Shape recognition method and shape recognition apparatus

Abstract

When recognizing the shape of a semiconductor wafer, it possesses robustness that makes it less susceptible to noise caused by edges, and allows for quick and easy shape recognition. [Solution] A shape recognition method performed by a shape recognition device that solves the above-mentioned problems, comprising: an acquisition step of acquiring coordinate values ​​in a first direction and a second direction in an image of the upper surface of an elliptical shape recognition target; a center coordinate calculation step of calculating the center coordinate of the ellipse based on the distribution of coordinate values; and an axis length calculation step of calculating the axis length of the ellipse based on the distribution of coordinate values.

Description

Technical Field

[0001] This embodiment relates to a shape recognition method and a shape recognition apparatus.

Background Art

[0002] Conventionally, a shape recognition method in an apparatus for determining the shape and position of an ellipse of a semiconductor wafer has been disclosed, which is a technique of extracting coordinate values of an edge portion of the semiconductor wafer.

[0003] However, in the prior art, it may be difficult to accurately recognize the shape of a semiconductor wafer because it is greatly affected by noise due to noise added to the edge.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] One of the objectives of this embodiment is to provide a shape recognition method and a shape recognition apparatus that have robustness against the influence of noise by an edge and can quickly and simply recognize a shape.

Means for Solving the Problems

[0006] According to one embodiment, there is provided a shape recognition method executed by a shape recognition apparatus, including: an acquisition step of acquiring coordinate values in a first direction X and a second direction Y in an image obtained by imaging an upper surface of an object to be recognized having an elliptical shape; a center coordinate calculation step of calculating a center coordinate of the elliptical shape based on the distribution of the coordinate values; and an axis length calculation step of calculating a length of an axis of the elliptical shape based on the distribution of the coordinate values.

Brief Description of the Drawings

[0007] [Figure 1] Figure 1 is a block diagram showing an example of the hardware configuration of a shape recognition device according to an embodiment. [Figure 2] Figure 2 is a block diagram showing an example of the configuration of a shape recognition device according to the first embodiment. [Figure 3] Figure 3 is an example of an input image according to the first embodiment. [Figure 4] Figure 4 shows an example of raster-format coordinate information represented by each pixel from the input image in the first embodiment. [Figure 5] Figure 5 shows an example of the distribution of the number of pixels on the semiconductor wafer relative to a given coordinate, based on the coordinate information acquired according to the first embodiment. [Figure 6] Figure 6 is a flowchart showing the procedure for deriving the center coordinates and axis lengths of the ellipse in the first embodiment. [Figure 7] Figure 7 is a block diagram showing an example of the configuration of a shape recognition device according to the second embodiment. [Figure 8] Figure 8 shows an example of an input image according to the second embodiment. [Figure 9] Figure 9 shows an example of a method for deriving the correction amount in the second embodiment, in which an edge point cloud is extracted and the equation of the ellipse is calculated. [Figure 10] Figure 10 shows an example of a method for deriving the correction amount in the second embodiment, illustrating the coordinate information of a semiconductor wafer where an orientation flat exists. [Figure 11] Figure 11 is a flowchart showing an example of the procedure for deriving the center coordinates and axis lengths of an ellipse in the second embodiment. [Figure 12] Figure 12 is an example of a flowchart showing the procedure for deriving the correction amount in the second embodiment. [Modes for carrying out the invention]

[0008] The semiconductor wafer shape recognition method and apparatus according to the embodiments will be described in detail below with reference to the attached drawings. However, the present invention is not limited to these embodiments.

[0009] (First embodiment) Figure 1 is a block diagram showing an example of the hardware configuration of a shape recognition device according to an embodiment. In this embodiment, an image of a semiconductor wafer is sent to the shape recognition device 1 by a video source or camera. The shape recognition device 1 has a hardware configuration similar to that of a general information processing device such as a PC or server. As an example, the shape recognition device has a CPU (Central Processing Unit) 61, ROM (Read Only Memory) 62, RAM (Random Access Memory) 63, HDD (Hard Disk Drive) / SSD (Solid State Drive) 64, and I / F 65. The CPU 61, ROM 62, RAM 63, HDD / SSD 64, and I / F 65 can communicate with each other via a bus 69, etc.

[0010] The CPU 61 is the processing unit (processor). The CPU 61 controls the operation of the entire shape recognition device 1. The ROM 62 is a read-only, non-volatile storage medium. Programs such as firmware are stored in the ROM 62. The RAM 63 is a volatile storage medium that allows for high-speed reading and writing of information. The RAM 63 is used as a workspace for the CPU 61 when it processes (calculates) information. The HDD / SSD 64 is a non-volatile storage medium that allows for reading and writing of information. The HDD / SSD 64 stores the OS (Operating System), various control programs, application programs, etc.

[0011] In addition to the CPU61, various other computing means (processors) such as GPUs (Graphics Processing Units), ASICs (Application Specific Integrated Circuits), and FPGAs (Field-Programmable Gate Arrays) can be used as appropriate.

[0012] Note that as the HDD / SSD 64, either an HDD or an SSD may be used, or only one of the HDD and SSD may be used. Also, the HDD / SSD 64 is not limited to HDDs and SSDs, and storage devices such as flash memories and CD-ROM drives can also be appropriately used.

[0013] The I / F 65 is an interface circuit that connects the bus 69 to various hardware and networks, etc., and controls this connection (communication). An LCD (Liquid Crystal Display) 71, an operation unit 72, and a dedicated device 73 are connected to the I / F 65.

[0014] The LCD 71 is a visual user interface (display device) for the user to check the state of the shape recognition device 1. The operation unit 72 is a user interface (input device) such as a keyboard or a mouse for the user to input information into the shape recognition device 1. Note that the LCD 71 and the operation unit 72 may be integrally configured as a touch panel display. Here, the LCD 71 is an example of a display unit.

[0015] FIG. 2 is a block diagram showing an example of the functional configuration of the shape recognition device 1 according to the first embodiment. As shown in FIG. 2, the shape recognition device 1 includes an imaging unit 10, a system control unit 11, a user I / F unit 14, a network I / F unit 15, and a mechanism control unit 16.

[0016] The user I / F unit 14 is an I / F for connecting the system control unit 11 and the operation unit 72. The network I / F unit 15 is an I / F for connecting the system control unit 11 to a network such as a LAN. The mechanism control unit 16 is a control unit for the operations of the shape recognition device 1 such as the reading operation of the semiconductor wafer.

[0017] The imaging unit 10 provides an input image to the system control unit 11. The imaging unit 10 may be any device that can acquire image data, such as a camera. Based on the input image acquired by the imaging unit 10, the system control unit 11 determines the shape of the semiconductor wafer, that is, the center coordinates and the lengths of the axes, as a shape recognition method. The system control unit 11 includes an image preprocessing unit 12 and an ellipse calculation unit 13.

[0018] The image preprocessing unit 12 acquires data of coordinates with and without a semiconductor wafer as shown in FIG. 4 described later from the input image acquired by the imaging unit 10 in order to determine the center coordinates and the lengths of the axes. By being integrated within its own FPGA (Field Programmable Gate Array) or ASIC (Application Specific Integrated Circuit), the image preprocessing unit 12 can achieve high processing speed and can be easily realized within hardware.

[0019] As shown in FIG. 2, the image preprocessing unit 12 includes a data acquisition unit 120 and a wafer presence / absence recognition unit 121. The data acquisition unit 120 acquires the input image acquired by the imaging unit 10. FIG. 3 is an example of an input image according to the first embodiment. The imaging unit 10 such as a camera acquires an input image of the semiconductor wafer and sends the input image to the shape recognition device in association with the coordinate information in the first direction X and the second direction Y as shown in FIG. 4 described later. In this embodiment, the shape of the semiconductor wafer is assumed to be an elliptical shape 300. This is because an ellipse has higher expressiveness than a circle, so the shape of the semiconductor wafer can be determined more accurately. From this input image, the shape recognition device 1 determines the center coordinates 301 (cx, cy), the length a302 of the axis in the first direction X, and the length b303 of the axis in the second direction Y. In addition to the input image as described above, wafer data acquired at a fixed pitch like pixels may be used. For example, data such as density or film thickness, which are characteristics of the wafer, acquired at a fixed pitch may also be used.

[0020] The wafer presence / absence recognition unit 121 determines whether or not a semiconductor wafer is present in the input image. Specifically, the wafer presence / absence recognition unit 121 determines whether or not a semiconductor wafer is present in the input image and converts the input image into binary image data in which the parts of the input image where a wafer is present are set to "1" and the other parts are set to "0". The data acquisition unit 120 acquires data like that shown in Figure 4, which will be described later, by linking the data acquired in the manner described above with the coordinate information. Figure 4 shows an example of raster-format coordinate information represented by each pixel from the input image in the first embodiment. The input image sent from the imaging unit 10 is sent to the image preprocessing unit 12 in the shape recognition device 1. From the input image, the portion 401 where the wafer exists is represented as "1" and the other portion 400 is represented as "0" by each pixel. The data of the pixels corresponding to "1" becomes data associated with coordinate information in the first direction X and the second direction Y, as shown in Figure 4.

[0021] The ellipse calculation unit 13 determines the shape of the semiconductor wafer, i.e., the center coordinates and axis lengths, based on the raster-format image data obtained as shown in Figure 4 above. The ellipse calculation unit 13 includes a distribution generation unit 130, a center coordinate calculation unit 131, an axis length calculation unit 132, and an output unit 133.

[0022] The distribution generation unit 130 calculates the distribution of the number of pixels present on the semiconductor wafer from the raster image data shown in Figure 4. Figure 5 is an example showing the distribution of the number of pixels on a semiconductor wafer for a given coordinate, based on coordinate information acquired according to the first embodiment. Specifically, the distribution generation unit 130 generates a distribution of the number of pixels on a semiconductor wafer for a given coordinate in the first direction X and the second direction Y, from raster-format coordinate information data as shown in Figure 4. Figure 5 shows the distribution for the first direction X (501-504) when the axis length a in the first direction X and the axis length b in the second direction Y are a=50, b=100, a=100, b=100, a=150, b=100, and a=200, b=100, respectively. The center coordinate calculation unit 131 and the axis length calculation unit 132 calculate the mean and standard deviation based on these distributions and calculate the shape of the semiconductor wafer (center coordinate (cx, cy), axis lengths in the first direction X and the second direction Y (a, b)).

[0023] The center coordinate calculation unit 131 calculates the center coordinates of the semiconductor wafer based on the distribution shown in Figure 5. The center coordinate calculation unit 131 obtains the coordinate distribution of the semiconductor wafer as shown in Figure 5 for the first direction X and the second direction Y, respectively, and calculates the mean values ​​MEAN(X) and MEAN(Y) for each coordinate based on this distribution. Furthermore, the center coordinate calculation unit 131 defines the center coordinate of the semiconductor wafer as the mean value of the coordinate distribution (cx, cy) using equation (1.1).

[0024]

number

[0025] Here, MEAN is a function that calculates the average value. X and Y are the coordinate values ​​of the data in the first direction and the data in the second direction for each pixel, respectively.

[0026] The axis length calculation unit 132 calculates the axis length of the semiconductor wafer based on the distribution shown in Figure 5. Similarly, the axis length calculation unit 132 obtains the coordinate distribution of the semiconductor wafer as shown in Figure 5 for the first direction X and the second direction Y, respectively, and calculates the standard deviations STD(X) and STD(Y) for each coordinate based on this distribution. Furthermore, the axis length calculation unit 132 defines the axis lengths of the semiconductor wafer as a and b, respectively, by using equation (1.2) and doubling the standard deviation of the coordinate distribution.

[0027]

number

[0028] Here, STD is the function for calculating the standard deviation. X and Y are the coordinate values ​​of the data in the first direction and the second direction, respectively, for each pixel.

[0029] The output unit 133 transmits the information obtained in this way, such as the center coordinates (cx, cy) of the semiconductor wafer and the axis lengths a and b, to the exposure apparatus 17.

[0030] Next, the shape recognition procedure in the shape recognition device 1 of the first embodiment will be described. Figure 6 is a flowchart of the procedure for deriving the center coordinates and axis lengths of the ellipse in the first embodiment. First, the imaging unit 10 acquires an input image of the semiconductor wafer that is the target of shape recognition (S600).

[0031] Next, the wafer presence / absence recognition unit 121 checks the presence or absence of a wafer from the acquired input image. If the wafer is not present (S601: No), the process terminates. On the other hand, if the wafer is present (S601: Yes), the process continues. Subsequently, the distribution generation unit 130 acquires the coordinate distribution of the semiconductor wafer in the first direction X and the second direction Y from the acquired input image (S602). Based on the acquired coordinate distribution, the center coordinate calculation unit 131 calculates the mean values ​​MEAN(X) and MEAN(Y) of the distribution. Then, using the calculated mean values, the center coordinates (cx, cy) of the semiconductor wafer are calculated using equation (1.1) (S603). This step corresponds to the "center coordinate calculation step". Next, the axis length calculation unit 132 calculates the standard deviations STD(X) and STD(Y) from the acquired coordinate distribution. The axis length calculation unit 132 uses equation (1.2) to determine the lengths of the first direction X and the second direction Y axes of the semiconductor wafer, a and b, from the obtained standard deviation values ​​(S604). This step corresponds to the "axis length calculation step".

[0032] As described above, in this embodiment, semiconductor wafer shape recognition is performed by calculating only the mean and standard deviation, resulting in low computational cost and rapid processing. For example, when determining edge areas from an input image as in the comparative example, the size and brightness of pixels have an influence, which can introduce noise into the edges and raise concerns about incorrect edge determination. Furthermore, in semiconductor wafers, sawtooth patterns are added to the edges during manufacturing, further increasing the impact. This embodiment uses coordinate values ​​including those within the wafer, not just specific edges, thus providing robustness against edge noise.

[0033] The shape recognition method according to this embodiment includes an acquisition step of acquiring coordinate values ​​in a first direction X and a second direction Y in an image of the upper surface of an elliptical shape recognition target; a center coordinate calculation step of calculating the center coordinates of the elliptical shape based on the distribution of the coordinate values; and an axis length calculation step of calculating the axis lengths of the elliptical shape based on the distribution of the coordinate values.

[0034] Specifically, in the shape recognition device 1 according to this embodiment, an input image of a semiconductor wafer acquired by the imaging unit 10 is obtained, and based on the acquired input image, distribution coordinates of the semiconductor wafer in a first direction and a second direction perpendicular to each other are generated, and the mean and standard deviation are calculated from the acquired coordinate distribution. Based on these acquired values, the center coordinates of the semiconductor wafer and the lengths of the axes in the two directions are determined, and the shape is recognized. As a result, the shape of the semiconductor wafer can be recognized without being affected by noise in the edge portion of the semiconductor wafer.

[0035] Furthermore, in the shape recognition method according to this embodiment, the center coordinate calculation step calculates the average value from the distribution of coordinate values ​​in the first and second directions and uses the calculated average value as the center coordinate, and the axis length calculation step calculates the standard deviation from the distribution of coordinate values ​​in the first and second directions and calculates the axis length as twice the calculated standard deviation. As a result, shape recognition inspection and analysis can be processed quickly with limited time and computing resources.

[0036] (Second embodiment) In the first embodiment, the center coordinates and axis lengths were determined assuming that the semiconductor wafer was elliptical in shape. This second embodiment, however, shows a method for recognizing the shape of a semiconductor wafer in which orientation flats and the like exist. An orientation flat is a flat area formed at the bottom of an elliptical shape, as shown in Figure 8, to indicate the crystal orientation of the semiconductor wafer. In addition to orientation flats, other shapes that are fixed each time, such as notches, can also be calculated in the same way as in this embodiment.

[0037] Figure 7 is a block diagram showing an example of the configuration of the shape recognition device 701 according to the second embodiment. As shown in Figure 7, the shape recognition device 701 according to the second embodiment comprises a system control unit 711, an image preprocessing unit 712, and an ellipse calculation unit 713. The image preprocessing unit 712 also includes a correction amount setting unit 714, and the ellipse calculation unit 713 includes a center coordinate calculation unit 715 and an axis length calculation unit 716. In the described Figure 2, the same reference numerals are used for components identical to those in the first embodiment, and their descriptions are omitted.

[0038] Similar to the first embodiment, an input image as shown in Figure 8 is acquired from the imaging unit 10. Figure 8 is an example of an input image according to the embodiment in the second embodiment. The semiconductor wafer shape assumed in this embodiment is a shape 800 in which an orientation flat portion 804 exists, which is a flat portion at the bottom of an ellipse. The shape of the semiconductor wafer is recognized by determining the center coordinates 801 of this shape 800, the length a802 of the axis in the first direction X, and the length b803 of the axis in the second direction Y.

[0039] (An example of a method for calculating the correction amount) The correction amount setting unit 714 derives the correction amount used when determining the center coordinates and axis lengths in cases where orientation flats, etc., are present. An example of how to calculate the correction amount is described below. First, the imaging unit 10 acquires an input image of a semiconductor wafer sample with an orientation flat, as shown in Figure 8, in order to determine the correction amount. Next, the wafer presence / absence recognition unit 121 extracts edge point cloud data of the semiconductor wafer sample based on the acquired input image, and the data acquisition unit 120 acquires that data.

[0040] Figure 9 shows an example of how to derive the correction amount in the second embodiment, in which an edge point cloud is extracted and the equation of an ellipse is calculated. The data acquisition unit 120 extracts an edge point cloud 900 containing the orientation flat portion 901 as shown in Figure 9 from the acquired input image. The correction amount setting unit 714 derives the equation of an ellipse 902, as shown in equation (2.1), using the least squares method based on the coordinate data of the edge point cloud excluding this orientation flat portion.

[0041]

number

[0042] Here, cx' and cy' are the coordinates of the center of the ellipse equation, and a' and b' are the lengths of the axes in the first direction X and the second direction Y of the ellipse equation.

[0043] Once the equation of the ellipse is determined, the wafer presence / absence recognition unit 121 recognizes the parts of the semiconductor wafer that are present and those that are not from the input image of the semiconductor wafer sample, and associates this with coordinate information in the first direction X and the second direction Y, thereby obtaining image data as shown in Figure 10. Figure 10 is an example showing the coordinate information of a semiconductor wafer with an orientation flat, as an example of a method for deriving the correction amount in the second embodiment.

[0044] Once image data as shown in Figure 10 is acquired, the distribution generation unit 130 acquires the coordinate distribution of the semiconductor wafer in the first direction X and the second direction Y, respectively. The center coordinate calculation unit 715 and the axis length calculation unit 716 derive the mean values ​​MEAN(X) and MEAN(Y) and the standard deviations STD(X) and STD(Y), respectively, from the acquired coordinate distribution. Next, the correction amount setting unit 714 uses equations (2.2) and (2.3) to determine the correction amounts kx, ky, ka, and kb from the center coordinates (cx', cy') and axis lengths a' and b' of the ellipse equation 902 described above.

[0045]

number

[0046] In this way, when recognizing the shape of a semiconductor wafer that is not elliptical, such as one with an orientation flat, it is necessary to calculate a correction amount in advance.

[0047] Returning to the explanation of Figure 7, the center coordinate calculation unit 715 calculates the center coordinates of the semiconductor wafer where the orientation flat is located. Similar to the first embodiment, the distribution generation unit 130 calculates the distribution of the number of pixels on the semiconductor wafer for the coordinates in the first and second directions. Then, the center coordinate calculation unit 715 calculates the average values ​​MEAN(X) and MEAN(Y) for the first (X) and second (Y) directions based on the generated coordinate distribution. The center coordinate calculation unit 715 uses equation (2.4) to determine the center coordinates (cx, cy) of the semiconductor wafer from the correction amount described above and the average value of the coordinate distribution generated by the distribution generation unit 130.

[0048]

number

[0049] Here, MEAN is a function that calculates the average value. X and Y are the coordinate values ​​of the data in the first direction and the data in the second direction for each pixel, respectively. kx and ky are correction amounts that take into account the effect of the orientation flat.

[0050] The axis length calculation unit 716 calculates the standard deviations STD(X) and STD(Y) of the coordinate distribution for each coordinate, based on the distribution of the number of pixels present on the semiconductor wafer for the coordinates in the first and second directions generated by the distribution generation unit 130. The axis length calculation unit 716 uses equation (2.5) to determine the lengths a and b of the first direction X and second direction Y axes of the semiconductor wafer from the correction amount described above and the standard deviation of the coordinate distribution generated by the distribution generation unit 130.

[0051]

number

[0052] Here, STD is the function for calculating the standard deviation. X and Y are the coordinate values ​​of the data in the first and second directions for each pixel, respectively. ka and kb are correction amounts that take into account the effect of the orientation flat.

[0053] In this embodiment, a method for recognizing the shape of a semiconductor wafer, including an orientation flat, will be explained with reference to Figure 11. Figure 11 is a flowchart showing an example of the procedure for deriving the center coordinates and axis lengths of an ellipse in the second embodiment. First, the correction amount setting unit 714 sets a correction amount corresponding to the shape of the semiconductor wafer, such as an orientation flat, in the shape recognition device 1 using the method described above (S1100). This step corresponds to the "correction amount calculation step".

[0054] After the derived correction amount is saved in the correction amount setting unit 714, the imaging unit 10 acquires an input image of the semiconductor wafer (S1101). Next, the wafer presence / absence recognition unit 121, after acquiring the input image, recognizes the parts of the input image where the semiconductor wafer is present and parts where it is not. The mechanism control unit 16 then asks the user to confirm whether to terminate the process if the semiconductor wafer is not present (S1102: No), and continues the process if the semiconductor wafer is present (S1102: Yes).

[0055] The distribution generation unit 130 generates a coordinate distribution (S1103) based on the binary data of the parts of the semiconductor wafer that are present and those that are not, obtained above, and the data linked to the coordinates of the first direction X and the second direction Y shown in Figure 8.

[0056] The center coordinate calculation unit 715 calculates the average values ​​MEAN(X) and MEAN(Y) for each coordinate based on the generated coordinate distribution. Then, using equation (2.4), the center coordinate calculation unit 715 calculates the corrected center coordinates (cx, cy) of the semiconductor wafer from the average values ​​of the coordinate distribution and kx and ky set by the correction amount setting unit 714 (S1104).

[0057] The axis length calculation unit 716 calculates the standard deviations STD(X) and STD(Y) for each coordinate based on the generated coordinate distribution. Then, using equation (2.5), the axis length calculation unit 716 calculates the lengths a and b of the first direction X and second direction Y axes of the semiconductor wafer from the standard deviation of the coordinate distribution and the correction amount setting unit 714 (S1105).

[0058] Once the center coordinates and axis lengths of the ellipse are determined, the user checks the calculated values ​​on the LCD71. If the user is not satisfied with the calculated values, the mechanism control unit 16 recaptures the input image and recalculates the center coordinates and axis lengths, and re-acquires the input image (S1106: No). Alternatively, if the values ​​are satisfactory, the mechanism control unit 16 completes the job and terminates the process (S1106: Yes).

[0059] Figure 12 is an example of a flowchart showing the procedure for deriving the correction amount in the second embodiment. First, the imaging unit 10 acquires an input image of a semiconductor wafer sample with an orientation flat in order to determine the correction amount (S1200).

[0060] Next, the wafer presence / absence recognition unit 121 extracts edge point cloud data of the semiconductor wafer sample based on the acquired input image. If extraction is not possible, it acquires the input image again (S1201: No), and if successful, it continues processing (S1201: Yes). The data acquisition unit 120 extracts edge point cloud 900 containing the orientation flat portion 901 as shown in Figure 9 from the acquired input image. The correction amount setting unit 714 derives the ellipse equation 902 as shown in equation (2.1) using the least squares method based on the coordinate data of the edge point cloud excluding this orientation flat portion (S1202). The mechanism control unit 16 acquires the input image again if it recognizes that the semiconductor wafer does not exist from the acquired input image (S1203: No), and continues processing if it recognizes that the semiconductor wafer exists (S1203: Yes).

[0061] After acquiring image data as shown in Figure 10, the distribution generation unit 130 acquires the coordinate distribution of the semiconductor wafer in the first direction X and the second direction Y, respectively (S1204). The center coordinate calculation unit 715 and the axis length calculation unit 716 derive the mean values ​​MEAN(X) and MEAN(Y) and the standard deviations STD(X) and STD(Y), respectively, from the acquired coordinate distribution (S1205). Next, the correction amount setting unit 714 uses equations (2.2) and (2.3) to determine the correction amounts kx, ky, ka, and kb from the center coordinates (cx', cy') and axis lengths a' and b' of the ellipse equation 902 described above (S1206).

[0062] It should be noted that the derivation of the correction amount shown in Figure 12 does not need to be performed for each wafer. For example, if a batch consists of wafers produced under the same conditions, Figure 12 can be applied to a few wafers within the batch to calculate the correction amount, and the remaining wafers in the batch can then follow that correction amount.

[0063] In the shape recognition method according to this embodiment, the center coordinate calculation step calculates the center coordinate by correcting the average value calculated from the distribution of coordinate values ​​in the first direction X and the second direction Y with a correction amount, and the axis length calculation step calculates the axis length by correcting the standard deviation calculated from the distribution of coordinate values ​​in the first direction X and the second direction Y with a correction amount. As a result, even in the case of semiconductor wafers with orientation flats or notches formed on them, the center coordinate and axis length can be determined from the average value and standard deviation of the coordinate distribution calculated based on the input image.

[0064] Furthermore, the shape recognition method according to this embodiment includes a correction amount calculation step which extracts edge point cloud data from the shape recognition target and calculates correction amounts for the center coordinates and axis length of the shape recognition target calculated from equation (2.1) derived by the least squares method, the mean and standard deviation obtained from the distribution of coordinate values ​​in the first direction X and the second direction Y, and equations (2.2) to (2.3). The center coordinate calculation step calculates the center coordinates of the shape recognition target by correcting the mean value calculated from the distribution of coordinate values ​​with the correction amount of the center coordinates calculated in the correction amount calculation step, and the axis length calculation step calculates the axis length of the shape recognition target by correcting the standard deviation calculated from the distribution of coordinate values ​​with the correction amount of the axis length calculated in the correction amount calculation step. Thus, according to this embodiment, even when recognizing the shape of a semiconductor wafer with an orientation flat or notch formed, the correction amount can be determined without comparing it with an input image of an elliptical semiconductor wafer before the orientation flat or notch is formed.

[0065] Furthermore, the shape recognition device 1,701 utilizing the shape recognition method according to this embodiment includes an acquisition unit that acquires coordinate values ​​of a first direction X and a second direction Y in an image of the upper surface of an elliptical shape recognition target, a center coordinate calculation unit that calculates the center coordinates of the ellipse based on the distribution of the coordinate values, and an axis length calculation unit that calculates the axis lengths of the ellipse based on the distribution of the coordinate values. As a result, according to this embodiment, there is robustness against edge noise of semiconductor wafers and it is possible to process quickly with limited time and computing resources.

[0066] In the above embodiment, a semiconductor wafer was used as an example of the object to be recognized for shape recognition, but the object to be recognized for shape recognition is not limited to a semiconductor wafer as long as it has an elliptical shape.

[0067] While several embodiments of the present invention have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These novel embodiments can be carried out in a variety of other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims of the invention and its equivalents.

[0068] 1 Shape recognition device, 10 Imaging unit, 120 Data acquisition unit, 714 Correction amount setting unit, 715 Center coordinate calculation unit, 716 Axis length calculation unit.

Claims

1. A shape recognition method performed by a shape recognition device, An acquisition step to obtain coordinate values ​​in a first direction and a second direction in an image of the upper surface of an elliptical shape recognition target, A center coordinate calculation step, which calculates the center coordinates of the ellipse based on the distribution of the coordinate values, A step of calculating the axis length of the elliptical shape based on the distribution of the coordinate values, A shape recognition method including [details omitted].

2. The center coordinate calculation step involves calculating an average value from the distribution of the coordinate values ​​in the first direction and the second direction, and using the calculated average value as the center coordinate. The axial length calculation step involves calculating the standard deviation from the distribution of the coordinate values ​​in the first and second directions, and calculating the axial length as twice the calculated standard deviation. The shape recognition method according to claim 1.

3. The aforementioned center coordinate calculation step involves calculating the center coordinate by correcting the average value calculated from the distribution of the coordinate values ​​in the first and second directions with a correction amount. The aforementioned axis length calculation step calculates the length of the axis by correcting the standard deviation calculated from the distribution of the coordinate values ​​in the first and second directions with a correction amount. The shape recognition method according to claim 1.

4. The system further comprises extracting edge point cloud data from the shape recognition target, deriving equation (1) by the least squares method, the mean and standard deviation obtained from the distribution of coordinate values ​​X and Y in the first and second directions, and a correction amount calculation step that calculates the correction amount for the center coordinates and axis length of the shape recognition target calculated from equations (2) to (5), [Math 1] (Here, cx' and cy' are the coordinates of the center of the ellipse equation, a' and b' are the axis lengths of the ellipse equation, and kx, ky, ka, and kb are the correction quantities.) The center coordinate calculation step calculates the center coordinate of the shape recognition target by correcting the average value calculated from the distribution of the coordinate values ​​with the correction amount of the center coordinate calculated in the correction amount calculation step. The shaft length calculation step involves calculating the length of the shaft of the shape recognition target by correcting the standard deviation calculated from the distribution of the coordinate values ​​with the correction amount of the shaft length calculated in the correction amount calculation step. The shape recognition method according to claim 3, including the following:

5. The shape recognition method described above is for a semiconductor wafer. The shape recognition method according to claim 1.

6. An acquisition unit that acquires coordinate values ​​in a first direction and a second direction in an image of the upper surface of an elliptical shape recognition target, A center coordinate calculation unit calculates the center coordinates of the ellipse based on the distribution of the aforementioned coordinate values, An axis length calculation unit that calculates the length of the axis of the ellipse based on the distribution of the coordinate values, A shape recognition device equipped with the following features.

Images