noise subtracting unit , An optical measurement device that subtracts the interpolated noise signal from the light intensity signal of each pixel of interest,
The pixel-of-interest extraction unit forcibly identifies a position separated by a predetermined pixel width from the one end when the other end of the target pixel cannot be recognized after recognizing the one end of the target pixel. Or an optical measuring device capable of performing either of the processing that considers that the one end is not recognized from the beginning.

Images