JP2009032829A - Substrate detection device and substrate processing apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a substrate detection device which can accurately detect the holding state of a substrate under processing. <P>SOLUTION: The substrate detection device includes: a picked-up image generation means to photograph a held substrate plural times under different exposure conditions and generate a plurality of picked-up image data; an image synthesizing means to synthesize the picked-up image data and generate a synthesized image data; a holding position specification means to specify the holding position of the substrate in a holding means on the basis of the synthesized image; and a holding condition determination means to compare the holding position of the specified substrate with the supposed holding position and to generate a detection result indicating that the holding condition of the substrate is normal or not according to that both positions are matched or not. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an apparatus for detecting a holding state of a substrate in a substrate processing apparatus that performs predetermined processing such as cleaning and etching on the substrate.

  A substrate loading / unloading apparatus for transferring a substrate to / from a substrate storage container having a plurality of storage shelves in which substrates are stored in a substrate processing apparatus, and storing a plurality of stages of substrates regardless of the type of the substrate storage container Devices that can detect the presence or absence of a substrate at each position are already known (see, for example, Patent Document 1).

  In the apparatus disclosed in Patent Document 1, it is obtained by imaging the substrate storage container with the illumination turned off from the image data obtained by imaging the substrate storage container with the illumination turned on. By subtracting the image data, the image data of the edge portion of the substrate is extracted, and the presence / absence of the substrate at each accommodation position is determined based on the image data.

JP 11-31725 A

  The technique disclosed in Patent Document 1 detects the presence or absence of a substrate in the substrate container by imaging the substrate container on which the substrate is placed in the substrate carry-in / out section of the substrate processing apparatus. The part for which the presence or absence of the substrate is required in the substrate processing apparatus is not limited to this. That is, in the middle of a certain processing process, it is important to grasp whether or not the substrate is held in the same manner as when the substrate is loaded without dropping off. For example, in the case of a batch type substrate processing apparatus that performs chemical processing on a substrate by immersing a plurality of substrates in a chemical solution, the substrate is immersed when the plurality of substrates are pulled up after being immersed in the chemical solution. It is necessary to confirm that the substrate is held at the same holding position as before. If there is a difference in the holding state of the substrate before and after the immersion, the substrate has dropped into the chemical solution for some reason during the chemical treatment, which causes a problem.

  However, the technique disclosed in Patent Document 1 is not always applicable to substrate detection in the middle of such a processing process. For example, taking the chemical treatment process as an example, when the technique disclosed in Patent Document 1 is applied, the imaging means and the illumination means are provided in a mode in which they are directly exposed to the chemical treatment atmosphere. Since it is difficult from the viewpoint of the chemical resistance of the means, it is necessary to devise the arrangement relationship between the imaging means and the illumination means, the chemical tank and the holding means, and further the setting of the detection execution position.

  In addition, when imaging processing is performed with the arrangement configuration of each means described above being suitable, the captured image data obtained is not necessarily the same as that obtained in the technique according to Patent Document 1, and thus the holding state The data processing performed to determine whether or not the image data must be in accordance with the characteristics of the acquired captured image data.

  The present invention has been made in view of the above problems, and a substrate detection apparatus capable of accurately and simply detecting a holding state of a substrate in the middle of a processing process in a substrate processing apparatus, and the same It aims at providing the substrate processing apparatus provided with.

  In order to solve the above-mentioned problems, the invention of claim 1 is an apparatus for detecting a holding state of a substrate held by a predetermined holding means capable of holding a plurality of substrates in a substrate processing apparatus that performs a predetermined process on a substrate. In the predetermined detection execution position, the illumination unit that irradiates illumination light to the substrate held by the holding unit and the substrate that is irradiated with the illumination light are imaged a plurality of times under different exposure conditions. On the basis of a captured image generating means for generating a plurality of captured image data, an image combining means for combining the plurality of captured image data to generate composite image data, and a composite image represented by the composite image data, The holding position specifying means for specifying the holding position of the substrate in the holding means, the holding position of the substrate specified by the holding position specifying means, and the holding position of the substrate Compared with an assumed holding position that is assumed in advance, a detection result indicating that the holding state of the substrate is normal is generated when both match, and when the two do not match, the holding state of the substrate is set. Holding state determining means for generating a detection result indicating that there is an abnormality.

  A second aspect of the present invention is the substrate detection apparatus according to the first aspect, wherein the image synthesizing unit includes pixel values of pixels at the same position in an image represented by each of the plurality of captured image data. A value determined based on a product of a maximum pixel value within a range less than a predetermined upper limit value and an exposure time for generating captured image data that gives the pixel value is a pixel value at a corresponding position of the composite image. The composite image data is generated.

  According to a third aspect of the present invention, there is provided the substrate detection apparatus according to the second aspect, wherein the image composition means determines a value determined based on a logarithmic value of a product of the maximum pixel value and the exposure time. The composite image data is generated so that the pixel value becomes the pixel value at the corresponding position.

  A fourth aspect of the present invention is the substrate detection apparatus according to the first aspect, wherein the image synthesizing unit averages pixel values of pixels at the same position in an image represented by each of the plurality of captured image data. The composite image data is generated so that the value becomes a pixel value at a corresponding position of the composite image.

  A fifth aspect of the present invention is the substrate detection apparatus according to the first aspect, wherein the image synthesizing unit is configured to center pixel values of pixels at the same position in an image represented by each of the plurality of captured image data. The composite image data is generated so that the value becomes a pixel value at a corresponding position of the composite image.

  A sixth aspect of the present invention is the substrate detection apparatus according to any one of the first to fifth aspects, wherein the holding position specifying means previously determines pixel values of pixels constituting an image of the substrate in the composite image. By performing an operation of integrating in the circumferential direction of the predetermined substrate along the predetermined arrangement direction of the substrate, peak profile data indicating a change in the integrated value with respect to the arrangement direction is generated, and then the peak A holding position of the substrate is specified based on a peak position appearing in profile data.

  A seventh aspect of the present invention is the substrate detection apparatus according to any one of the first to sixth aspects, wherein the detection execution position is a predetermined position inside a chamber that performs the predetermined processing in the substrate processing apparatus. In addition, the captured image generation means images the substrate held at the detection execution position via the chamber wall.

  According to an eighth aspect of the present invention, there is provided a substrate processing apparatus comprising: the substrate detection apparatus according to the seventh aspect; a substrate transfer unit that functions as the holding unit; and a chemical solution tank provided in the chamber. As the treatment, a chemical solution treatment is performed by immersing the substrate in the chemical solution tank, and the substrate detection device detects a holding state of the substrate held by the substrate transfer means after the chemical solution treatment. And

  According to invention of Claim 1 thru | or 8, the picked-up image production | generation means which produces | generates several picked-up image data by picking up the board | substrate with which illumination light was irradiated several times on different exposure conditions, and several picked-up image Since the image synthesizing unit that synthesizes the data to generate the composite image data and the holding position specifying unit that specifies the holding position of the substrate in the holding unit based on the composite image expressed by the composite image data, the illumination unit Decrease in the accuracy of specifying the holding position of the substrate due to the occurrence of light and darkness due to the holding position in the captured image obtained by imaging the substrate held by the holding means depending on the positional relationship between the holding position and the detection execution position And substrate detection with higher accuracy can be realized.

  In particular, according to the second and third aspects of the invention, since the holding position of the substrate can be specified with good specific accuracy regardless of the distance from the illumination means, the substrate detection with higher accuracy is realized. it can.

  In particular, according to the invention of claim 7, the detection execution position is a predetermined position inside the chamber where the predetermined processing is performed in the substrate processing apparatus, and the captured image generating means is detected via the chamber wall. Since the substrate held by the camera is imaged, it is difficult to arrange the illumination unit and the captured image generation unit in the processing tank. Therefore, the holding unit depends on the positional relationship between the illumination unit, the captured image generation unit, and the detection execution position. Even when the captured image obtained by imaging the substrate held on the substrate is bright and dark due to the holding position, the decrease in the accuracy of specifying the holding position of the substrate is suppressed, and the substrate with higher accuracy. Detection can be realized.

  In particular, according to the invention of claim 8, since the captured image generating means must be arranged outside the chamber because it is not preferable to be exposed to a chemical atmosphere, the illumination means and the captured image generation Even if the captured image obtained by imaging the substrate held by the holding means may be bright and dark due to the holding position depending on the positional relationship between the means and the detection execution position, the holding position of the substrate is specified. Since a decrease in accuracy is suppressed, substrate detection with higher accuracy can be realized.

<Outline of configuration of substrate processing apparatus>
FIG. 1 is a diagram schematically showing a substrate processing apparatus 100 according to an embodiment of the present invention. The substrate processing apparatus 100 mainly includes a cleaning device 101 and a substrate detection device 102. The substrate processing apparatus 100 includes a main host 103 that controls the entire apparatus in an integrated manner, but some of the constituent elements also function as constituent elements of the substrate detection apparatus 102 as described later.

  The cleaning apparatus 101 includes at least one chamber 110 having a chemical tank 111 (FIG. 1 illustrates the case where one chamber 110 is provided). The cleaning apparatus 101 immerses a plurality of substrates (hereinafter simply referred to as “substrates”) in a batch in a predetermined chemical solution SOL such as phosphoric acid or sulfuric acid water stored in a chemical solution tank 111 provided in the chamber 110. This is an apparatus for performing a so-called batch type cleaning process for cleaning a substrate.

  In addition, the cleaning apparatus 101 includes a transport unit (transport chuck) 114 that transports the substrate W between the chamber 110 and the outside (for example, another chamber 110). Although not shown in FIG. 1, the transport unit 114 has a plurality of holding grooves 115 (see FIG. 9), and holds the substrate W in an upright position by each holding groove 115. Transport multiple sheets at the same time. It is assumed that the transport unit 114 holds the substrate W in the posture a shown in FIG. 1 and does not hold the substrate W in the posture b. The chamber 110 is provided with a lifter 112 that moves up and down in the z-axis direction as indicated by an arrow AR1 while holding the substrate W. In the cleaning process in the chamber 110, the lifter 112 provided in the chamber 110 receives the substrate W transported to the predetermined substrate delivery position above the chamber 110 by the transport unit 114, and the lifter 112 holds the received substrate W. This is realized by descending into the chemical bath 111 and immersing the substrate W in the chemical solution SOL stored in the chemical bath 111 until a predetermined time elapses. After a predetermined time has elapsed, the substrate W that has been immersed in the chemical solution SOL is lifted from the chemical solution SOL when the lifter 112 is raised, transferred to the transfer means 114 at the transfer position, and transferred to the other by the transfer means 114. . The transport operation by the transport unit 114 and the lifting / lowering operation of the lifter 112 are realized by a predetermined drive unit (not shown) that operates based on the control of the transport control unit 170 provided in the main host 103. Further, an overflow tank 113 is provided on the outer periphery of the chemical liquid tank 111. The chemical liquid SOL overflowed by the immersion of the substrate W is collected in the overflow tank 113 and circulated and supplied to the chemical liquid tank 111 through a circulation path (not shown). Is done.

  In the substrate processing apparatus 100 according to the present embodiment, at least a part of the side surface (tank wall) of the chamber 110 is made of a light-transmitting material, for example, transparent acrylic, because it is necessary to perform the imaging process described below. Consists of.

  The substrate detection apparatus 102 is an apparatus that detects the holding state (substrate detection process) of the substrate W that has finished a predetermined cleaning process in each chamber 110. In the present embodiment, the detection of the holding state of the substrate W is to specify the holding position and the holding number of the substrate W that is lifted by the lifter 112 after being immersed in the chemical solution SOL in the chamber 110, and before the immersion. This is a process for checking whether or not the same number exists. In other words, it is a process for confirming that the substrate W has not dropped into the chemical solution tank 111 during the immersion in the chemical solution SOL. Specifically, after the cleaning process is performed in the chamber 110, when it is transferred from the lifter 112 to the transport unit 114 (in a state where the transport unit 114 takes the posture a and holds the substrate in FIG. 1). The edge portion of the substrate W is imaged, predetermined image processing is performed on the obtained captured image, and the presence or absence of the substrate is confirmed based on the obtained image processing result.

  The substrate detection apparatus 102 includes, as components responsible for imaging processing for substrate detection, an imaging unit 120 such as a CCD camera or a CMOS camera for imaging the substrate W, and illumination light to the substrate W during imaging. The illumination means 121 which consists of LED etc. for irradiating is provided.

  2 and 3 are diagrams for explaining the arrangement relationship between the imaging unit 120 and the illumination unit 121. FIG. 2 is a top view (xy plan view) of the substrate W held by the transfer means 114 in the upper part of the chamber 110 as seen from the positive z-axis direction (vertically upward), and FIG. 3 is the substrate W of FIG. FIG. 3 is a diagram (zx plan view) seen from the negative y-axis direction (viewed from the left side of FIG. 1). However, illustration of the chemical tank 111, the lifter 112, and the transport means 114 is omitted. 2 and 3, the one-dot chain line indicates the arrangement direction P of the substrates W held by the transport unit 114. That is, FIG. 2 and FIG. 3 show a case where the arrangement direction P of the substrates W coincides with the x-axis direction. Note that the arrangement relationship between the imaging unit 120 and the illumination unit 121 shown in FIG. 1 with respect to the chamber 110 is for convenience of illustration, and does not reflect the actual arrangement position in the substrate processing apparatus 100.

  As shown in FIG. 3, both of the imaging unit 120 and the illumination unit 121 are disposed on the outer upper side of the chamber 110. This is because it is necessary to keep the interior of the chamber 110 at a high degree of cleanliness, and to prevent the imaging unit 120 and the illumination unit 121 from being adversely affected by direct exposure to a chemical solution atmosphere. As described above, at least a part of the side surface of the chamber 110 is made of a light-transmitting material, for example, transparent acrylic so that imaging under such an arrangement relationship can be performed satisfactorily. Preferably, the chamber 110 or the chemical bath 111 is shielded by a shielding means (not shown) except when the substrate W is carried in and out.

  In addition, as shown in FIGS. 2 and 3, the imaging unit 120 and the illumination unit 121 are arranged so as to sandwich the chamber 110 when viewed in the x-axis direction. In the present embodiment, the illumination unit 121 is arranged in the zx plane including the arrangement direction P, whereas the imaging unit 120 is arranged at a location shifted from the plane. Actually, the arrangement position of the illumination unit 121 may be adjusted as appropriate in accordance with the arrangement position of the imaging unit 120.

  Returning to FIG. 1, the substrate detection apparatus 102 further includes a control unit 130 that performs control of imaging processing and image processing based on the captured image. The control unit 130 is electrically connected to the imaging unit 120, the illumination unit 121, and the main host 103.

  The control unit 130 includes an image input unit 140, an image processing unit 150, a host IF 160, and a memory M1.

  The image input unit 140 controls imaging by the imaging unit 120. Specifically, an imaging command (imaging control signal) is given to the imaging unit 120, and an image signal is acquired from the imaging unit 120 to generate captured image data. In the present embodiment, the captured image data is generated as multi-value gradation data in which each pixel takes any gradation value (pixel value) from 0 to 255. In the captured image represented by the captured image data, the edge portion of the substrate W is usually observed bright (white) and the other portion dark (black).

  The image processing unit 150 receives the captured image data generated by the image input unit 140 and performs processing for specifying the holding position of the substrate W held by the transport unit 114 based on the captured image data. By specifying the holding position, the number of substrates W held by the transport unit 114 is also specified. As a result of the processing in the image processing unit 150, the holding position of the substrate W actually held by the transport unit 114 at the time after being subjected to the cleaning process is specified (in this case, the number of held sheets is also specified by itself). Therefore, the processing in the image processing unit 150 corresponds to the substrate detection processing in the substrate detection apparatus 102. The obtained processing result is referred to as detection board information. The configuration of the image processing unit 150 and details of various processes relating to the generation of detection board information performed in the image processing unit 150 will be described later.

  The host IF 160 is a communication interface with the main host 103. An instruction to execute imaging by the imaging unit 120, which is given from the main host 103 in accordance with the conveyance timing in the conveyance unit 114, is given to the image input unit 140 via the host IF 160. Further, the result of the image processing in the image processing unit 150 is transmitted to the main host 103 through the host IF 160.

  The memory M1 is a storage area for storing various information related to the substrate detection process. The memory M1 acquires information necessary for the substrate detection processing in advance or as necessary during execution of the substrate detection processing, and stores this. Information stored in the memory M1 is appropriately read during the substrate detection process.

  When the cleaning apparatus 101 includes a plurality of chambers 110, the imaging unit 120 and the illumination unit 121 are provided corresponding to the respective chambers 110, and the control unit 130 centralizes all the imaging unit 120 and the illumination unit 121. Control.

  The main host 103 is a control unit that comprehensively controls the entire substrate processing apparatus 100. The main host 103 includes the transport control unit 170 that controls the operations of the lifter 112 and the transport unit 114 as described above, as well as the communication unit 180, the substrate information holding unit 190, the detection processing command unit 200, and the substrate presence / absence determination. Part 210.

  The communication unit 180 is a communication interface used when various signals are exchanged with the control unit 130 and other units of the substrate processing apparatus 100.

  The substrate information holding unit 190 is information regarding the number of substrates W transferred to the chamber 110 by the transfer unit 114 and the position of each substrate in a state where the maximum number of substrates that can be held by the substrate holding unit of the transfer unit 114 is held ( Hereinafter referred to as “processed substrate information”). In other words, the processing substrate information is information about the number of substrates W held and the holding position before the cleaning process. In other words, the information about the holding position of the substrate W before the cleaning process that the substrate information holding unit 190 has is the holding position (assumed holding) that is assumed as the holding position of the substrate W after the cleaning process. Position).

  The detection processing instruction unit 200 instructs the substrate detection apparatus 102 to execute a substrate detection process. Specifically, when a signal indicating that the substrate W has been transferred from the lifter 112 to the transport unit 114 is given from the transport control unit 170 after the cleaning process is completed, the detection processing command unit 200 controls the substrate detection apparatus. A signal for instructing execution of substrate detection processing is given to the unit 130.

  The substrate presence / absence determination unit 210 is configured so that all the substrates W subjected to the cleaning process are based on the processing substrate information held in the substrate information holding unit 190 and the detection substrate information generated by the image processing unit 150. It is determined whether or not it is held by the conveying means 114 after the cleaning process. That is, it is determined whether there is a substrate dropped in the chamber 110 for some reason. The substrate presence / absence determination unit 210 compares the holding position of the substrate specified by the substrate detection process with the holding position held in advance by the substrate information holding unit 190 as the holding position of the substrate. A detection result that the holding state of W is normal is generated, and if the two do not match, a detection result that there is an abnormality in the holding state of the substrate W is generated.

  The substrate information holding unit 190, the detection processing command unit 200, and the substrate presence / absence determination unit 210 are all components that perform processing related to substrate detection. In the present embodiment, the substrate information holding unit 190, the detection processing command unit 200, and the substrate presence / absence determination unit 210 function as components of the substrate detection apparatus 102. To do.

<Configuration of image processing unit>
FIG. 4 is a diagram showing a more detailed configuration of the image processing unit 150. The image processing unit 150 receives the captured image data generated by the image input unit 140, and specifies the holding position and the number of substrates W held by the transport unit 114 from the captured image data (that is, the transport unit 114). The holding state of the held substrate W is detected), and the result is output as detection substrate information. The detection of the holding state of the substrate W is performed based on the peak position in the one-dimensional data by generating one-dimensional data (peak profile data) for specifying the position of the substrate W from the captured image data.

  FIG. 5 is a diagram illustrating an example of the captured image data D1 of the substrate W captured by the imaging unit 120 and generated by the image input unit 140. In FIG. 5, an image of an edge portion of the substrate W (an image mainly including a reflected light component from the edge portion) is observed in white. Note that the two coordinate axes in the figure are specified and used in the substrate detection process described later, and the captured image data D1 itself does not necessarily have information on the coordinate axes concerned.

  FIG. 6 is a diagram illustrating an example of the one-dimensional data D2 generated by the image processing unit 150. The one-dimensional data D2 is integrated in a position coordinate (array direction coordinate) in the array direction of the substrate W specified by a method described later and the circumferential direction of the substrate W similarly specified by the method described later for each position coordinate. This is data indicating a relationship with pixel values (integrated pixel values) constituting the edge portion of the substrate W. In FIG. 6, the former is taken on the horizontal axis and the latter is taken on the vertical axis. Note that the orthogonal coordinate space shown in FIG. 6 corresponds to a transformation of the coordinate space consisting of the two coordinate axes shown in FIG. Since the peaks appearing in the one-dimensional data D2 shown in FIG. 6 correspond to the individual substrates W held by the transport unit 114, the holding position of the substrate W is specified by specifying each peak position. Will be.

  As shown in FIG. 4, the image processing unit 150 includes, as components for realizing the substrate detection processing, a detection power determination processing unit 151, an image composition unit 152, a detection processing unit 153, and a correction processing unit 154. And.

  The detection power determination processing unit 151 determines whether the captured image data generated by the image input unit 140 may be used for generating detection board information, that is, the captured image. It is responsible for processing (detection power determination processing) for determining whether or not the captured image represented by the data is an image having sufficient power for detecting the holding state of the substrate. For example, captured image data obtained in a state where dirt is attached to the chamber 110 may not have a contrast (sharpness) that can accurately specify the holding position of the substrate in the detection processing unit 153. The power determination processing unit 151 performs processing for distinguishing the captured image data in order to eliminate such poor quality captured image data and use only captured image data having a good contrast for the substrate detection process. Specifically, the power determination processing unit 151 digitizes the degree of image contrast in the captured image data D1, and determines that the image has power when the value exceeds a predetermined threshold. Details of the power determination process will be described later.

  The image composition unit 152 composes a plurality of captured image data obtained from the image input unit 140 by a predetermined method, and generates one image data (composite image data).

  In the substrate processing apparatus 100, since the illumination unit 121 and the substrate W are in the positional relationship shown in FIGS. 2 and 3, when the imaging unit 120 performs imaging by irradiating the substrate W with light from the illumination unit 121. In the obtained captured image, brightness and darkness occur in the image of the substrate W according to the distance from the illumination means 121. That is, since a large amount of reflected light is obtained from the substrate W close to the illumination unit 121, the image of the edge portion of the substrate W in the captured image is bright, but the amount of reflected light obtained from the substrate W far from the illumination unit 121 is small. Therefore, since the image of the edge portion of the substrate W becomes dark, the problem that the accuracy of specifying the position of the substrate W varies depending on the distance from the illumination unit 121 may occur. If the exposure amount and the exposure time are increased, the image of the substrate W far from the illumination unit 121 becomes bright, but in this case, the amount of reflected light from the substrate W near the illumination unit becomes too large and becomes saturated. Occurs.

  FIG. 7 is a diagram schematically showing changes in captured image data when the exposure amount is changed by changing the exposure time (that is, the shutter speed). The captured image data is originally two-dimensional data, but is shown here as one-dimensional data for the sake of simplicity. In FIG. 7, the larger the value on the horizontal axis, the farther from the illumination unit 121 in the arrangement of the substrates W held by the transport unit 114. In FIG. 7, the position PT1 indicates the position (edge position) of the edge portion of the substrate W on the side close to the illumination means 121, the position PT3 indicates the edge position of the substrate W on the side far from the illumination means 121, The position PT2 indicates the edge position of the substrate W located between the two. On the other hand, the position BG1 indicates the position between the edge portions of the substrate W on the side close to the illumination means 121 (inter-edge position), and the position BG3 indicates the position between the edges of the substrate W on the side far from the illumination means 121. , BG2 indicates the position between the edges of the substrate W in the middle of both.

  FIG. 7A shows a peak when the substrate W far from the illumination unit 121 is exposed with an appropriate exposure amount by relatively slowing down the shutter speed (exposure time = 1/30 seconds). It is a figure which shows a profile. In this case, the difference (peak bottom difference) between the pixel value at the position PT1 and the pixel value at the position BG1 is “5”, the difference “110” between the pixel value at the position PT2 and the pixel value at the position BG2, and the position It is significantly smaller than the difference “100” between the pixel value at PT3 and the pixel value at position BG3. This means that the amount of reflected light from the substrate W is saturated at the substrate position close to the illumination unit 121 due to overexposure, and a sufficient peak-bottom difference cannot be obtained.

  FIG. 7C shows a peak when the substrate W closer to the illumination means 121 is exposed with an appropriate exposure amount by relatively increasing the shutter speed (exposure time = 1/120 seconds). It is a figure which shows a profile. In this case, since the pixel value at the position PT3 which is the edge position far from the illumination unit 121 is as small as “50”, the peak bottom difference is as small as “40”.

  FIG. 7B shows a substrate in the central portion of the array by setting the shutter speed to an intermediate level between the case of FIG. 7A and the case of FIG. 7C (exposure time = 1/60 seconds). It is a figure which shows the peak profile at the time of making it expose with W by appropriate exposure amount. In this case, although the peak bottom difference can be obtained to some extent regardless of the position, the peak bottom difference at the position that is the edge position far from the illumination unit 121 is slightly small as “50”.

  In the substrate processing apparatus 100 according to the present embodiment, in order to solve such a problem, the situation in which the transport unit 114 holds the substrate W is imaged a plurality of times with different exposure amounts, and a plurality of captured images obtained. The image composition unit 152 synthesizes the data into one image data, and the synthesized image data generated thereby can be used as detection image data for subsequent processing. By using this composite image data for the detection processing of the substrate W, the difference in distance from the illumination unit 121 for each substrate W held by the transport unit 114 has an influence on the accuracy of specifying the holding position of the substrate W. It is suppressed. Details of the composite image data generation process (image composition process) will be described later.

  The detection processing unit 153 uses the captured image data obtained by the image input unit 140 or the synthesized image data obtained by the image synthesis unit 152 as detection image data, and uses the detection image data as shown in FIG. One-dimensional data D2 is generated, and processing for specifying a peak position in a peak profile represented by the one-dimensional data D2 is performed. The peak position specified by the detection processing unit 153 corresponds to the substrate holding position in the transport unit 114. Details of the method for generating one-dimensional data from the detection image data will be described later.

  The correction processing unit 154 changes the holding position of the substrate W in the transport unit 114 in the vertical direction (gravity direction) depending on the number of substrates W held (that is, depending on the load applied to the transport unit 114). Processing for correcting the deviation of the peak position of the one-dimensional data due to the above is performed.

  8, 9, 10 and 11 are diagrams for explaining the shift of the peak position of the one-dimensional data. FIG. 8 is a diagram illustrating a state in which the transport unit 114 holds the substrate W. Although FIG. 8 illustrates the case where the transport unit 114 has 16 holding grooves 115, the number of the holding grooves 115 provided in the transport unit 114 is not limited to this. FIG. 8A shows a case where 13 substrates W are held, and FIG. 8B shows a case where 3 substrates W are held. A state where a large number of substrates W are held (when holding a large number of substrates) is illustrated, and a state where the latter holds a relatively small number of substrates W (when a small number of substrates are held) is illustrated. In FIG. 8, each holding groove 115 is given a number for identification (substrate slot number).

  FIG. 9 is a diagram showing the correspondence between the number of substrates W held when the substrate W is held as shown in FIG. 8 and the holding position of the substrate W in the vertical direction (gravity direction). FIG. 9A corresponds to the case of holding a large number of substrates in FIG. 8A, and FIG. 9B corresponds to the case of holding a small number of substrates in FIG. As shown in FIG. 9, the center of gravity P2 of the substrate W in the latter is positioned higher in the vertical direction (gravity direction) than the center of gravity P1 of the substrate W in the former where the number of substrates held is large. This is because the latter has a smaller load on the conveying means 114.

  FIG. 10 is a diagram showing one-dimensional data obtained when the transport unit 114 holds the substrate W in the manner shown in FIG. In FIG. 10, the peak profile indicated by a solid line represents one-dimensional data when a large number of substrates W are held as shown in FIG. 8A, and the peak profile indicated by a broken line is shown in FIG. Represents one-dimensional data when a small number of substrates W are held. FIG. 11 is a diagram showing a determination result when the presence / absence of a substrate is determined without correcting the peak position based on the one-dimensional data of FIG. In FIG. 11, ◯ indicates a case where it is determined that a substrate is present in the holding groove 115 of the substrate slot number, and X is a case where it is determined that there is no substrate in the holding groove 115 of the slot number. Show.

  From the peak profile represented by the solid line in FIG. 10, it is assumed that the holding groove 115 corresponds from the peak position as indicated by a dotted line above FIG. The determination result shown in FIG. 11 also shows the content corresponding to this. That is, “0”, “2”, “3”, “4”, “5”, “7”, “8”, “10”, “11”, “12”, “13”, “14”, It is determined that the substrate W is held in the holding groove 115 with the substrate slot number “15”. This is a result that matches the actual holding state shown in FIG.

  On the other hand, from the peak profile represented by the broken line in FIG. 10, the substrate is inserted into the holding groove 115 to which the substrate slot numbers “9”, “10”, and “12” are attached as shown in FIG. It is determined that W is held. However, this is a result that does not match the actual holding state shown in FIG. The actual holding state shown in FIG. 8B is to hold the substrate W in the holding groove 115 to which the substrate slot numbers “7”, “8”, and “10” are attached. If the substrate W is held in the holding slot 115 having the substrate slot number, it is determined that the substrate W is held in the holding groove 115 having the substrate slot number “9”. This is because the peak derived from the substrate W held in the holding groove 115 having the substrate slot number of “7”, which should originally exist at the position x1 shown in FIG. 10, is the holding of the substrate slot number of “9”. This is due to the presence at the position x2 corresponding to the peak position when the substrate W is held in the groove 115.

  The results shown in FIGS. 10 and 11 mean that, as shown in FIG. 9, the holding groove 115 that is determined to hold the substrate W varies depending on the number of held sheets. In the substrate processing apparatus 100 according to the present embodiment, a case where a predetermined number of substrates W are held at the peak position specified by the detection processing unit 153 by the correction processing unit 154 in order to avoid such a problem. Correction processing for correcting as a reference can be performed. Details of the correction processing will be described later.

<Outline of processing in substrate detection apparatus>
FIG. 12 is a diagram showing a schematic flow of processing performed in the substrate detection apparatus 102.

  First, the substrate W after the cleaning process, which is the target of the detection process, is imaged, and captured image data is generated (step S1). When the captured image data is obtained, detection power determination processing is performed by the detection power determination processing unit 151 (step S2).

  If there is no problem as a result of the detection power determination process, the image composition unit 152 acquires image data for detection (step S3). When the imaging is performed only once, the captured image data generated by the image input unit 140 is used as detection image data as it is. On the other hand, as described above, in order to increase the accuracy of specifying the peak position in the one-dimensional data, composite image data is generated from a plurality of captured image data obtained by imaging with different exposure amounts by the imaging unit 120, The composite image data can also be used as detection image data.

  When the detection image data is obtained, the detection processing unit 153 generates one-dimensional data (detection one-dimensional data) used for specifying the holding position of the substrate W based on the detection image data. By specifying the position of the peak appearing in the one-dimensional data, a process (substrate detection process) for specifying the holding position of the substrate W held by the transport unit 114 is performed (step S4). Further, the correction processing unit 154 performs correction processing for correcting the peak position of the substrate W specified by the detection processing unit 153 (step S5). Thereby, the holding position of the substrate W is correctly specified. When the above-described shift in the peak position is not a problem and the holding position of the substrate can be correctly specified, the correction process is omitted and the peak of the substrate W specified by the detection processing unit 153 is omitted. It is also possible to specify the holding position from the position.

  Information on the holding position (and the number of sheets) of the substrate W held on the transport unit 114 specified in this way is sent to the main host 103 via the host IF 160 as detection substrate information. In the main host 103, the substrate presence / absence determination unit 210 compares the detected substrate information received from the communication unit 180 with the processing substrate information held by the substrate information holding unit 190, thereby determining whether or not there is a substrate at each holding position. Is determined (step S6). Specifically, the substrate presence / absence determination unit 210 has a predetermined threshold value based on the position where the holding position of the substrate W described in the detection substrate information is described as the holding position of the substrate in the processing substrate information. If it is included in the range, it is determined that the substrate W exists at the position.

  As a result, it is determined whether or not the substrate W subjected to the cleaning process is held by the transport unit 114 after the process.

<Imaging process>
FIG. 13 is a diagram illustrating a flow of the imaging process of the substrate W performed in step S1 of FIG.

  First, the substrate W that has been held in the lifter 112 and immersed in the chemical solution SOL in the chemical solution tank 111 and has been subjected to the cleaning process is pulled up from the chemical solution tank 111 as the cleaning process ends (step S11). Subsequently, the substrate W is transferred from the lifter 112 to the transport unit 114 (step S12). At this time, a signal indicating that the delivery has been completed is given from the transport control unit 170 to the detection processing command unit 200. In response to this, the detection processing command unit 200 issues an imaging command (step S13). The imaging command is transmitted to the host IF 160 in the control unit 130 through the communication unit 180, and the host IF 160 gives this to the image input unit 140.

  Receiving the imaging command, the image input unit 140 turns on the illumination unit 121 (step S14). From the illumination unit 121, illumination light having a predetermined light amount passes through the transparent portion on the side surface of the chamber 110 and is irradiated toward the substrate W held by the transport unit 114. When the illumination light is irradiated in this way, the edge portion (edge portion) on the upper side in the vertical direction (z-axis direction) of the substrate W reflects the illumination light toward the imaging means 120, and the main surface transmits the illumination light to the chamber 110. Is reflected in the bottom direction (negative z-axis direction in FIGS. 1 to 3).

  Under this situation, the image input unit 140 instructs the imaging unit 120 to image the substrate W (step S15). Specifically, the imaging unit 120 reads the number of times of imaging set in advance and stored in the memory M1 and the exposure amount (exposure time, shutter speed) at the time of each imaging, and executes the imaging based on the read conditions. To instruct.

  The imaging unit 120 images the substrate W a predetermined number of times with the designated exposure amount (step S16). When composite image data is used as detection image data, imaging is performed a plurality of times with different exposure amounts. An image signal (imaging signal) obtained by imaging by the imaging unit 120 is transmitted to the image input unit 140. The image input unit 140 generates captured image data based on the received image signal (step S17). As described above, since the reflected light at the edge portion of the substrate W is directed toward the imaging means 120, the edge portion of the substrate W is observed brightly in the captured image. The obtained captured image data is stored in the memory M1.

<Power detection processing>
Next, the power determination process performed in step S2 of FIG. 12 will be described in detail. As described above, the detection power determination process is a process of determining whether the captured image data is sufficient to specify the holding position of the substrate in the transport unit 114. Such power determination processing is executed in the power determination processing unit 151.

  FIG. 14 is a diagram illustrating the flow of the power determination process. When performing the power determination process, first, the captured image data generated in step S1 is acquired by the image processing unit 150 (step S21). The captured image data acquired in this way and used as a target of the power determination process is particularly referred to as power determination target data, and an image represented by the power determination target data is also referred to as a determination target image. .

  When the power determination target data is acquired, processing for enhancing the edge of the determination target image (edge enhancement processing) is performed (step S22). The edge of the determination target image refers to an area where the pixel value (density value) changes greatly in the image. In the present embodiment, this edge enhancement processing is performed by applying a general edge enhancement filter such as a Sobel filter to the power determination target data. This is a process for further clarifying the boundary between a bright part mainly including the reflected light component of the substrate W and a dark part mainly including other than the reflected light component of the substrate W in the determination target image. This corresponds to pre-processing for processing performed in steps.

  Next, a density histogram is generated based on the power determination target data that has been subjected to the edge enhancement processing in this way (step S23). The density histogram is obtained by taking all the pixel values (density values) that can be taken by the determination target image on the horizontal axis and counting the number of pixels (frequency) of the determination target image for each pixel value. FIG. 15 is a diagram illustrating density histograms for two different detection power determination target data. In FIG. 15, it is assumed that the larger the pixel value, the whiter (brighter).

  When the density histogram is obtained, the binarization threshold value and the class separation degree are calculated based on the density histogram (step S24). The binarization threshold is a threshold for binarizing a determination target image that provides a density histogram. The class separation degree is an index representing the degree of validity of the binarization threshold. The class separation degree indicates that the larger the value, the more the two classes are separated at the binarization threshold, in other words, the higher the bimodality of the pixel value histogram. In the present embodiment, the binarization threshold value and the class separation degree are calculated based on the discriminant analysis method.

Specifically, it is assumed that the determination target image is expressed by gradation values (levels) at L levels, and the probability of taking a pixel value (class) of level k or less when level k is a binarization threshold is ω. ₁ , where the probability of taking pixel values (classes) of k + 1 or more is ω ₂ , and the average value of the number of pixels for each class is μ ₁ and μ ₂ , k expressed by the following (Equation 1), A level k that maximizes the value (interclass variance) σ _B ² that is a function of is used as the binarization threshold.

However, if the number of pixels at level i is n _i , the total number of pixels is N, and the probability distribution at level i is P _i ,

It is. (Equation 2) is the 0th-order moment of the concentration distribution up to level k.

Using

It is. (Equation 4) is the first moment of the density distribution up to level k, and (Equation 5) is the average value of the number of pixels for all pixel values.

Further, when σ _B ² is maximum, the value given by the following (Equation 8) is the class separation degree. As a result of the calculation according to (Equation 8), the class separation degree is obtained as a value between 0 and 1.

However, σ _T ² is the variance of the number of pixels for all pixel values, and is given by the following (Equation 9).

Actually, by sequentially changing the value of k in (Equation 1), the value of σ _B ² is sequentially obtained, and k giving the maximum value is adopted as the binarization threshold. And class separation degree (eta) is calculated | required from (Equation 8) using the value of k at this time. In other words, the binarization threshold is determined so that the class separation degree η is maximized.

  Looking at the two pixel value histograms shown in FIG. 15, in the pixel value histogram of FIG. 15A, there are locations where the frequency (number of pixels) is maximized on the low pixel value side and the high pixel value side. The low threshold value side and the high pixel value side are clearly separated because the binarization threshold value is located where the intermediate frequency becomes minimum. In such a case, high class separation can be obtained. In the pixel value histogram, the higher the pixel value side, the brighter the pixel value. In the case of FIG. 15A, the pixels that appear to represent the edge portion of the substrate form the local maximum on the high pixel value side, and the other background portions are the low pixels. It seems to form the maximum on the value side. On the other hand, in the pixel value histogram of FIG. 15B, the frequency decreases monotonously as it goes to the higher pixel value side, and a considerable frequency is counted even in the neighboring pixel values of the binarization threshold. In such a case, the class separation will be low. In this case, in the captured image, the edge portion of the substrate W and other portions are not clearly separated.

  When the binarization threshold and the class separation degree are obtained, next, a sharpness value is calculated (step S25). The sharpness value is a value representing the degree of contrast of the determination target image. The sharpness value is calculated by multiplying the average value (referred to as edge average value) for pixel values having a value equal to or higher than the binarization threshold by the class separation degree.

  With regard to the two pixel value histograms shown in FIG. 15, although the class separation is different as described above, the edge average values are the same value m. Therefore, if the edge average value m is directly used as a criterion for determining the power, both cannot be identified. However, even in such a case, since the class separation degree is a different value reflecting the shape of each pixel value histogram, the value obtained by multiplying the edge average value by the class separation degree, that is, the sharpness value is used. By using the detection power criterion, the detection power can be determined more accurately.

  When the sharpness value is obtained, the power detection is performed (step S26). Specifically, the obtained sharpness value is compared with a reference value stored in advance in the memory M1 before performing the substrate detection process. Such a reference value can be obtained by, for example, assuming in advance a state that is an allowable limit of contamination of the chamber 110 due to the vaporized chemical solution SOL and the like, and calculating a sharpness value in such a state. Alternatively, the sharpness value in an ideal state may be calculated, and a value obtained by multiplying this value by a predetermined coefficient (allowable limit coefficient) may be used as the reference value.

  When the sharpness value is larger than the reference value (YES in step S26), it is determined that the detection power determination target data has a sufficient detection power for specifying the holding position of the substrate, so step S3 in FIG. Will proceed to. On the other hand, when the sharpness value is equal to or less than the reference value (NO in step S26), it is determined that the detection power determination target data does not have detection power. In this case, the subsequent substrate detection process is stopped. For example, when the detection result determination unit 151 notifies the main host 103 of the result through the host IF 160, the main host 103 performs predetermined processing such as cleaning the chamber 110 instead of instructing execution of the subsequent substrate detection processing. Predetermined warning processing (such as generation of warning sound or execution of warning display) is performed to prompt the user to take action.

  In the substrate processing apparatus according to the present embodiment, by performing such a detection power determination process, the quality of the captured image is not sufficient, and the captured image data that is not suitable for the detection of the substrate is used for the substrate detection process. Can be excluded from the target. Thereby, the substrate detection with high accuracy is realized. Also, since the captured image data with sufficient contrast is used for substrate detection, substrate detection is possible without creating an image that emphasizes the image of the edge portion of the substrate using the image data with the illumination light turned off. Is possible.

  Note that, when performing the power determination process, if the imaging unit 120 performs a plurality of times of image capturing, even if the first captured image data is the target of the power determination process, Alternatively, the detection power determination process may be performed for all or all captured image data. Alternatively, it is also possible to make a power determination on one piece of image data obtained by the combining process performed in S3.

  In addition, for example, when the quality of the captured image data is always kept good, such as when the wall surface of the chemical tank 111 is always kept normal, the power determination process is unnecessary. That is, the detection power determination process is not an essential process for specifying the substrate position, and may be omitted depending on circumstances. Information about whether or not the power determination process is necessary is given, for example, by an operator giving a predetermined input instruction to the main host 103, and stored in the memory M1.

<Image composition processing>
Next, the image composition process that is performed when the detection image data is acquired in step S3 in FIG. 12 will be described. The image composition processing is executed in the image composition unit 152. Note that a plurality of captured image data to be subjected to image synthesis processing is particularly referred to as synthesis target image data, and a captured image represented by each synthesis target image data is specifically referred to as a synthesis target image. An image expressed by the composite image data is particularly referred to as a composite image.

  Although several methods can be applied to the image composition processing, in this embodiment, a case where a method called a high dynamic range method (HDR method) is used will be described as an example. First, among the pixel values of pixels at the same position in each of a plurality of compositing target images (assuming values of 0 to 255), the maximum pixel value within a range less than a predetermined upper limit value Is identified. A value of about 240 to 250 is set as the upper limit value. The reason why the upper limit value is determined in this way is to exclude pixels in which the pixel value is saturated. Then, the true brightness is calculated using the specified pixel value, and a value obtained by normalizing the true brightness between 0 and 255, for example, is adopted as the pixel value at the pixel position of the compositing target image. Note that the exposure time e when imaging and the upper limit value for specifying the maximum pixel value are stored in the memory M1 in advance.

  FIG. 16 is a diagram showing a specific processing flow in the image composition processing. First, a plurality of compositing target image data to be subjected to image compositing processing is acquired (step S31). In the following description, it is assumed that a compositing target image represented by each compositing target image data is composed of N pixels, and an arbitrary pixel among them is designated as a jth pixel (1 ≦ j ≦ N ). First, j = 1 is set to process the first pixel (step S32).

  Hereinafter, a case where the process for the jth pixel is performed will be described. First, the maximum value o (j) within the range of the upper limit value is specified from the pixel values of the j-th pixel of each compositing target image, and exposure when the compositing target image that gives the maximum value is captured. Time e (j) is read from the memory M1 (step S33).

  Subsequently, the true brightness i (j) of the jth pixel is calculated from the following relational expression (step S34).

  Such processing is performed for all N pixels (step S35). That is, when there is a pixel for which the true brightness i is not obtained, the processes of step S33 and step S34 are repeated for the j + 1-th pixel (step S38).

  When true brightness i (j) is obtained for all N pixels (YES in step S35), processing for normalizing the value of i (j) is performed for all pixels (step S36). ). For example, the maximum value of i (1) to i (N) i (j) is iM, the minimum value is im, and the normalized pixel value of i (j) is I (j). Then, I (j) can be determined by the following arithmetic expression.

  The value of I (j) obtained in this way represents the pixel value in the jth pixel of the composite image. That is, the composite image data is generated by associating the arrangement position of each pixel with I (j) for each pixel (step S37).

  FIG. 17 is a diagram illustrating a result in a case where composite image data is generated based on three different captured image data schematically illustrated as one-dimensional data in FIG. FIG. 17A shows the result when the high dynamic range method is applied. Note that FIG. 17B is a generation example of composite image data in a modified example described later. Also in FIG. 17, it is assumed that the larger the value on the horizontal axis, the farther from the illumination unit 121 in the arrangement of the substrates W held by the transport unit 114.

  In the composite image data shown in FIG. 17A, the peak bottom difference is “148” at a position close to the illumination unit 121, and the peak bottom difference is “85” at a position far from the illumination unit. It can be seen that a relatively good peak-bottom difference is obtained regardless of the position, in any case.

  As described above, since the composite image data is generated as having a good peak-bottom difference, more accurate substrate detection can be realized by using the composite image data as a target of detection processing. In addition, since detection image data having a peak-bottom difference is used for substrate detection, an image in which the image of the edge portion of the substrate is emphasized using image data in a state where the illumination light is turned off is not created. In both cases, the substrate can be detected.

<Substrate detection processing>
Next, a detailed description will be given of the substrate detection process (strictly, the process of specifying the substrate position in the transport unit 114) performed in step S4 of FIG. FIG. 18 is a diagram showing the flow of the substrate detection process. First, for example, detection image data such as captured image data D1 shown in FIG. 5 is acquired (step S41). As described above, the detection image data may be an aspect using one captured image data obtained by the imaging process, or may be an aspect using composite image data. An image expressed by the detection image data is particularly referred to as a detection image.

  Next, one-dimensional data (peak profile data) is generated based on the detection image data (step S42). The generation of such one-dimensional data is performed by using the pixel value of a pixel (a pixel constituting a white portion in the case of FIG. 5) determined to constitute an image of the substrate W (image of an edge portion) in the detection image. This operation is performed along the arrangement direction of the substrates W, and the obtained integrated value is associated with the coordinates in the arrangement direction. For example, one-dimensional data D2 as shown in FIG. 6 is generated.

  In the present embodiment, the circumferential direction and the array direction used for generating this one-dimensional data are determined in advance using a reference detection image (reference detection image), and in the detection process, It is assumed that one-dimensional data is generated based on the predetermined circumferential direction and arrangement direction, regardless of the detection image data.

  FIG. 19 is a diagram showing a flow of processing for determining the arrangement direction and the circumferential direction based on the reference detection image. 20 and 21 are diagrams schematically showing images handled in such processing.

  First, in a state in which the substrate W is stored in all the holding grooves 115 of the transport unit 114 (all stored state), imaging by the imaging unit 120 is performed in the same manner as the imaging process described above, thereby providing a reference detection image. Image data (reference detection image data) is generated (step S421). FIG. 20A is a schematic diagram of the reference detection image IM1. The reference detection image IM1 is usually a multi-value gradation image in which the edge portion of the substrate W is observed to be bright (white) and the other portion is dark (black).

  When the reference detection image data is generated, brightness correction (step S422) is performed so that the edge portion of the substrate W is easily extracted, and the reference detection image is binarized (step S423). FIG. 20B is a schematic diagram of a lightness-corrected image IM2 obtained by performing lightness correction on the reference detection image IM1 shown in FIG. Further, FIG. 20C is a schematic diagram of a binarized image IM3 obtained by performing binarization on the lightness-corrected image IM2 shown in FIG. 20B.

  When the binarized image IM3 is obtained, a labeling process for uniquely identifying a closed region having a high pixel value (a region that appears white in the binarized image IM3) is performed (step S424). A known technique can be applied to the labeling process.

  The labeled region (labeling region) mainly represents the edge portion of the substrate W (this is referred to as an edge portion label region), but there is also a noisy region included in the original reference detection image. (This is referred to as a noise label region). Since such a noise label region is unnecessary for specifying the edge portion, a filtering process for removing it (replacement with a low pixel value) is performed on the binarized image IM3 (step S425). The filtering process is performed based on the position and size of each label area, the main axis direction, and the like. FIG. 20D is a schematic diagram of the filtered image IM4 obtained by performing the filtering process on the binarized image IM3 illustrated in FIG. As illustrated in FIG. 20D, in the post-filtering image IM4, only the edge label region exists as a region having a high pixel value.

When the filtered image IM4 is obtained, the barycentric position of each edge label area is specified (step S426). The coordinates of pixels belonging to a certain edge part label area R are (x, y), the total number of pixels belonging to the edge part label area is M, and the center of gravity of the edge part label area is (x _g , y _g ). Then, the center of gravity of the edge portion label area is obtained by the following relationship.

However, R represents the label area | region used as the object which calculates | requires a gravity center.

  When the center of gravity of each edge label area is obtained, the arrangement direction of the substrates W held by the transport unit 114 is specified. Specifically, the centroid position of each edge portion label area is obtained, and a straight line (regression line) that approximates the distribution of each centroid position is specified as a straight line that gives the arrangement direction of the substrates W (step S427). The regression line can be obtained by a known method such as a least square method. FIG. 21A illustrates a straight line y = ax + b indicating the arrangement direction obtained based on the distribution of the centroids G1 to G7 of the label regions L1 to L7 existing in the filtered image IM4 illustrated in FIG. It is a figure to do.

When the arrangement direction is specified, the circumferential direction (main axis direction) of the substrate W held by the transport unit 114 is specified. Specifically, after obtaining the angle (main axis angle θ) representing the main axis direction for each edge part label area, a value representative of the main axis angle θ for all edge part label areas is conveyed. An angle that defines the circumferential direction of the substrate W held by the means 114 (referred to as a circumferential angle θ ₀ ) is determined (step S428).

  The principal axis angle θ for each edge portion label area is obtained as follows.

  However,

It is. When the principal axis angle θ for all the edge part label regions is obtained, the circumferential direction angle can be defined by adopting, for example, an average value or a median value thereof.

Information on the arrangement direction and the circumferential direction obtained as described above (specifically, the slope a and the intercept b of the straight line y = ax + b specifying the arrangement direction, and the circumferential angle specifying the circumferential direction) θ ₀ ) is stored in the memory M1.

Using the arrangement direction and the circumferential direction specified in this way, the pixel values of the pixels determined to constitute the image of the edge portion of the substrate W in the detection image are integrated in the direction of the circumferential angle θ _0. By performing the operation along the straight line y = ax + b determined by specifying the arrangement direction, the generation of the one-dimensional data in step S42 shown in FIG. 18 is realized.

  However, since the edge portion in the detection image has a curved shape such as a bow shape or a crescent shape, if all pixels constituting the edge portion are to be integrated, one-dimensional data may give an unnecessary peak. Can happen. In the present embodiment, in order to avoid this, the integration of pixels in the circumferential direction is limited to a predetermined range in which the edge portion is linearly approximated. FIG. 22 is a diagram illustrating the integration range of such pixels. FIG. 22 shows a case where the integration of pixel values in the circumferential direction is performed only for a region (shaded portion) belonging to the range α centered on a straight line y = ax + b that specifies the arrangement direction. Yes. Note that the integrated range is preferably set so that at least the notch portion of the substrate W is included.

  When the one-dimensional data as shown in FIG. 6 is obtained by the process as described above, the holding position of the substrate W in the transport unit 114 can be specified by specifying the peak position. However, in the one-dimensional data, in addition to the components derived from the reflected light from the edge portion of the substrate W, reflected light components such as disturbances such as light transmitted from the outside into the chamber 110 and scattered light at the time of imaging, etc. Due to the presence, noise components are superimposed on the background portion. In the present embodiment, for the purpose of specifying a more accurate peak position, a process of removing a background component corresponding to background noise from the obtained one-dimensional data is performed (step S43).

  FIG. 23 is a diagram showing a flow of processing for removing background noise. First, for each data point on the one-dimensional data, a local region (predetermined range in the arrangement direction coordinates) centered on the data point is set, and the data point is determined from the pixel values belonging to the range of the local region. A pixel value (background pixel value) corresponding to the background noise is determined (step S431). Further, as a method for determining the background pixel value for each local region, it is a preferable example to use the minimum value of each pixel value in the local region.

  FIG. 24 is a diagram for explaining a method for determining a background pixel value using a partially enlarged version of the one-dimensional data D2 shown in FIG. In FIG. 24, the size d of the local region is set to be approximately the same as the peak width of the peak appearing in the one-dimensional data. The size d may be adjusted as appropriate.

FIG. 24 illustrates background pixel values a _bk , b _bk , and c _bk for three data points A, B, and C. The data point A is a data point that does not include a peak in the local region. Data point B is a data point at the peak position. Data point C is a data point including a part of a peak in a local region. Although there are significant differences in the accumulated pixel values at these data points, the differences between the background pixel values are much smaller.

  When the background pixel value is obtained, the background pixel value is subtracted from the pixel value for each data point (step S432).

  By plotting the subtraction values obtained for all the data points of the one-dimensional data, one-dimensional data (one-dimensional data for detection) from which background noise has been removed is generated (step S433).

  FIG. 25 is a diagram illustrating a case where the one-dimensional data for detection is generated by setting the size d of the local region to be approximately equal to the peak width for the one-dimensional data D2 illustrated in FIG. FIG. 25A is a diagram showing background pixel value data Db indicating the relationship between the background pixel value obtained based on the one-dimensional data for detection and the coordinate value in the arrangement direction coordinate. That is, FIG. 25A represents the background noise in the one-dimensional data D2 shown in FIG. FIG. 25A shows that the background noise can be understood as a smooth curve with little fluctuation. On the other hand, FIG. 25B is a diagram showing detection one-dimensional data D3 which is one-dimensional data obtained by subtracting the background pixel value at each data point from the one-dimensional data D2 shown in FIG. is there. When FIG. 6 is compared with FIG. 25B, in FIG. 25B in which the detection one-dimensional data D3 is plotted, background noise is suitably removed, and peaks are clearly observed in the entire range. I understand that.

  FIG. 26 and FIG. 27 are diagrams showing the results when the size d of the local region is changed for the one-dimensional data D2 shown in FIG.

  FIG. 26 is a diagram illustrating a result when the size d of the local region is set smaller than the peak width. FIG. 26A is a diagram showing background noise. Unlike the case shown in FIG. 25A, fine fluctuations appear. This is a result which means that the influence of the peak appears in the background pixel value depending on the data point. Accordingly, in the background pixel value subtraction result shown in FIG. 26B, not only the background noise component but also the peak component is subtracted, and the peak becomes small.

  FIG. 27A is a diagram showing a result when the size d of the local region is set larger than the peak width. As shown in FIG. 27 (a), the background noise is captured as a smooth curve, but the level of the background pixel value itself is smaller than that shown in FIG. 25 (a). Accordingly, the background noise component still remains in the one-dimensional data after the background pixel value removal shown in FIG. That is, this indicates that background noise has not been sufficiently removed.

  These results indicate that it is preferable that the size d of the local region is approximately the same as the peak width.

  Returning to FIG. 18, the peak position of the obtained one-dimensional data for detection D3 is detected (step S44). Specifically, the peak position is specified by differentiating the peak profile expressed by the obtained one-dimensional data for detection D3 and obtaining the maximum point. The peak position obtained by such processing corresponds to the position of the substrate W held by the transport unit 114. The obtained peak position information is stored in the memory M1 as detection board information.

  As described above, in the present embodiment, the substrate arrangement direction and the circumferential direction are obtained in advance, and the pixel values of the pixels that are determined to constitute the image of the edge portion of the substrate are set in the circle. Since the one-dimensional data for detection is generated by performing the operation of accumulating in the circumferential direction along the arrangement direction, each time the captured image data is obtained, the arrangement direction of the substrate and the circumferential direction are determined from the contents of the captured image data. Need not be calculated, and the substrate detection process is simplified. Thereby, for example, even when the obtained image of the edge portion has a complicated shape and it is complicated to directly specify the arrangement direction and the circumferential direction, such arrangement direction Necessary one-dimensional data can be generated without specifying the circumferential direction. At this time, by performing the integration range only within the range in which the image of the edge portion is linearly approximated in the circumferential direction, it is possible to suppress unnecessary peaks from appearing in the one-dimensional data for detection.

  Further, by generating one-dimensional data for detection from which background noise is removed, the peak position corresponding to the holding position of the substrate can be specified more accurately. In addition, the background noise is removed without creating an image that emphasizes the image of the edge portion of the substrate using the image data in a state where the illumination light is turned off.

<Correction process>
Finally, the correction process in step S5 of FIG. 12 will be described. As described above, such correction processing is performed by the transport unit 114 due to the shift of the holding position of the substrate W as shown in FIG. 9 depending on the total weight of the substrate W held by the transport unit 114. This is a process for correcting the shift of the peak position of the one-dimensional data for detection corresponding to each substrate W held in each holding groove 115.

  This correction process is based on the assumption that all the holding positions assumed for the individual substrates W held in the holding grooves 115 vary linearly only depending on the total weight of the substrates W. This is the process to be performed. Since the weight of the individual substrates W to be held is considered to be constant, after all, this assumption depends only on the number of the substrates W held by the transport means 114. Equivalent to the assumption that

  FIG. 28 shows the peak position specified by the detection processing unit 153 for the substrate W held in the holding groove 115 with the transport unit 114 under this assumption (this is particularly referred to as a target substrate), and the transport unit. FIG. 11 is a diagram showing a correspondence relationship with the number of substrates W held by 114.

  As shown in FIG. 28, when the transport means 114 holds N1 and N2 (N1 ≠ N2) substrates W, the holding positions (positions in the arrangement direction coordinates) for the target substrate are X1 and X2, respectively. Assuming that the holding position when holding the Nk substrates W is Xk, the value of Xk is obtained by the following equation.

  For example, when the transport unit 114 has N holding grooves 115, N1 = 2 (corresponding to the case where the substrate W is held only in the holding grooves 115 at both ends), N2 = N (all holdings) When the holding positions X1 and X2 in each case are obtained in advance from one-dimensional data, the position Xk obtained from the above formula is an arbitrary Nk number of substrates. When W is held, the holding position of the target substrate is indicated.

  Further, in this case, the difference value Xk−X2 (actually a value obtained from the first term on the right side of Equation 14) is the holding position of the target substrate when N substrates W are held, and Nk This represents a deviation from the holding position when the substrate W is held.

  That is, the peak position for the one-dimensional data when the Nk substrates W are held, which is specified by the detection processing unit 153, is the above-described peak position when the N substrates W are held. Therefore, the substrate W is held in all the holding grooves 115 by subtracting the above-described difference value from the coordinate value indicating the specified peak position. Thus, a coordinate value indicating a position assumed to be detected is obtained.

  The actual holding position of the target substrate (that is, the position of the holding groove 115 holding the target substrate or the substrate slot number) in a state where the substrate W is held in all the holding grooves 115 is one-dimensional data. Since the position X2 always corresponds to the position X2 obtained from the above, if the above difference value is subtracted from the position Xk of the target substrate when any Nk substrates W are held, the position Xk is detected. It is possible to reliably identify the holding position at which the peak indicates the presence of the substrate W. That is, the calculation process corresponds to a correction process for the substrate of interest.

  However, in the case where the substrate is held in each holding groove 115, it is complicated to make the correction by using the equation (15). Therefore, in the present embodiment, the above-described relationship is applied only to the correction of the peak position for the substrate W (these are referred to as correction reference substrates) held in two specific holding grooves 115 of the transport unit 114. The substrate W (target substrate) held in the other holding groove 115 is based on the positional relationship between the two correction reference substrates and the target substrate and the corrected holding positions of the two correction reference substrates. The holding position is corrected. Note that the position of the holding groove 115 where the correction reference substrate is held is not particularly limited, but in view of the accuracy of correction, a mode in which the holding grooves 115 at both ends are selected is most preferable. Therefore, in the following description, a case where the correction reference substrate is held at both ends of the N holding grooves 115 will be described as an example. In this case, the target substrate is the substrate W held in the holding groove 115 between the holding grooves 115 held between the two correction reference substrates.

  FIG. 29 is a diagram showing a flow of correction processing in such a case. FIG. 30 is a diagram showing a correspondence relationship between the holding positions of the two correction reference substrates and the target substrate held by the transport unit 114 and the number of substrates W held by the transport unit 114. In FIG. 30, the transport means 114 having N holding grooves 115 holds a total of two substrates W in the holding grooves 115 at both ends (that is, when only the correction reference substrate is held). The peak positions for the correction reference substrate held at the first end of the holding groove 115 are each Xa1, and the peak positions for the correction reference substrate held at the second end are each Xb1. Further, when the transport unit 114 holds the substrate W in all the holding grooves 115, the peak positions of the correction reference substrates held at the first end of the holding groove 115 are Xa2, respectively. It is assumed that the peak positions for the correction reference substrates held at the ends are Xb2.

  In the correction processing, first, prior to the actual processing, the detection processing unit 153 performs the substrate detection processing when these two substrates are accommodated and when all the N substrates are accommodated. Thus, the peak positions Xa1, Xb1, Xa2, and Xb2 are obtained in advance, and the results are stored in the memory M1 (steps S51 and S52). The values of Xa1, Xb1, Xa2, and Xb2 are repeatedly used when substrate detection is performed under the same conditions for the substrates W that are transported by the same transporting unit 114.

  When actually executing the correction process, first, information on the peak position corresponding to the holding position of the substrate W in the holding state, which is the target of the substrate detection process, specified by the detection processing unit 153 is acquired. (Step S53). In this case, information on the number of stored sheets Nk is also acquired from the number of peak position data (that is, the number of peaks).

  Here, as shown in FIG. 30, the peak position of the target substrate held in an arbitrary holding groove 115 when Nk substrates are held is Xk, and held at the first end and the second end. The peak positions of the corrected reference substrate are Xak and Xbk, respectively, and the peak position of the target substrate when the substrate W is held in all the holding grooves 115 is Xk ′.

  The peak position Xak of the correction reference substrate held at the first end is a point on the line segment La obtained by setting X1 = Xa1 and X2 = Xa2 in Equation 14, and is held at the second end. The peak position Xbk of the correction reference substrate is a point on the line segment Lb obtained by setting X1 = Xb1 and X2 = Xb2 in Equation 14.

  On the other hand, the peak position of the target substrate is also assumed to change linearly. The peak position Xk when the Nk substrates W are held exists between the position Xak and the position Xbk, while the peak position Xk ′ when the substrates W are held in all the holding grooves 115 is Since it exists between the position Xa2 and the position Xb2, it is reasonable to think that the line segment connecting the two indicates a change in the peak position for the substrate of interest. This means that the following proportional relationship holds.

  From this equation, the peak position Xk 'is obtained as follows.

  Since the right side is all known, by substituting the respective values, the substrate W that gives the peak position Xk when the transport means 114 holds Nk substrates, and the substrate W is placed in all the holding grooves 115. The peak position Xk ′ when it is held is specified (step S54). The specified peak position Xk ′ always corresponds to the actual holding position in a state where the substrate W is held in all the holding grooves 115 (contains within a certain threshold range based on the actual holding position. Therefore, by specifying the peak position Xk ′ from the above equation for all the acquired peaks, it is possible to specify which holding groove of the transfer means 114 holds the Nk substrates W ( Step S55).

  As described above, when the peak position specified based on the one-dimensional data indicated by the broken line in FIG. 10 is represented as it is as the holding position of the substrate W without correction, as shown in FIG. Although the holding position is erroneously determined, the correct holding position can be specified by performing the correction processing as described above for the specified peak position. That is, improvement in the accuracy of specifying the holding position of the substrate is realized.

<Modified example of composite image processing>
In the above-described embodiment, the composite image data is generated using the high dynamic range method (HDR method), but according to the equation (10) used at that time, the pixel for each composition target image When the maximum value o (j) is constant, the true brightness i (j) increases as the exposure time e (j) decreases. In general, the smaller the exposure time, the worse the accuracy of the captured image. Therefore, when many pixel values of the compositing target image obtained with a short exposure time are adopted as o (j), as a result, Accuracy will deteriorate. Therefore, in order to solve such a problem, the common logarithm value log ₁₀ i (j) of i (j) may be calculated instead of i (j), and the obtained value may be normalized.

  Alternatively, instead of the high dynamic range method, an average value of pixel values of pixels at the same position in each of a plurality of composition target images may be set as a pixel value at a corresponding position of the composite image. FIG. 17B is an example of generating composite image data in such a case.

  Or you may make it make the median value of the pixel value of the pixels in the same position in each of several synthetic | combination object images be a pixel value in the corresponding position of a synthesized image.

  In the above-described embodiment, the composite image data is generated using a plurality of captured image data captured with different exposure amounts. Instead, the exposure amount is kept constant, and the illumination unit 121 The aspect which produces | generates synthetic | combination image data using the some picked-up image data which adjusted the light quantity of this illumination light and imaged with a different light quantity may be sufficient. In this case, in the above-described embodiment, the control of the imaging unit 120 in the image input unit 140 is simplified.

It is a figure which shows typically the substrate processing apparatus 100 which concerns on embodiment of this invention. It is a figure for demonstrating the arrangement | positioning relationship between the imaging means 120 and the illumination means 121. FIG. It is a figure for demonstrating the arrangement | positioning relationship between the imaging means 120 and the illumination means 121. FIG. 2 is a diagram illustrating a more detailed configuration of an image processing unit 150. FIG. It is a figure which shows an example of the captured image data D1 of the board | substrate W. FIG. It is a figure which shows an example of the one-dimensional data D2. It is a figure which shows typically the change of the captured image data at the time of changing exposure amount by changing exposure time (shutter speed). It is a figure explaining the shift | offset | difference of the peak position of one-dimensional data. It is a figure explaining the shift | offset | difference of the peak position of one-dimensional data. It is a figure explaining the shift | offset | difference of the peak position of one-dimensional data. It is a figure explaining the shift | offset | difference of the peak position of one-dimensional data. It is a figure which shows the schematic flow of a board | substrate detection process. It is a figure which shows the flow of the imaging process of the board | substrate W. It is a figure which shows the flow of a detection power determination process. It is a figure which illustrates the density | concentration histogram about two different detection power determination object data. It is a figure which shows the flow of the specific process in an image composition process. It is a figure which shows the result at the time of producing | generating synthetic | combination image data based on three different captured image data. It is a figure which shows the flow of a board | substrate detection process. It is a figure which shows the flow of the process which determines an arrangement | sequence direction and a circumferential direction based on the image for a reference | standard detection. It is a figure which shows typically the image handled in the process which determines an arrangement | sequence direction and a circumferential direction. It is a figure which shows typically the image handled in the process which determines an arrangement | sequence direction and a circumferential direction. It is a figure which illustrates the integration range of the pixel at the time of determining an arrangement direction. It is a figure which shows the flow of the process which removes background noise from one-dimensional data. It is a figure explaining the method of determining a background pixel value. It is a figure which shows the case where the size of a local region is made into the same grade as a peak width, and the one-dimensional data for a detection are produced | generated. It is a figure which shows the result at the time of changing the size of a local region. It is a figure which shows the result at the time of changing the size of a local region. FIG. 10 is a diagram illustrating a correspondence relationship between a peak position specified by a detection processing unit 153 and the number of substrates W held by a transport unit 114 for a target substrate held in a holding groove 115 provided with a transport unit 114. It is a figure which shows the flow of a correction process. FIG. 6 is a diagram illustrating a correspondence relationship between the holding positions of two correction reference substrates and a target substrate held by a transport unit 114 and the number of substrates W held by the transport unit 114.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Substrate processing apparatus 101 Cleaning apparatus 102 Substrate detection apparatus 103 Main host 110 Chamber 111 Chemical solution tank 112 Lifter 113 Overflow tank 114 Conveying means 115 Holding groove 120 Imaging means 121 Illuminating means 130 Control part D1 Captured image data D2 One-dimensional data D3 For detection One-dimensional data Db Background pixel value data G1 to G7 Center of gravity IM1 Image for reference detection IM2 Image after brightness correction IM3 Binary image IM4 Image after filtering L1 to L7 Label region SOL Chemical solution W Substrate

Claims (8)

An apparatus for detecting a holding state of a substrate held by a predetermined holding unit capable of holding a plurality of substrates in a substrate processing apparatus that performs predetermined processing on a substrate,
Illuminating means for irradiating illumination light to the substrate held by the holding means at a predetermined detection execution position;
Picked-up image generation means for generating a plurality of picked-up image data by picking up the substrate irradiated with the illumination light a plurality of times under different exposure conditions;
Image combining means for combining the plurality of captured image data to generate combined image data;
Holding position specifying means for specifying the holding position of the substrate in the holding means based on the composite image represented by the composite image data;
The holding position of the substrate specified by the holding position specifying means is compared with an assumed holding position that is assumed in advance as the holding position of the substrate, and when both match, the holding state of the substrate is normal. Generating a detection result, and if the two do not match, holding state determination means for generating a detection result that the holding state of the substrate is abnormal;
The board | substrate detection apparatus characterized by the above-mentioned. It is a board | substrate detection apparatus of Claim 1, Comprising:
The image synthesizing unit is configured to capture a maximum pixel value and a pixel value within a range less than a predetermined upper limit among pixel values of pixels at the same position in an image represented by each of the plurality of captured image data. Generating the composite image data so that a value determined based on a product of the exposure time for generating the image data is a pixel value at a corresponding position of the composite image;
A substrate detection apparatus. It is a board | substrate detection apparatus of Claim 2, Comprising:
The image composition means generates the composite image data such that a value determined based on a logarithmic value of a product of the maximum pixel value and the exposure time is a pixel value at the corresponding position of the composite image;
A substrate detection apparatus. It is a board | substrate detection apparatus of Claim 1, Comprising:
The composite image data is such that the image composition means has an average value of pixel values of pixels at the same position in an image represented by each of the plurality of captured image data as a pixel value at a corresponding position of the composite image. Generate
A substrate detection apparatus. It is a board | substrate detection apparatus of Claim 1, Comprising:
The composite image data is set so that the median of the pixel values of the pixels at the same position in the image represented by each of the plurality of captured image data is the pixel value at the corresponding position of the composite image. Generate
A substrate detection apparatus. A substrate detection apparatus according to any one of claims 1 to 5,
The holding position specifying unit performs an operation of accumulating pixel values of pixels constituting an image of the substrate in the composite image in a predetermined circumferential direction of the substrate along a predetermined arrangement direction of the substrates. By generating peak profile data indicating a change in integrated value with respect to the arrangement direction, and specifying the holding position of the substrate based on the peak position appearing in the peak profile data,
A substrate detection apparatus. The substrate detection apparatus according to any one of claims 1 to 6,
The detection execution position is a predetermined position inside a chamber that performs the predetermined processing in the substrate processing apparatus, and the captured image generation unit is held at the detection execution position via a wall of the chamber. Image the board,
A substrate detection apparatus. A substrate detection apparatus according to claim 7,
A substrate transfer means that functions as the holding means;
A chemical bath provided in the chamber;
With
As the predetermined processing, chemical processing is performed by immersing the substrate in the chemical bath, and the substrate detection device detects the holding state of the substrate held by the substrate transporting means after the chemical processing.
A substrate processing apparatus.