Detection method, detection system, and storage medium

Abstract

The application provides a detection method, comprising: performing optimization operation on the size of a reference area along each direction to be optimized, respectively, wherein the optimization operation comprises: obtaining a judgment function about the size of the reference area along the direction to be optimized, the judgment function being positively or negatively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target value to be detected; performing optimization processing on the size of the reference area along the direction to be optimized through the judgment function, so that the judgment function of the reference area satisfies a preset condition, and the optimized size of the reference area along the direction to be optimized is obtained; and obtaining the target value to be detected of the target to be detected according to the detection information of the reference area after the optimization operation. Through the optimization operation on the reference area by the judgment function, the reference area can avoid losing too much useful detection information, so that the signal-to-noise ratio of the detection information or the robustness of the detection result can be improved.

Description

Technical Field

[0001] This invention relates to the field of semiconductor wafer inspection equipment, and in particular to an inspection method, inspection system and storage medium. Background Technology

[0002] During photolithography, excessive dimensional errors in the formed pattern can negatively impact the performance of the resulting semiconductor device, potentially leading to short circuits or open circuits. Therefore, the fabrication of semiconductor devices requires the inspection of the formed pattern (including its center position, size, and alignment errors), and the accuracy of this inspection directly affects the quality of the final semiconductor device. This is particularly true in multi-patterning processes, where the smaller pattern size places even higher demands on inspection accuracy.

[0003] In the process of multi-layer patterning, the positional change vector between the current layer pattern and the corresponding pattern in the previous layer of the exposure field on the wafer is defined as the alignment accuracy. Alignment accuracy is one of the fundamental metrics of photolithography, and its tolerance is typically one-third of the critical dimension. In photolithography, each layer prepares a specific mark at a specific location, called an alignment mark. The IBO (Image-Based Alignment Measurement) method uses an optical imaging device to obtain images corresponding to the alignment marks of two layers, thereby calculating the alignment error.

[0004] The existing overlay measurement process has low detection accuracy and is increasingly unable to meet the requirements of multi-layer patterning. Summary of the Invention

[0005] To address the above problems, this invention proposes a detection method that can reduce the influence of noise signals, improve the signal-to-noise ratio of the detection information and the robustness of the target value, thereby improving the detection quality.

[0006] The technical solution of the present invention provides a detection method, comprising: acquiring detection information of a test object, wherein the test object has a test target and a reference point, and the detection information includes the correspondence between the reference position of the reference point on the surface of the test object and the detection parameters;

[0007] A reference area is set for the detection information, and the detection information in the reference area is used to obtain the target value of the target to be tested. The reference area includes one or more directions to be optimized.

[0008] The dimensions of the reference region along each direction to be optimized are optimized, and the optimization operation includes the following steps:

[0009] A judgment function is obtained regarding the size of the reference region along the direction to be optimized. The judgment function is positively or negatively correlated with the signal-to-noise ratio of the detection information of the reference region or the robustness of the target value to be measured. The size of the reference region along the direction to be optimized is optimized using the judgment function to make the judgment function of the reference region meet preset conditions, thereby obtaining the optimized size of the reference region along the direction to be optimized. The preset conditions include: the optimized judgment function value of the reference region makes the signal-to-noise ratio or robustness greater than or equal to a preset value, or makes the signal-to-noise ratio or robustness have a maximum value.

[0010] The target value of the target to be tested is obtained based on the detection information of the reference area after the optimization operation.

[0011] Optionally, the optimized reference area judgment function value makes the signal-to-noise ratio or robustness greater than or equal to a preset value, including: when the judgment function is positively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target value to be measured, the optimized reference area judgment function value is greater than the preset judgment function value; or when the judgment function is negatively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target value to be measured, the optimized reference area judgment function value is less than the preset judgment function value.

[0012] Maximizing the signal-to-noise ratio or robustness includes: when the judgment function is positively correlated with the signal-to-noise ratio of the detection information in the reference region or the robustness of the target value to be measured, the optimized judgment function in the reference region has a maximum value; or when the judgment function is negatively correlated with the signal-to-noise ratio of the detection information in the reference region or the robustness of the target value to be measured, the optimized judgment function in the reference region has a minimum value.

[0013] Optionally, the step of optimizing the size of the reference region along the direction to be optimized using the judgment function, so that the judgment function of the reference region satisfies the preset conditions, includes:

[0014] Multiple alternative sizes are set for the reference region along the direction to be optimized;

[0015] The judgment function value for each candidate size is obtained according to the judgment function, and the maximum or minimum value among the judgment function values ​​of each candidate size is the reference judgment function value.

[0016] The optimized size of the reference region along the direction to be optimized is obtained based on the judgment function value of each candidate size. The optimized size makes the signal-to-noise ratio or robustness greater than or equal to the signal-to-noise ratio or robustness determined by the reference judgment function value.

[0017] Optionally, the one or more directions to be optimized include a first direction, and the judgment function is a first judgment function; the first judgment function includes the functional relationship between the signal-to-noise ratio of the detection information of the reference region and the size of the reference region along the direction to be optimized;

[0018] The optimization operation for each direction to be optimized includes: optimizing the first direction using a first judgment function;

[0019] Setting multiple different alternative sizes for the reference region along the direction to be optimized includes: setting multiple different alternative sizes for the reference region along the first direction;

[0020] The detection information of the reference area is composed of multiple first reference groups, each of which includes multiple reference positions that have the same position in the first compression direction; the reference positions in different first reference groups have different positions in the first compression direction.

[0021] In the step of optimizing the first direction, the step of obtaining the judgment function value of each candidate size according to the judgment function includes: compressing the detected parameters of some or all reference positions in the first reference group to obtain a first compression value; repeating the compression process for each first reference group to obtain the first compression value of each first reference group, and obtaining a first correspondence between the position of each first reference group along the second direction and the first compression value, wherein the second direction is perpendicular to the first compression direction; obtaining the ratio between the sum of the first compression values ​​and the discrete values ​​in the first correspondence under each candidate size, and obtaining the judgment function value of each candidate size.

[0022] The discrete value represents the degree of dispersion of the first compression value within the reference region.

[0023] Optionally, the first direction has a non-zero angle with the first compression direction; or,

[0024] Compression processing is performed on the detected parameters of some or all reference positions within the first reference group, including: compression processing is performed on the detected parameters of each reference position within the first reference group.

[0025] Optionally, the detection parameters include grayscale value or light intensity value;

[0026] The sum includes the sum of all first compression values ​​in the reference region, or the square of the sum of all first compression values ​​in the reference region, or the sum of the squares of all first compression values ​​in the reference region.

[0027] The photometric discrete values ​​include: the variance, root mean square, or standard deviation of all first compression values ​​within the reference region.

[0028] Optionally, the step of obtaining the optimized size of the reference region along the direction to be optimized based on the judgment function values ​​of each candidate size, wherein the optimized size makes the signal-to-noise ratio or robustness greater than or equal to the reference judgment function value, includes: fitting the correspondence between the candidate sizes and the judgment function values ​​to obtain a fitting function; obtaining the candidate size of the reference region corresponding to the extreme value of the fitting function to obtain the optimized size; or...

[0029] The optimal judgment function value is obtained by finding the maximum value among all judgment function values; the optimal size is obtained by finding the alternative size of the reference area corresponding to the optimized judgment function value.

[0030] Optionally, the one or more directions to be optimized include a third direction, and the judgment function is negatively correlated with the signal-to-noise ratio of the detection information in the reference region or the robustness of the target value to be measured; the judgment function is a second judgment function;

[0031] The optimization operation for each direction to be optimized includes: optimizing the first direction using a second judgment function;

[0032] Setting multiple different alternative sizes for the reference area along the direction to be optimized includes: setting multiple different alternative sizes for the reference area along a third direction;

[0033] In the optimization operation of the first direction through the second judgment function, the judgment function value of each candidate size is obtained according to the judgment function, and the maximum value among the judgment function values ​​of each candidate size is the reference judgment function value, including:

[0034] The target value of the object to be tested is obtained based on the detection information of the reference area of ​​each alternative size;

[0035] A movable window is set for the alternative sizes, and the movable window includes a preset number of adjacent alternative sizes;

[0036] Obtain the target value of multiple adjacent candidate sizes in the moving window;

[0037] The discrete values ​​of the target value to be measured are obtained from multiple adjacent candidate sizes in the moving window to obtain the judgment function value; the moving window is moved to include multiple other adjacent candidate sizes, and the step of obtaining the judgment function value is repeated until multiple judgment function values ​​are obtained, and the minimum value among the multiple judgment function values ​​is the reference judgment function value; the median value of the candidate sizes in the moving window of the reference judgment function value is obtained to obtain the optimized size.

[0038] Optionally, the detection information of the reference area consists of multiple second reference groups, each of which includes multiple reference positions having the same position in the second compression direction; the reference positions in different second reference groups have different positions in the second compression direction.

[0039] Obtaining the target value of the test object based on the detection information of the reference area of ​​the alternative size includes: compressing the detected parameters at all reference positions of the second reference group to obtain a second compression value; repeating the compression process for each second reference group to obtain the second compression value of each second reference group, and obtaining a second correspondence between the position of each second reference group along the test direction and the second compression value, wherein the test direction is perpendicular to the second compression direction; and obtaining the target value of the test object along the test direction based on the second correspondence.

[0040] Optionally, the second compression direction is parallel to or has an acute angle with the direction to be optimized;

[0041] If the target value to be measured includes the size of the target along the direction to be measured, the target value to be measured along the direction to be measured is obtained according to the second correspondence, including: obtaining the gradient value of the second compression value at each reference position in the second correspondence to obtain the gradient relationship; obtaining the reference positions corresponding to the two maxima of the gradient relationship in the reference region to obtain the two extreme positions; obtaining the distance between the two extreme positions to obtain the size of the target along the direction to be measured.

[0042] Optionally, the direction to be measured is parallel to the direction to be optimized or has an acute angle between them;

[0043] The detection information of the reference area is composed of multiple second reference groups, each of which includes multiple reference positions that have the same position in the second compression direction; the reference positions in different second reference groups have different positions in the second compression direction.

[0044] Obtaining the target value of the test object based on the detection information of the reference area of ​​the alternative size includes: compressing the detected parameters at some or all reference positions of the second reference group to obtain a second compressed value; repeating the compression process for each second reference group to obtain the second compressed value of each second reference group, and obtaining a second correspondence between the position of each second reference group along the test direction and the second compressed value, wherein the test direction is perpendicular to the second compression direction; performing autocorrelation on the second correspondence within each reference position of the reference area to obtain an autocorrelation function; obtaining the time delay variable that maximizes the value of the autocorrelation function as a position variable; and obtaining the position information of the center of the test object based on the position variable.

[0045] Optionally, the target value to be measured includes: alignment error between different targets to be measured; obtaining the target value of the object to be measured based on the detection information of the reference area of ​​the candidate size further includes: after obtaining the position information of the center of the target to be measured based on the position variable, performing difference processing on the position information of different targets to be measured to obtain the alignment error between the centers of the targets to be measured; or,

[0046] The target value to be measured includes the position information of the center of the target along the direction to be measured; the reference area includes at least two sub-reference areas; the at least two sub-reference areas include a first sub-reference area and a second sub-reference area set at intervals, the first sub-reference area includes at least a portion of the edge contour of one end of the target along the direction to be measured, and the second sub-reference area includes at least a portion of the edge contour of the other end of the target along the direction to be measured.

[0047] Optionally, the one or more directions to be optimized include a first direction and a third direction;

[0048] The optimization operation for each direction to be optimized includes: optimizing the size of the reference area along the first direction; and optimizing the size of the reference area along the third direction.

[0049] The target to be tested includes at least two marker units arranged symmetrically at the center, and the marker units are rectangular; the reference area is rectangular, the first direction is the length direction of the marker unit, the third direction is the width direction of the marker unit, and the third direction is perpendicular to the first direction.

[0050] Optionally, the discrete value of the target value to be measured is: the variance, root mean square, or standard deviation of the target values ​​to be measured for multiple alternative sizes; the median includes the mean or median.

[0051] Optionally, the shape of the reference area is the same as the shape of the target to be tested; setting a reference area for the detection information includes: setting a reference area for the detection information such that the center of the reference area coincides with the center of the detection information of the target to be tested.

[0052] Optionally, the test object includes multiple test targets, which are located in different layers of the test object or formed in different processes; the test targets include multiple detection marks symmetrically arranged about the center of symmetry of the test targets; the test target value includes: alignment error between the centers of symmetry of different test targets;

[0053] Setting a reference area for the detection information, such that the center of the reference area coincides with the center of the detection information of the target to be tested, includes: setting a sub-reference area group for the detection information of each detection mark, wherein the shape of the sub-reference area group is the same as the shape of the detection mark, and the reference area includes one or more of the sub-reference area groups; and ensuring that the center of symmetry of each sub-reference area group coincides with the center of symmetry of the detection information of the detection mark.

[0054] Optionally, the detection information of the reference area consists of multiple third reference groups, each third reference group including multiple reference positions having the same position in the third compression direction; the reference positions in different third reference groups have different positions in the third compression direction;

[0055] To ensure that the center of symmetry of each sub-reference group coincides with the center of symmetry of the detection information of the detection mark, the following steps are taken: compressing the detection parameters at all reference positions of the third reference group to obtain a third compressed value; repeating the compression process for each third reference group to obtain the third compressed value for each third reference group, thereby obtaining a third correspondence between the position of each third reference group along the test direction and the third compressed value, wherein the test direction is perpendicular to the third compression direction.

[0056] Obtain the autocorrelation function of the third correspondence, and obtain the time delay variable that maximizes the value of the autocorrelation function as the position variable; obtain the reference position of the center of the reference area based on the position variable;

[0057] Alternatively, the detection mark may include multiple mark units; the sub-reference area group may include multiple sub-reference areas, such that the center of symmetry of each sub-reference area group coincides with the center of symmetry of the detection information of the detection mark; including: aligning the center of each sub-reference area with the center of each mark unit.

[0058] Optionally, the compression process includes: weighting the detection information, squaring the weights, or squaring the weights; the weighting is to sum the parameters after multiplying them by a weighting coefficient.

[0059] Optionally, the detection information is a detection image, and the detection parameter is a pixel grayscale or light intensity value.

[0060] A detection system, characterized in that it comprises: a processor, the processor being used to execute the above-described detection method.

[0061] The present invention also provides a storage medium storing one or more computer instructions, which are used to implement the detection method for the analyte described in any one of the above-mentioned methods.

[0062] In the detection method provided by the technical solution of the present invention, by setting a reference area for the detection information, the target to be tested is detected only by the detection information within the reference area, which can reduce the influence of noise signals at the edge of the detection pattern on the detection structure. Furthermore, by setting a judgment function to optimize the reference area, it is possible to avoid the reference area losing too much useful detection information, thereby improving the signal-to-noise ratio of the detection information or improving the robustness of the detection results, and thus improving the detection quality.

[0063] Furthermore, by setting multiple alternative dimensions along the direction to be optimized for the reference area, and obtaining the judgment function value for each alternative dimension according to the judgment function, the judgment function and the alternative dimensions can be correlated, thereby enabling the acquisition of optimized alternative dimensions based on the judgment function value and improving the detection quality.

[0064] Furthermore, by compressing the detection information of the reference area along a first compression direction parallel to a first direction, the detection information of the detection area can be transformed into one-dimensional data. The first compression value is correlated with the candidate size, and a first judgment function is constructed based on the ratio between the sum and discrete values ​​of the reference area detection information. The value of this judgment function is positively correlated with the signal-to-noise ratio (SNR) of the detection information. Therefore, the first judgment function is a one-dimensional function representing the SNR associated with the candidate size. This allows for the optimization of the candidate size to obtain a function with a high SNR, thereby improving detection accuracy.

[0065] Furthermore, by compressing some or all of the detected parameters of the reference region along the first compression direction, which is parallel to the second direction, the first compression direction is parallel to the second direction. By obtaining the first compression value of each reference position along the first direction of the reference region, the reference position is different when the candidate size is different. Then, by constructing a first judgment function based on the ratio between the sum value and the discrete value of the reference region detection information, the reference judgment function value can correlate the candidate size with the signal-to-noise ratio. Thus, the candidate size can be optimized through the first judgment function to obtain a function with a high signal-to-noise ratio, thereby improving the detection accuracy.

[0066] Furthermore, by setting multiple alternative sizes for the reference area and obtaining a target value to be measured based on the detection information within the reference area of ​​each alternative size; and constructing a second judgment function based on the dispersion of the target value to be measured in the reference area of ​​each alternative size, the second judgment function can establish a relationship between the robustness of the reference area size and the target value to be measured, thereby optimizing the reference area size through the second judgment function and improving the robustness of the detection method. Attached Figure Description

[0067] The present invention will be specifically described below with reference to the accompanying drawings and embodiments. The advantages and implementation methods of the present invention will become more apparent from this description. The content shown in the drawings is for illustrative purposes only and does not constitute any limitation on the present invention. The drawings are schematic only and are not strictly drawn to scale. In the drawings:

[0068] Figure 1 This is a schematic diagram of the structure of each step in an embodiment of the detection method of the present invention;

[0069] Figure 2 and Figure 3 This is a schematic diagram of the target and reference area in an embodiment of the detection method of the present invention;

[0070] Figures 4 to 6 This is a flowchart of the specific steps of an optimized operation of an embodiment of the detection method of the present invention;

[0071] Figure 7 This is a schematic diagram of the moving window process in an embodiment of the detection method of the present invention. Detailed Implementation

[0072] The present invention provides a detection method, comprising:

[0073] The detection method can optimize the size of the reference area and improve the signal-to-noise ratio or robustness by constructing a judgment function to associate the reference area with the signal-to-noise ratio of the detection information or the target value to be measured.

[0074] The detection method of the present invention will be described in detail below.

[0075] Figure 1 This is a flowchart of each step in an embodiment of the detection method for the analyte provided by the technical solution of the present invention.

[0076] refer to Figure 1 The present invention provides a detection method, comprising:

[0077] SP1, acquire detection information of the test object, wherein the test object has a test target and a reference point, and the detection information includes the correspondence between the reference position of the reference point on the surface of the test object and the detection parameters;

[0078] SP2 sets a reference area for the detection information. The detection information in the reference area is used to obtain the target value of the target to be tested. The reference area includes one or more directions to be optimized.

[0079] SP3 optimizes the size of the reference region along each optimization direction. The optimization steps include: obtaining a judgment function for the size of the reference region along the optimization direction, where the judgment function is positively or negatively correlated with the signal-to-noise ratio (SNR) of the detection information of the reference region or the robustness of the target value; optimizing the size of the reference region along the optimization direction using the judgment function to satisfy preset conditions, thereby obtaining the optimized size of the reference region along the optimization direction. The preset conditions include: the optimized reference region judgment function value makes the SNR or robustness greater than or equal to a preset value; or makes the SNR or robustness have a maximum value.

[0080] SP4 obtains the target value of the target to be tested based on the detection information of the reference area after the optimization operation.

[0081] By setting a reference area for the detection information, the target to be tested can be detected only using the detection information within the reference area. This reduces the impact of noise signals at the edges of the detection pattern on the detection structure. Furthermore, by setting a judgment function to optimize the reference area, excessive useful detection information can be avoided, thereby improving the signal-to-noise ratio of the detection information or the robustness of the detection results, and ultimately improving the detection quality.

[0082] Figure 2 and Figure 3 This is a schematic diagram of the target and reference area in an embodiment of the detection method of the present invention; Figures 4 to 6 This is a flowchart of the specific steps of an optimized operation of an embodiment of the detection method of the present invention; Figure 7 This is a schematic diagram of the moving window process in an embodiment of the detection method of the present invention.

[0083] The following combination Figures 2 to 7 The technical solution of the present invention will be described in detail below.

[0084] Please refer to Figure 2 Execute step SP1 to obtain the detection information of the test object. The test object has a test target 110 and a reference point. The detection information includes the correspondence between the reference position of the reference point on the surface of the test object and the detection parameters.

[0085] In this embodiment, the object under test is a wafer. In other embodiments, the object under test may also be a chip, a screen panel, or a semiconductor device.

[0086] In this embodiment, the target to be tested 110 is an overlay mark. In other embodiments, the target to be tested 110 can be a metal thin film, a TSV hole, a semiconductor thin film, or a positioning mark, etc.

[0087] The test object includes multiple test targets 110, which are located in different layers of the test object or formed in different processes. Each test target 110 includes multiple detection marks symmetrically arranged about a center of symmetry. The reference area includes multiple sub-reference area groups. The test target 110 value includes the alignment error between the centers of symmetry of different test targets 110 along the test direction. Specifically, in this embodiment, the test object includes two test targets 110, and a test target 110 and a test target 110.

[0088] In this embodiment, the at least one optimization direction includes a first direction and a third direction. In other embodiments of the present invention, the at least one optimization direction may include only the first direction or the third direction.

[0089] The target to be tested 110 includes: at least one detection mark arranged in a centrally symmetrical manner; the reference area includes at least one sub-reference area group arranged in a centrally symmetrical manner, and one sub-reference area group corresponds to one detection mark; each detection mark includes at least one mark unit; the sub-reference area group includes at least one sub-reference area 120, and one sub-reference area 120 corresponds to one mark unit.

[0090] Specifically, in this embodiment, the target to be tested 110 includes two centrally symmetrically arranged detection marks, namely a first detection mark 110a and a second detection mark 110b. Correspondingly, the reference area includes two centrally symmetrically arranged sub-reference area groups; each detection mark includes two mark units, and the sub-reference area group includes two sub-reference areas 120.

[0091] The marking unit is rectangular; the sub-reference area 120 is rectangular, the first direction is the length direction, the third direction is the width direction, and the third direction is perpendicular to the first direction. In other embodiments of the present invention, the first direction is the width direction, and the third direction is the length direction.

[0092] In this embodiment, the arrangement of two sub-reference regions 120 in one sub-reference region group is perpendicular to the arrangement of two sub-reference regions 120 in another sub-reference region group along the second direction. The arrangement direction of the two sub-reference regions 120 in each sub-reference region group is parallel to the width direction of the sub-reference region 120.

[0093] In other embodiments, the object to be tested includes only one target 110, which is a positioning marker and is cross-shaped or ring-shaped.

[0094] Please refer to Figure 3Execute step SP2, set a reference area for the detection information according to each reference position of the object to be tested, the detection information of the reference area is used to obtain the value of the target 110 to be tested, and the reference area includes one or more directions to be optimized.

[0095] The technical solution of this application sets a reference area for the detection information and obtains the value of the target 110 of the target 110 using only the detection information within the reference area. This avoids the influence of noise in the edge detection information of the target 110 on the value of the target 110, thereby improving the detection accuracy.

[0096] In this embodiment, the detection information is a detection image of the object to be tested, and the detection parameter is the pixel grayscale value of the detection image. In other embodiments of the present invention, the detection information is the correspondence between the reference position of each point on the surface of the object to be tested and the detection parameter. The detection reference is a light intensity value, voltage value, or current value, etc.

[0097] The value of the target 110 is the positional or geometrical information of the target 110. The geometrical information related to the target 110 includes the width, thickness, or length of the target 110. The positional information of the target 110 includes the offset between the centers of two targets 110 and the reference position of the target 110.

[0098] This embodiment uses the value of the target to be tested 110 as an example of the offset between the centers of two targets to be tested 110.

[0099] In this embodiment, there are multiple targets 110 to be tested. Correspondingly, there are multiple reference areas. Setting reference areas for the detection information includes: aligning the center of each reference area with the reference position of the center of the target 110 to be tested;

[0100] In one embodiment, each target 110 includes multiple detection markers; the reference area includes multiple sub-reference areas 120; aligning the center of each reference area with the center of each target 110 includes: In this embodiment, the target 110 includes multiple centrally symmetrical detection markers, and setting a reference area for the detection information includes: setting a sub-reference area 120 for the detection information of each detection marker; the reference area includes one or more of the sub-reference areas 120.

[0101] Specifically, the detection information of each target 110 includes multiple sub-reference area groups, each sub-reference area group includes one or more sub-reference areas 120, each sub-reference area group has a test direction, and each sub-reference area group is used to obtain the target 110 value in the corresponding test direction.

[0102] The sub-reference regions 120 of each sub-reference region group are centrally symmetrical, and the sub-reference regions 120 within the sub-reference region group are arranged along the width direction of the sub-reference regions 120.

[0103] The detection information is a matrix formed by the detection parameters of multiple reference positions.

[0104] The detection information of the reference area consists of multiple third reference groups, each of which includes multiple reference positions that have the same position in the third compression direction; the reference positions in different third reference groups have different positions in the third compression direction.

[0105] To make the center of the reference area coincide with the reference position of the center of the target 110 under test, the method includes: making the center of symmetry of each sub-reference area group coincide with the center of symmetry of the target 110 under test.

[0106] To make the center of symmetry of the sub-reference group coincide with the center of symmetry of the target 110, the following steps are taken: compressing the detected parameters at all reference positions of the third reference group to obtain a third compressed value; repeating the compression process for each third reference group to obtain the third compressed value of each third reference group, obtaining a third correspondence between the position of each third reference group along the test direction and the third compressed value, obtaining the autocorrelation function of the third correspondence with the test direction perpendicular to the third compression direction, and obtaining the time delay variable that maximizes the value of the autocorrelation function as the position variable; and obtaining the reference position of the center of the reference area based on the position variable.

[0107] The compression process includes:

[0108] The compression processing performed on the detected parameters at some or all reference positions along the third compression direction of the sub-reference area group is the third compression processing.

[0109] The direction to be measured is perpendicular to, parallel to, or has an acute angle with the direction to be optimized. Specifically, in this embodiment, each sub-reference area group has one direction to be measured; specifically, among the multiple sub-reference area groups, there is a first sub-reference area group and a second sub-reference area group; each sub-reference area group has two mutually perpendicular directions to be optimized, respectively along the long side and the short side of the sub-reference area 120. In this embodiment, the direction to be measured of the first sub-reference area group is parallel to and perpendicular to the two directions to be optimized of the first sub-reference area group, respectively; the direction to be measured of the second sub-reference area group is parallel to and perpendicular to the two directions to be optimized of the second sub-reference area group, respectively.

[0110] The detection method of the invention includes: repeatedly aligning the center of symmetry of the sub-reference groups with the center of symmetry of the target 110, including: aligning the center of symmetry of each sub-reference group with the center of symmetry of the target 110.

[0111] The compression process refers to compressing multiple detection parameters at all positions along the vertical compression direction of the sub-reference area group reference image into a single detection parameter to obtain a compressed value.

[0112] The compression process includes: calculating a weighted value, squaring the weighted values, or squaring the weighted values. The weighting is achieved by multiplying the parameters by a weighting coefficient and then summing the results; when all weighting coefficients are 1, the weighting is a summation; when all weighting coefficients are 1 and divided by the number of parameters, the weighting is an average.

[0113] In this embodiment, the compression process includes summing multiple detection parameters at all positions along the direction perpendicular to the test direction of the sub-reference area group reference image to obtain the third compression value.

[0114] In other embodiments of the present invention, compressing the detected parameters at all reference positions of the third reference group includes: arbitrarily selecting one detection parameter from the detected parameters at all positions of the sub-reference area group reference image along the third compression direction as the third compression value.

[0115] exist Figure 2 and Figure 3 In the embodiment shown, the test object includes two test targets 110; each test target 110 includes two sub-reference area groups, the two sub-reference area groups include a first sub-reference area group and a second sub-reference area group, the long sides of the sub-reference areas 120 of the first sub-reference area group are all along the X direction, and the short sides of the sub-reference areas 120 of the second reference area group are all along the Y direction.

[0116] In other embodiments of the present invention, a target to be tested 110 may include two or more sub-reference groups.

[0117] The value of the target 110 to be tested needs to be obtained by acquiring the alignment error between the center positions of the target 110 in different layers. The center of each target 110 to be tested needs to be determined by the position of the long side of each marker unit. Therefore, the sub-reference area 120 needs to cover the long side edge of the image of the marker unit. Since there is a lot of noise near the short side of the marker unit, the interference can be reduced by exposing the short side of the marker unit in the sub-reference area 120.

[0118] In another embodiment, the sub-reference area group includes a plurality of sub-reference areas 120, such that the center of symmetry of each sub-reference area group coincides with the center of symmetry of the detection information of the detection mark, including: aligning the center of each sub-reference area 120 with the center of each mark unit.

[0119] The detection information is the detection image, and the detection parameter is the pixel grayscale or light intensity value. Specifically, in this embodiment, the detection parameter is the pixel grayscale.

[0120] In other embodiments, the step of aligning the center of the reference area with the center of the target 110 includes: obtaining a reference position of the edge contour of the target 110 in the detection information; obtaining a reference position of the center of the target 110 in the detection information based on the edge contour; and setting a reference area such that the center of the reference area coincides with the reference position of the center of the target 110 in the detection image.

[0121] In this embodiment, the sub-reference area 120 covers at least a portion of the long side of the marking unit and exposes the short side of the marking unit.

[0122] Continue to refer to Figure 3 Execute step SP3, and perform optimization operations on each direction to be optimized. The optimization operations include the following steps:

[0123] SP31, obtain a judgment function about the size of the reference region along the direction to be optimized, the judgment function being positively or negatively correlated with the signal-to-noise ratio of the detection information of the reference region or the robustness of the target 110 value;

[0124] SP32 optimizes the size of the reference region along the direction to be optimized by the judgment function, so that the judgment function of the reference region meets the preset conditions, and obtains the optimized size of the reference region along the direction to be optimized. The preset conditions include: the optimized reference region judgment function value makes the signal-to-noise ratio or robustness greater than or equal to a preset value, or makes the signal-to-noise ratio or robustness have a maximum value.

[0125] The optimized reference area judgment function value makes the signal-to-noise ratio or robustness greater than or equal to a preset value, including: when the judgment function is positively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target 110 value, the optimized reference area judgment function value is greater than the preset judgment function value; or when the judgment function is negatively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target 110 value, the optimized reference area judgment function value is less than the preset judgment function value.

[0126] The optimal value of the signal-to-noise ratio or robustness includes: when the judgment function is positively correlated with the signal-to-noise ratio of the detection information in the reference area or the robustness of the target 110 value, the optimized judgment function in the reference area has a maximum value; or when the judgment function is negatively correlated with the signal-to-noise ratio of the detection information in the reference area or the robustness of the target 110 value, the optimized judgment function in the reference area has a minimum value.

[0127] The optimized reference area judgment function has a maximum value when it is greater than or equal to the maximum value among all the judgment values ​​obtained by the judgment function; the optimized reference area judgment function has a minimum value when it is less than or equal to the minimum value among all the judgment values ​​obtained by the judgment function.

[0128] The technical solution of the present invention can optimize the dimensions of the reference area in several or more directions, and the direction of the dimension to be optimized is the direction to be optimized.

[0129] Figure 4 and Figure 5 This is a flowchart of the optimized operation steps in one embodiment of the detection method of the present invention.

[0130] The following combination Figure 4 and Figure 5 The specific steps for optimization are explained in detail.

[0131] refer to Figure 4 Execute step S31, set a judgment function about the size of the reference area, the judgment function being positively or negatively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target 110 value.

[0132] The judgment function characterizes the functional relationship between the size of the reference region along the direction to be optimized and the signal-to-noise ratio of the detection information of the reference region, or the functional relationship between the size of the reference region along the direction to be optimized and the robustness of the target 110 value.

[0133] In this embodiment, the reference region includes multiple symmetrically arranged sub-reference regions 120, each sub-reference region 120 being rectangular; the dimension of the reference region along the optimization direction is the dimension of the sub-reference region 120 along the optimization direction. Optimizing the dimension of the reference region along the optimization direction using the judgment function includes: optimizing the dimension of the sub-reference region 120 along the optimization direction using the judgment function. In other embodiments, the reference region includes multiple asymmetrical or symmetrically arranged sub-reference regions 120, and the dimension of the reference region along the optimization direction is the dimension of any one of the sub-reference regions 120 along the optimization direction.

[0134] In this embodiment, the one or more directions to be optimized include a first direction and a third direction. The judgment function for optimizing the size of the reference region along the first direction is a first judgment function; the first judgment function is positively or negatively correlated with the signal-to-noise ratio of the detection information of the reference region; the judgment function for optimizing the size of the reference region along the third direction is a second judgment function, and the robustness of the measured target 110 value of the second judgment function is positively or negatively correlated.

[0135] In this embodiment, the first direction is the length direction of the sub-reference area 120, and the third direction is the width direction of the sub-reference area 120.

[0136] In this embodiment, a coordinate system is established with the first direction and the second direction as coordinate axes. The second direction is perpendicular to the first direction, or the second direction and the first direction have an acute angle between them.

[0137] The first judgment function is the relationship between the sum of signal strength and the sum of noise strength of the detected information.

[0138] Specifically, the sum of the detection parameters at all positions along the first direction of the detection information in the reference area is the compressed value; there is a corresponding relationship between each position and the compressed value in the second direction, and the second direction is perpendicular to the first direction; the function value of the first judgment function is positively correlated with the first ratio between the sum of the detection parameters of each compressed value and the discrete value in the corresponding relationship.

[0139] The positive correlation value of the first ratio is the first ratio itself, the logarithm of the first ratio, or the 1 / 2 power of the first ratio. In this embodiment, the positive correlation value is the logarithm of the first ratio, as follows:

[0140]

[0141]

[0142] Where SN is the first judgment function, n is the coordinate along the second direction, and m is the coordinate in the first direction; I(n,m) is the detection parameter with reference position (n,m); σ n (I(m,n)) represents the standard deviation of the detection parameter with coordinate n along the second direction in the reference area detection information. Alternatively, in other embodiments, n is the coordinate along the first direction, m is the coordinate in the second direction; n is an integer, m is an integer, N is the number of pixels in the reference area along the first direction, the coordinates of the detection information are in units of the distance between adjacent reference positions, and N0 is the minimum coordinate of each position in the reference area.

[0143] Specifically, the detection information is the detection image, and I(n,m) is the pixel grayscale value of the pixel at the reference position (n,m); σ n (I(m,n)) represents the standard deviation of the pixel grayscale values ​​of a column of pixels with coordinates n along the second direction in the reference area detection information, where n and m are both integers.

[0144] In other embodiments of the present invention, if the first judgment function is negatively correlated with the signal-to-noise ratio of the detection information, then the first judgment function is the sum of noise intensities divided by the sum of signal intensities of the detection information, or the first judgment function is the logarithm of the sum of noise intensities divided by the sum of signal intensities of the detection information.

[0145] In another embodiment of the present invention, the sum of the detected parameters at some or all reference positions along the second direction of the reference image in the reference region is a compressed value; there is a correspondence between each position in the first direction and the compressed value, and the correspondence between each position in the first direction and the compressed value is the first judgment function. The sum represents the sum, square, or square of the compressed values ​​in the reference region, and the discrete value represents the degree of dispersion of the compressed values ​​in the reference region.

[0146] The second judgment function is a discrete value of the target 110 obtained from a reference area with different sizes along a third direction.

[0147] refer to Figure 4 SP32 is executed, and the size of the reference region along the direction to be optimized is optimized by the judgment function, so that the judgment function of the reference region meets the preset conditions, and the optimized size of the reference region along the direction to be optimized is obtained. The preset conditions include: the optimized reference region judgment function value makes the signal-to-noise ratio or robustness greater than or equal to the preset value; or makes the signal-to-noise ratio or robustness have a maximum value.

[0148] Please refer to Figure 5 The step of optimizing the size of the reference region along the direction to be optimized using the judgment function, so that the judgment function of the reference region satisfies the preset conditions, includes:

[0149] Please refer to Figure 5 SP321 provides multiple alternative sizes for the reference region along the direction to be optimized;

[0150] The step of setting multiple alternative sizes for the reference area along the direction to be optimized includes: setting multiple reference areas, the centers of the multiple reference areas all coinciding with the center of the target 110 to be measured, and the sizes of the multiple reference areas along the direction to be optimized are different.

[0151] It should be noted that, in this embodiment, the detection method is used to obtain the offset between the centers of multiple targets 110 based on the detection information of the reference area; the sub-reference area groups are respectively used to obtain the offset components of the targets 110 along the width direction of the sub-reference area 120 in each sub-reference area group. Therefore, in this embodiment, the size of the sub-reference area 120 along the width direction is greater than the width of the reference position corresponding to the short side of the target unit; the size of the sub-reference area 120 along the length direction is less than the length of the reference position corresponding to the long side of the target unit.

[0152] Please refer to Figure 5 SP322, the judgment function value for each candidate size is obtained according to the judgment function, and the maximum value among the judgment function values ​​of each candidate size is the reference judgment function value;

[0153] Specifically, obtaining the judgment function value for each candidate size according to the judgment function includes: substituting multiple different candidate sizes into the judgment function to obtain the judgment function value for each candidate size.

[0154] Please refer to Figure 5 SP323 obtains the optimized size of the reference area along the direction to be optimized based on the judgment function value of each candidate size, wherein the optimized size makes the signal-to-noise ratio or robustness greater than or equal to the signal-to-noise ratio or robustness determined by the reference judgment function value.

[0155] In this embodiment, the optimization operation for each direction to be optimized includes: optimizing the size of the reference area along the first direction using the first judgment function; and optimizing the size of the reference area along the third direction using the second judgment function.

[0156] In other embodiments, the optimization operation for each direction to be optimized includes only: optimizing the size of the reference region along the first direction using the first judgment function; or optimizing the size of the reference region along the third direction using the second judgment function.

[0157] The optimization operation of the size of the reference region along the first direction by the first judgment function includes:

[0158] Please refer to Figure 5 SP321 provides multiple alternative sizes for the reference region along the direction to be optimized.

[0159] The first direction is the length direction of the sub-reference area 120.

[0160] In this embodiment, the detection information is the detection image, and the length of the reference area is the number of pixels in the sub-reference area 120 along the direction to be optimized.

[0161] In one specific embodiment, the alternative size is a positive integer between 3 and 10.

[0162] Please refer to Figure 5 SP322, obtain the judgment function value for each candidate size according to the judgment function, and the maximum value among the judgment function values ​​of each candidate size is the reference judgment function value.

[0163] The detection information is a matrix formed by multiple reference positions and detection information. The reference positions of the reference area can be divided and grouped along any direction; in the step of aligning the center of the reference area with the center of the detection mark, the detection parameters of the reference positions along the first compression direction are used as a first reference group.

[0164] Please refer to Figure 5 In step SP322, the detection information of the reference area consists of multiple first reference groups, each of which includes multiple reference positions having the same position in the first compression direction; the reference positions in different first reference groups have different positions in the first compression direction.

[0165] In the step of optimizing the first direction, the step of obtaining the judgment function value of each candidate size according to the judgment function includes: compressing the detected parameters of some or all reference positions within the first reference group to obtain a first compression value; repeating the compression process for each first reference group to obtain the first compression value of each first reference group, thus obtaining a first correspondence between the position of each first reference group along the second direction and the first compression value, wherein the second direction is perpendicular to the first compression direction; obtaining the ratio between the sum of the first compression values ​​and the discrete value in the first correspondence under each candidate size, thus obtaining the judgment function value of each candidate size; wherein the discrete value characterizes the degree of dispersion of the first compression value within the reference area. It should be noted that, in one embodiment, the direction to be measured is the second direction, and the direction perpendicular to the direction to be measured is the first direction; the compression process and the compression treatment are the same step. In another embodiment, the compression process and the compression treatment are different steps.

[0166] Compression processing is performed on the detected parameters of some or all reference positions within the first reference group, including: compression processing is performed on the detected parameters of all reference positions within the first reference group.

[0167] In this embodiment, the detected parameters of all reference positions within the first reference group are compressed to obtain a first compressed value. Furthermore, the first direction is the same as the first compression direction. In other embodiments, the first direction and the first compression direction have a non-zero angle.

[0168] The detected parameters of some reference positions within the first reference group are compressed, including arbitrarily obtaining one detected parameter from the first reference group as the first compressed value. Specifically, obtaining the judgment function value for each candidate size according to the judgment function includes: substituting multiple different candidate sizes into the first judgment function to obtain the judgment function value for each candidate size.

[0169] The first judgment function is:

[0170]

[0171]

[0172] Where N is the length of the reference region, that is, for reference regions with different alternative sizes, N takes a positive integer between 3 and 10. Substituting N, a positive integer between 3 and 10, into the first judgment function above, multiple judgment function values ​​can be obtained.

[0173] In other embodiments, when the first judgment function is another function, multiple judgment function values ​​are obtained in the same way.

[0174] SP323 obtains the optimized size of the reference area along the direction to be optimized based on the judgment function value of each candidate size, wherein the optimized size makes the signal-to-noise ratio or robustness greater than or equal to the signal-to-noise ratio or robustness determined by the reference judgment function value.

[0175] In this embodiment, obtaining the optimized size of the reference area along the direction to be optimized based on the judgment function value of each candidate size includes: obtaining the candidate size for which the reference judgment function value is obtained as the optimized size of the reference area along the direction to be optimized.

[0176] In other embodiments of the present invention, obtaining the optimized size of the reference region along the direction to be optimized based on the judgment function value of each candidate size includes: obtaining the extreme point of the judgment function; and taking the size of the reference region along the direction to be optimized corresponding to the extreme point as the optimized size. The extreme point is greater than or equal to the reference judgment function value. If the candidate size is a discrete value, obtaining the optimized size of the reference region along the direction to be optimized based on the judgment function value of each candidate size includes: obtaining the extreme point of the judgment function and the candidate size corresponding to the extreme point as the extreme size; and obtaining the candidate size with the smallest difference from the extreme size to obtain the optimized size.

[0177] During the optimization of the dimension of the reference region in the first direction using the first judgment function, the dimension of the reference region along the second direction is a fixed value. The optimization operation of the dimension of the reference region along the third direction using the second judgment function includes:

[0178] SP321 provides multiple alternative sizes for the reference region along the direction to be optimized, including providing multiple alternative sizes for the reference region along the third direction.

[0179] The one or more directions to be optimized include a third direction, and the judgment function is negatively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target 110 value; setting multiple different alternative sizes for the reference area along the direction to be optimized includes setting multiple different alternative sizes for the reference area along a third direction;

[0180] Specifically, the detection information of each target 110 includes multiple sub-reference area groups, each sub-reference area group includes one or more sub-reference areas 120, each sub-reference area group has a test direction, and each sub-reference area group is used to obtain the target 110 value in the corresponding test direction.

[0181] The sub-reference regions 120 of each sub-reference region group are centrally symmetrical, and the sub-reference regions 120 within the sub-reference region group are arranged along the width direction of the sub-reference regions 120.

[0182] The target to be tested 110 includes one or more detection mark groups, and each detection mark group includes multiple detection marks that are symmetrical about the center of the target to be tested 110; each detection mark is used to obtain the value of the target to be tested 110 in the corresponding test direction.

[0183] The third direction is the width direction of the sub-reference area 120. In another embodiment, the third direction can be the length direction of the sub-reference area 120.

[0184] In this embodiment, the detection information is the detection image, and the width of the reference area is the number of pixels in the reference area along the direction to be optimized.

[0185] In one specific embodiment, the alternative size is a positive integer between 1 and 5.

[0186] Please refer to Figure 6 and Figure 7 SP322, obtain the judgment function value for each candidate size according to the judgment function, and the maximum value among the judgment function values ​​of each candidate size is the reference judgment function value.

[0187] Specifically, the judgment function value for each candidate size is obtained according to the second judgment function, and the maximum or minimum value among the judgment function values ​​for each candidate size is the reference judgment function value.

[0188] Figure 6 and Figure 7 This is a flowchart of the steps for obtaining the judgment function value for each candidate size according to the judgment function in one embodiment of the detection method of the present invention.

[0189] The step of obtaining the judgment function value of each candidate size according to the second judgment function, and taking the maximum or minimum value of the judgment function value of each candidate size as the reference judgment function value, includes:

[0190] Please refer to Figure 6 S3221, the test target 110 value of the test object is obtained according to the detection information of the reference area of ​​each alternative size.

[0191] The detection information of the reference area consists of multiple second reference groups, each including multiple reference positions having the same position in the second compression direction; the reference positions in different second reference groups have different positions in the second compression direction. Obtaining the target value 110 of the test object based on the detection information of the reference area of ​​the candidate size includes: compressing the detected parameters at some or all reference positions of the second reference group to obtain a second compression value; repeating the compression process for each second reference group to obtain the second compression value of each second reference group, thus obtaining a second correspondence between the position of each second reference group along the test direction and the second compression value, wherein the test direction is perpendicular to the second compression direction; performing autocorrelation on the second correspondence within each reference position of the reference area to obtain an autocorrelation function; obtaining the time delay variable that maximizes the autocorrelation function value as a position variable; and obtaining the position information of the center of the target 110 based on the position variable.

[0192] Specifically, in this embodiment, the detected parameters at all reference positions of the second reference group are compressed to obtain a second compressed value. In other embodiments, the detected parameters at some reference positions of the second reference group are compressed, including selecting one detected parameter in the second reference group as the second compressed value.

[0193] By compressing the detected parameters of the second reference group, and by performing autocorrelation on the second correspondence within each reference position of the reference area to obtain the autocorrelation function, and by obtaining the time delay variable that maximizes the value of the autocorrelation function as the position variable, the size of the area to be tested in any direction can be optimized.

[0194] The value of the target to be tested 110 includes: the alignment error between different targets to be tested 110; obtaining the value of the target to be tested 110 of the object to be tested based on the detection information of the reference area of ​​the candidate size further includes: after obtaining the position information of the center of the target to be tested 110 based on the position variable, performing difference processing on the position information of different targets to be tested 110 to obtain the alignment error between the centers of the targets to be tested 110; or...

[0195] The value of the target 110 to be measured includes the position information of the center of the target 110 along the direction to be measured; the reference area includes at least two sub-reference areas 120; the at least two sub-reference areas 120 include a first sub-reference area 120 and a second sub-reference area 120 spaced apart, the first sub-reference area 120 includes at least a portion of the edge contour of one end of the target 110 in the direction to be measured, and the second sub-reference area 120 includes at least a portion of the edge contour of the other end of the target 110 in the direction to be measured.

[0196] In other embodiments, obtaining the target value 110 of the test object based on the detection information of the reference area of ​​the alternative size includes: compressing the detected parameters at all reference positions of the second reference group to obtain a second compressed value; repeating the compression process for each second reference group to obtain the second compressed value of each second reference group, thereby obtaining a second correspondence between the position of each second reference group along the test direction and the second compressed value, wherein the test direction is perpendicular to the second compression direction; and obtaining the target value 110 of the test object 110 along the test direction based on the second correspondence.

[0197] The second compression direction is parallel to or has an acute angle with the direction to be optimized; if the value of the target 110 to be tested includes the size of the target 110 to be tested along the direction to be tested, the value of the target 110 to be tested along the direction to be tested is obtained according to the second correspondence, including: obtaining the gradient value of the second compression value at each reference position in the second correspondence to obtain the gradient relationship; obtaining the reference positions corresponding to the two maxima of the gradient relationship in the reference area to obtain two extreme positions; obtaining the distance between the two extreme positions to obtain the size of the target 110 to be tested along the direction to be tested.

[0198] In this embodiment, the value of the target to be tested 110 includes: the alignment error between different targets to be tested 110; obtaining the value of the target to be tested 110 of the object to be tested based on the detection information of the reference area of ​​the alternative size, further includes: after obtaining the position information of the center of the target to be tested 110 based on the position variable, performing difference processing on the position information of different targets to be tested 110 to obtain the alignment error between the centers of the targets to be tested 110.

[0199] In another embodiment of the present invention, the value of the target 110 to be measured includes the position information of the center of the target 110 along the measurement direction, or the size of the target 110 along the measurement direction; the reference area includes at least two sub-reference areas 120; the at least two sub-reference areas 120 include a first sub-reference area 120 and a second sub-reference area 120, the first sub-reference area 120 includes at least a portion of the edge contour of one end of the target 110 in the measurement direction, and the second sub-reference area 120 includes at least a portion of the edge contour of the other end of the target 110 in the measurement direction; the first sub-reference area 120 and the second sub-reference area 120 are separated; that is, when the value of the target 110 to be measured includes the position information of the center of the target 110 along the measurement direction, by making the reference area include the partial edge contours at both ends of the target 110, while ignoring the middle part of the edge contour, the influence of the middle part of the edge contour during autocorrelation can be reduced, thereby improving the detection accuracy.

[0200] In this embodiment, the value of the target to be tested 110 is the alignment error between the centers of the two targets to be tested 110.

[0201] Specifically, obtaining the target value 110 of the test object based on the detection information of each reference area includes: obtaining the position information of the symmetry center of different detection targets based on the detection information of the reference area; and obtaining the alignment error between different detection targets based on the position information of the symmetry center of different detection targets. The alignment error includes multiple sub-offset errors along different test directions or alignment errors between the centers of different detection targets.

[0202] The compression processing performed on the detected parameters of the second reference group is the second compression processing.

[0203] In this embodiment, the direction to be measured is perpendicular to the direction to be optimized. The first compression process, the second compression process, and the third compression process are the same step. In other embodiments, the first compression process, the second compression process, and the third compression process are different steps.

[0204] In another embodiment of the present invention, obtaining the target value 110 of the test object based on the detection information of the reference area of ​​the alternative size includes: performing a two-dimensional autocorrelation operation on the detection information of the reference area to obtain a two-dimensional autocorrelation function, and obtaining the position information of the symmetry center of the test target based on the time delay variable with the largest value of the two-dimensional autocorrelation function.

[0205] In this embodiment, one-dimensional autocorrelation operations are performed on the detection information of the reference area along different test directions to obtain the position coordinates of the center of symmetry of the detected target along each test direction. The one-dimensional autocorrelation operation includes: compressing the detection parameters of the reference area along the direction perpendicular to the test direction, and obtaining a compressed detection information at each position along the test direction; obtaining the compressed detection information at each position along the test direction to obtain the correspondence between the position information of each point along the test direction and the compressed detection information, thus obtaining a function to be processed; performing autocorrelation operations on the function to be processed to obtain multiple one-dimensional autocorrelation functions; and obtaining the position coordinates of the center of symmetry of the detected target along each test direction based on the time delay variable with the largest value of each one-dimensional autocorrelation function, thus obtaining the position information. The autocorrelation performed on the second correspondence within each reference position of the reference area is one-dimensional autocorrelation.

[0206] The second correspondence can be a discrete correspondence between the position information of each point along the direction to be measured and the second compression value, or it can be a continuous functional correspondence between the position information of each reference point along the direction to be measured and the second compression value.

[0207] The alignment deviation of the center of symmetry of different detection targets obtained from the detection information of the reference area includes: obtaining the position information of the center of symmetry of the detection mark of different detection targets based on the detection information of each sub-reference area group; obtaining the sub-offset error of the detection target along different test directions based on the position information of the center of symmetry of the detection mark of different detection targets, and thus obtaining the alignment error.

[0208] Please refer to Figure 6 S3222, Set a moving window for the alternative size, the moving window includes a preset number of adjacent alternative sizes;

[0209] The preset quantity is 3 to 10, specifically, the preset quantity is 5 or 8.

[0210] Combined with reference Figure 7 S3223, obtain the values ​​of the target 110 with multiple adjacent alternative sizes in the moving window;

[0211] like Figure 7 As shown, W1 to W9 are nine candidate size values ​​arranged in ascending or descending order; S1 to S9 are the target 110 values ​​corresponding to the candidate sizes W1 to W9.

[0212] Specifically, the preset quantity is 5, and obtaining the target 110 values ​​of multiple adjacent candidate sizes in the moving window includes obtaining the offset errors of 5 adjacent candidate sizes in the moving window. For example, W1 to W5 correspond to offset errors S1 to S5 respectively.

[0213] S3224, obtain discrete values ​​of the target 110 values ​​of multiple adjacent candidate sizes in the moving window, and obtain the reference judgment function value;

[0214] like Figure 7 In the embodiment shown, W1 to W5 correspond to offset errors S1 to S5 respectively. The discrete values ​​of S1 to S5 are obtained to obtain the reference judgment function value.

[0215] The discrete value of the target 110 to be measured includes: the variance, root mean square deviation, or standard deviation of the target 110 values ​​of multiple candidate sizes. Specifically, the discrete value is the variance D of the target 110 values ​​of the candidate sizes; then the discrete value of S1 to S5 is D1.

[0216] S3225, Move the moving window to include multiple other adjacent alternative sizes, and repeat the step of obtaining reference judgment function values ​​until multiple reference judgment function values ​​are obtained, where the minimum value among the multiple reference judgment function values ​​is the optimized judgment function value;

[0217] like Figure 7In the illustrated embodiment, by moving the moving window, the moving window is made to include adjacent candidate sizes W2 to W6 respectively. The reference judgment function values ​​corresponding to the adjacent candidate sizes W2 to W6 are S2 to S6, and the reference judgment function values ​​D2 of S2 to S6 are obtained. The moving window is then moved further, making it include adjacent candidate sizes W3 to W7. The reference judgment function values ​​corresponding to the adjacent candidate sizes W3 to W7 are S3 to S7, and the reference judgment function values ​​D3 of S2 to S6 are obtained. The moving window is then moved further, making it include adjacent candidate sizes W3 to W7. For candidate sizes W4 to W8, the reference judgment function values ​​corresponding to adjacent candidate sizes W4 to W8 are S3 to S7, and the reference judgment function values ​​D3 for S2 to S6 are obtained; continue moving the moving window so that the moving window includes adjacent candidate sizes W5 to W9, the reference judgment function values ​​corresponding to adjacent candidate sizes W5 to W9 are S5 to S9, and the reference judgment function values ​​D4 for S2 to S6 are obtained; the minimum value among all reference judgment function values ​​D1 to D4 is the optimized judgment function value. For example, if D3 is the minimum value, then the optimized judgment function value is D3.

[0218] In this embodiment, the step size of the moving window is the same each time, and the specific step size of the moving window each time is a candidate size; in other embodiments, the step size of the moving window can be multiple candidate sizes, and the step size of the moving window each time can be different.

[0219] S3226, Obtain the median of the candidate sizes in the moving window of the optimization judgment function value to obtain the optimized size.

[0220] The median includes the median or mean.

[0221] Specifically, in Figure 7 In the illustrated embodiment, D3 is the value of the optimization judgment function, and the alternative sizes in the moving window of D3 are W3 to W7; when the median value is the median, the optimized size is W5; when the median value is the mean value, the optimized size is (W3+W4+W5+W6+W7) / 5.

[0222] By setting multiple alternative sizes for the reference area and obtaining a target value 110 for each alternative size based on the detection information within the reference area; and constructing a second judgment function based on the dispersion of the target value 110 for each alternative size reference area, the second judgment function can establish a relationship between the robustness of the reference area size and the target value 110, thereby optimizing the reference area size through the second judgment function and improving the robustness of the detection method.

[0223] SP4 obtains the target value 110 of the test object based on the detection information of the reference area after the optimization operation.

[0224] In this embodiment, the value of the target to be tested 110 is the alignment error between the two targets to be tested 110. In other embodiments, the value of the target to be tested 110 is the position information of the center of symmetry of the target to be tested 110.

[0225] Obtaining the target value 110 of the test object based on the detection information of each reference area includes: obtaining the position information of the symmetry center of different detection targets based on the detection information of the reference areas; and obtaining the alignment error between different detection targets based on the position information of the symmetry center of different detection targets. The alignment error includes multiple sub-offset errors along different test directions or offset scalars between the centers of different detection targets.

[0226] In this embodiment, obtaining the position information of the symmetry center of different detection targets based on the detection information of the reference area includes: obtaining the position information of the symmetry center of different detection targets by performing autocorrelation processing on the detection information of the reference area.

[0227] The autocorrelation processing of the detection information in the reference area includes: performing a two-dimensional autocorrelation operation on the detection information in the reference area to obtain a two-dimensional autocorrelation function; and obtaining the position information of the symmetry center of the detection target based on the time delay variable with the largest value of the two-dimensional autocorrelation function; or, performing a one-dimensional autocorrelation operation on the detection information in the reference area along different test directions to obtain the position coordinates of the symmetry center of the detection target along each test direction; the one-dimensional autocorrelation operation includes: performing a fourth compression processing on the detection information in the reference area along the direction perpendicular to the test direction, obtaining a compressed detection information at each position along the test direction; obtaining the compressed detection information at each position along the test direction to obtain the correspondence between the position information of each point along the test direction and the compressed detection information, thus obtaining the function to be processed; performing an autocorrelation operation on the function to be processed to obtain multiple one-dimensional autocorrelation functions; and obtaining the position coordinates of the symmetry center of the detection target along each test direction based on the time delay variable with the largest value of each one-dimensional autocorrelation function, thus obtaining the position information.

[0228] The function to be processed can be a discrete correspondence between the position information of each point along the direction to be tested and the compression detection information, or it can be a continuous functional correspondence between the position information of each point along the direction to be tested and the compression detection information.

[0229] Obtaining the position information of the symmetry center of different detection targets based on the detection information of the reference area includes: obtaining the position information of the symmetry center of the detection mark of different detection targets based on the detection information of each sub-reference area group; obtaining the sub-offset error of the detection target along different test directions based on the position information of the symmetry center of the detection mark of different detection targets, and thus obtaining the alignment error.

[0230] The present invention also provides a detection system for performing the above detection methods, including:

[0231] The acquisition module is used to acquire the detection information of the test object, which has a test target 110 and a reference point. The detection information includes the correspondence between the reference position of the reference point on the surface of the test object and the detection parameters.

[0232] The setting module is used to set a reference area for the detection information. The detection information in the reference area is used to obtain the target value of the target 110 to be tested. The reference area includes one or more directions to be optimized.

[0233] The optimization module is used to optimize the dimensions of the reference region along each direction to be optimized. The optimization operation includes the following steps:

[0234] A judgment function is obtained regarding the size of the reference region along the direction to be optimized. The judgment function is positively or negatively correlated with the signal-to-noise ratio of the detection information of the reference region or the robustness of the target 110 value. The size of the reference region along the direction to be optimized is optimized using the judgment function to make the judgment function of the reference region meet preset conditions, thereby obtaining the optimized size of the reference region along the direction to be optimized. The preset conditions include: the optimized judgment function value of the reference region makes the signal-to-noise ratio or robustness greater than or equal to a preset value, or makes the signal-to-noise ratio or robustness have a maximum value.

[0235] The processing module is used to obtain the target value of the target 110 based on the detection information of the reference area after the optimization operation.

[0236] The optimized reference area judgment function value makes the signal-to-noise ratio or robustness greater than or equal to a preset value, including: when the judgment function is positively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target 110 value, the optimized reference area judgment function value is greater than the preset judgment function value; or when the judgment function is negatively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target 110 value, the optimized reference area judgment function value is less than the preset judgment function value.

[0237] The optimal value of the signal-to-noise ratio or robustness includes: when the judgment function is positively correlated with the signal-to-noise ratio of the detection information in the reference area or the robustness of the target 110 value, the optimized judgment function in the reference area has a maximum value; or when the judgment function is negatively correlated with the signal-to-noise ratio of the detection information in the reference area or the robustness of the target 110 value, the optimized judgment function in the reference area has a minimum value.

[0238] The optimization module is specifically used for: setting multiple different candidate sizes for the reference region along the direction to be optimized; obtaining the judgment function value for each candidate size according to the judgment function, wherein the maximum or minimum value among the judgment function values ​​of each candidate size is the reference judgment function value; obtaining the optimized size of the reference region along the direction to be optimized according to the judgment function value of each candidate size, wherein the optimized size makes the signal-to-noise ratio or robustness greater than or equal to the signal-to-noise ratio or robustness determined by the reference judgment function value.

[0239] The optimization module is specifically used for: including a first direction among the one or more directions to be optimized; the judgment function is a first judgment function; the first judgment function includes the functional relationship between the signal-to-noise ratio of the detection information of the reference area and the size of the reference area along the direction to be optimized; performing optimization operations on each direction to be optimized includes: optimizing the first direction through the first judgment function; setting multiple different alternative sizes for the reference area along the direction to be optimized includes: setting multiple different alternative sizes for the reference area along the first direction;

[0240] The detection information of the reference area is composed of multiple first reference groups, each of which includes multiple reference positions that have the same position in the first compression direction; the reference positions in different first reference groups have different positions in the first compression direction.

[0241] In the step of optimizing the first direction, the optimization module is specifically used to: compress the detected parameters of some or all reference positions in the first reference group to obtain a first compression value; repeat the compression process for each first reference group to obtain the first compression value of each first reference group, and obtain a first correspondence between the position of each first reference group along the second direction and the first compression value, wherein the second direction is perpendicular to the first compression direction; and obtain the ratio between the sum of the first compression values ​​and the discrete values ​​in the first correspondence under each candidate size to obtain the judgment function value of each candidate size.

[0242] The discrete value represents the degree of dispersion of the first compression value within the reference region.

[0243] The first direction and the first compression direction have a non-zero angle; or, the detected parameters of some or all reference positions in the first reference group are compressed, including: the detected parameters of each reference position in the first reference group are compressed.

[0244] The detection parameters include grayscale values ​​or light intensity values; the sum includes the sum of each first compression value in the reference area, or the square of the sum of each first compression photometric value in the reference area, or the sum of the squares of each first compression value in the reference area; the photometric dispersion includes the variance, root mean square, or standard deviation of each first compression value in the reference area.

[0245] The compression process includes: weighting the detection information, squaring the weights, or squaring the weights; the weighting is to sum the parameters after multiplying them by a weighting coefficient.

[0246] Obtaining the optimized size of the reference region along the optimization direction based on the judgment function values ​​of each candidate size, wherein the optimized size makes the signal-to-noise ratio or robustness greater than or equal to the reference judgment function value includes: fitting the correspondence between the candidate sizes and the judgment function values ​​to obtain a fitting function; obtaining the candidate size of the reference region corresponding to the extreme value of the fitting function to obtain the optimized size; or, obtaining the extreme value among each judgment function value to obtain the optimized judgment function value; obtaining the candidate size of the reference region corresponding to the optimized judgment function value to obtain the optimized size.

[0247] The one or more directions to be optimized include a third direction; the judgment function is negatively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target 110 value; the judgment function is a second judgment function; optimizing each direction to be optimized includes: optimizing the first direction through the second judgment function; setting multiple different alternative sizes for the reference area along the direction to be optimized includes: setting multiple different alternative sizes for the reference area along the third direction.

[0248] In the optimization operation of the first direction through the second judgment function, the optimization module is specifically used for: obtaining the target 110 value of the test object based on the detection information of the reference area of ​​each candidate size; setting a moving window for the candidate size, the moving window including a preset number of adjacent candidate sizes; obtaining the target 110 values ​​of multiple adjacent candidate sizes in the moving window; obtaining discrete values ​​of the target 110 values ​​of multiple adjacent candidate sizes in the moving window to obtain a judgment function value; moving the moving window to include other adjacent candidate sizes, and repeating the step of obtaining the judgment function value until multiple judgment function values ​​are obtained, the minimum value among the multiple judgment function values ​​is the reference judgment function value; obtaining the median value of the candidate sizes in the moving window of the reference judgment function value to obtain the optimized size.

[0249] In one embodiment, the detection information of the reference area consists of multiple second reference groups, each including multiple reference positions having the same position in the second compression direction; the reference positions in different second reference groups have different positions in the second compression direction; obtaining the target value 110 of the test object based on the detection information of the reference area of ​​the candidate size includes: compressing the detected parameters at all reference positions of the second reference group to obtain a second compression value; repeating the compression process for each second reference group to obtain the second compression value of each second reference group, thereby obtaining a second correspondence between the position of each second reference group along the test direction and the second compression value, wherein the test direction is perpendicular to the second compression direction; and obtaining the target value 110 of the test object 110 along the test direction based on the second correspondence.

[0250] In another embodiment, the second compression direction is parallel to or has an acute angle with the direction to be optimized; if the value of the target 110 to be tested includes the size of the target 110 to be tested along the direction to be tested, the value of the target 110 to be tested along the direction to be tested is obtained according to the second correspondence, including: obtaining the gradient value of the second compression value at each reference position in the second correspondence to obtain the gradient relationship; obtaining the reference positions corresponding to the two maxima of the gradient relationship in the reference region to obtain two extreme positions; obtaining the distance between the two extreme positions to obtain the size of the target 110 to be tested along the direction to be tested.

[0251] The direction to be measured is parallel to or has an acute angle with the direction to be optimized.

[0252] In one embodiment, the detection information of the reference region consists of a plurality of second reference groups, each second reference group including a plurality of reference positions having the same position in the second compression direction; the reference positions in different second reference groups have different positions in the second compression direction;

[0253] Obtaining the target value 110 of the test object based on the detection information of the reference area of ​​the alternative size includes: compressing the detected parameters at some or all reference positions of the second reference group to obtain a second compressed value; repeating the compression process for each second reference group to obtain the second compressed value of each second reference group, and obtaining a second correspondence between the position of each second reference group along the test direction and the second compressed value, wherein the test direction is perpendicular to the second compression direction; performing autocorrelation on the second correspondence within each reference position of the reference area to obtain an autocorrelation function; obtaining the time delay variable that maximizes the value of the autocorrelation function as a position variable; and obtaining the position information of the center of the target 110 based on the position variable.

[0254] The value of the target to be tested 110 includes: the alignment error between different targets to be tested 110; obtaining the value of the target to be tested 110 of the object to be tested based on the detection information of the reference area of ​​the candidate size further includes: after obtaining the position information of the center of the target to be tested 110 based on the position variable, performing difference processing on the position information of different targets to be tested 110 to obtain the alignment error between the centers of the targets to be tested 110; or...

[0255] The value of the target 110 to be measured includes the position information of the center of the target 110 along the direction to be measured; the reference area includes at least two sub-reference areas 120; the at least two sub-reference areas 120 include a first sub-reference area 120 and a second sub-reference area 120 spaced apart, the first sub-reference area 120 includes at least a portion of the edge contour of one end of the target 110 in the direction to be measured, and the second sub-reference area 120 includes at least a portion of the edge contour of the other end of the target 110 in the direction to be measured.

[0256] The one or more directions to be optimized include a first direction and a third direction;

[0257] The optimization operation for each direction to be optimized includes: optimizing the size of the reference area along the first direction; and optimizing the size of the reference area along the third direction.

[0258] The target to be tested 110 includes at least two marker units arranged symmetrically at the center, and the marker units are rectangular; the reference area is rectangular, the first direction is the length direction of the marker unit, the third direction is the width direction of the marker unit, and the third direction is perpendicular to the first direction.

[0259] The discrete values ​​of the target 110 to be measured are: the variance, root mean square, or standard deviation of the target 110 values ​​of multiple alternative sizes; the median includes the mean or median.

[0260] The shape of the reference area is the same as the shape of the target 110 to be tested; setting a reference area for the detection information includes: setting a reference area for the detection information such that the center of the reference area coincides with the center of the detection information of the target 110 to be tested.

[0261] The test object includes multiple test targets 110, which are located in different layers of the test object or formed in different processes. Each test target 110 includes multiple detection marks symmetrically arranged about the center of symmetry of the test target 110. The value of each test target 110 includes the alignment error between the centers of symmetry of different test targets 110.

[0262] In one embodiment, setting a reference area for the detection information, such that the center of the reference area coincides with the center of the detection information of the target 110 to be tested, includes: setting a sub-reference area group for the detection information of each detection mark, wherein the shape of the sub-reference area group is the same as the shape of the detection mark, and the reference area includes one or more of the sub-reference area groups; and ensuring that the center of symmetry of each sub-reference area group coincides with the center of symmetry of the detection information of the detection mark.

[0263] The detection information of the reference area is composed of multiple third reference groups, each of which includes multiple reference positions having the same position in the third compression direction; the reference positions in different third reference groups have different positions in the third compression direction.

[0264] In another embodiment, aligning the center of symmetry of each sub-reference group with the center of symmetry of the detection information of the detection marker includes: compressing the detection parameters at all reference positions of the third reference group to obtain a third compressed value; repeating the compression process for each third reference group to obtain the third compressed value of each third reference group, thereby obtaining a third correspondence between the position of each third reference group along the test direction and the third compressed value; obtaining the autocorrelation function of the third correspondence with the test direction perpendicular to the third compression direction, and obtaining the time delay variable that maximizes the value of the autocorrelation function as the position variable; and obtaining the reference position of the center of the reference region based on the position variable.

[0265] Alternatively, the detection mark may include multiple mark units; the sub-reference area group may include multiple sub-reference areas 120, such that the center of symmetry of each sub-reference area group coincides with the center of symmetry of the detection information of the detection mark; including: aligning the center of each sub-reference area 120 with the center of each mark unit.

[0266] The detection information is the detection image, and the detection parameter is the pixel grayscale or light intensity value.

[0267] The detection system of the present invention is used to perform the detection method described in any of the above embodiments.

[0268] This application also provides a storage medium for performing the detection method of any of the above embodiments.

[0269] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. A detection method, characterized in that, include: Acquire detection information of the test object, wherein the test object has a test target and a reference point, and the detection information includes the correspondence between the reference position of the reference point on the surface of the test object and the detection parameters; A reference area is set for the detection information, the center of the reference area coincides with the center of the detection information of the target to be tested, the detection information of the reference area is used to obtain the target value of the target to be tested, and the reference area includes one or more directions to be optimized, including a first direction; The dimensions of the reference region along each direction to be optimized are optimized, and the optimization operation includes the following steps: A judgment function is obtained regarding the size of the reference region along the direction to be optimized. The judgment function is positively or negatively correlated with the signal-to-noise ratio of the detection information of the reference region or the robustness of the target value to be measured. The size of the reference region along the direction to be optimized is optimized using the judgment function to make the judgment function of the reference region meet preset conditions, thereby obtaining the optimized size of the reference region along the direction to be optimized. The preset conditions include: the optimized judgment function value of the reference region makes the signal-to-noise ratio or robustness greater than or equal to a preset value, or makes the signal-to-noise ratio or robustness have a maximum value. The target value of the target to be tested is obtained based on the detection information of the reference area after the optimization operation; The step of optimizing the size of the reference region along the optimization direction using the judgment function to make the judgment function of the reference region meet preset conditions includes: setting multiple different candidate sizes for the reference region along the optimization direction; obtaining the judgment function value for each candidate size according to the judgment function, wherein the maximum or minimum value among the judgment function values ​​of each candidate size is a reference judgment function value; obtaining the optimized size of the reference region along the optimization direction according to the judgment function values ​​of each candidate size, wherein the optimized size makes the signal-to-noise ratio or robustness greater than or equal to the signal-to-noise ratio or robustness determined by the reference judgment function value; The judgment function is a first judgment function; the first judgment function includes the functional relationship between the signal-to-noise ratio of the detection information of the reference area and the size of the reference area along the direction to be optimized; performing optimization operations on each direction to be optimized includes: performing optimization operations on the first direction through the first judgment function; setting multiple different alternative sizes for the reference area along the direction to be optimized includes: setting multiple different alternative sizes for the reference area along the first direction; The detection information of the reference area consists of multiple first reference groups, each first reference group including multiple reference positions having the same position in the first compression direction; the reference positions in different first reference groups have different positions in the first compression direction; in the step of optimizing the first direction, the step of obtaining the judgment function value of each candidate size according to the judgment function includes: compressing the detection parameters of some or all reference positions in the first reference group to obtain a first compression value; repeating the compression process for each first reference group to obtain the first compression value of each first reference group, obtaining a first correspondence between the position of each first reference group along the second direction and the first compression value, wherein the second direction is perpendicular to the first compression direction; obtaining the ratio between the sum of each first compression value and the discrete value in the first correspondence under each candidate size, respectively, to obtain the judgment function value of each candidate size; wherein the discrete value characterizes the degree of dispersion of the first compression value in the reference area.

2. The detection method according to claim 1, characterized in that, The optimized reference area judgment function value makes the signal-to-noise ratio or robustness greater than or equal to a preset value, including: when the judgment function is positively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target value to be measured, the optimized reference area judgment function value is greater than the preset judgment function value; or when the judgment function is negatively correlated with the signal-to-noise ratio of the detection information of the reference area or the robustness of the target value to be measured, the optimized reference area judgment function value is less than the preset judgment function value. Maximizing the signal-to-noise ratio or robustness includes: when the judgment function is positively correlated with the signal-to-noise ratio of the detection information in the reference region or the robustness of the target value to be measured, the optimized judgment function in the reference region has a maximum value; or when the judgment function is negatively correlated with the signal-to-noise ratio of the detection information in the reference region or the robustness of the target value to be measured, the optimized judgment function in the reference region has a minimum value.

3. The detection method according to claim 1, characterized in that, The first direction and the first compression direction have a non-zero angle; or, Compression processing is performed on the detected parameters of some or all reference positions within the first reference group, including: compression processing is performed on the detected parameters of each reference position within the first reference group.

4. The detection method according to claim 1, characterized in that, The detection parameters include grayscale values ​​or light intensity values; the sum includes the sum of each first compression value in the reference area, or the square of the sum of each first compression photometric value in the reference area, or the sum of the squares of each first compression value in the reference area. The photometric discrete values ​​include: the variance, root mean square, or standard deviation of each first compression value within the reference region.

5. The detection method according to claim 1, characterized in that, The step of obtaining the optimized size of the reference region along the direction to be optimized based on the judgment function values ​​of each candidate size, wherein the optimized size makes the signal-to-noise ratio or robustness greater than or equal to the reference judgment function value, includes: fitting the correspondence between the candidate sizes and the judgment function values ​​to obtain a fitting function; obtaining the candidate size of the reference region corresponding to the extreme value of the fitting function to obtain the optimized size; or... The optimal judgment function value is obtained by finding the maximum value among all judgment function values; the optimal size is obtained by finding the alternative size of the reference area corresponding to the optimized judgment function value.

6. The detection method according to any one of claims 1 to 5, characterized in that, The one or more directions to be optimized include a third direction; the judgment function is negatively correlated with the signal-to-noise ratio of the detection information in the reference region or the robustness of the target value to be measured; the judgment function is a second judgment function. The optimization operation for each direction to be optimized includes: optimizing the first direction using a second judgment function; Setting multiple different alternative sizes for the reference area along the direction to be optimized includes: setting multiple different alternative sizes for the reference area along a third direction; In the optimization operation of the first direction through the second judgment function, the judgment function value of each candidate size is obtained according to the judgment function, and the maximum value among the judgment function values ​​of each candidate size is the reference judgment function value, including: The target value of the object to be tested is obtained based on the detection information of the reference area of ​​each alternative size; A movable window is set for the alternative sizes, and the movable window includes a preset number of adjacent alternative sizes; Obtain the target value of multiple adjacent candidate sizes in the moving window; The discrete values ​​of the target to be measured in multiple adjacent candidate sizes within the moving window are obtained to obtain the judgment function value; The moving window is moved to include multiple other adjacent alternative sizes, and the step of obtaining the judgment function value is repeated until multiple judgment function values ​​are obtained. The minimum value among the multiple judgment function values ​​is the reference judgment function value. The median value of the alternative sizes in the moving window with the reference judgment function value is obtained to obtain the optimized size.

7. The detection method according to claim 6, characterized in that, The detection information of the reference area consists of multiple second reference groups, each of which includes multiple reference positions having the same position in the second compression direction. The reference positions in different second reference groups are located in the second compression direction; Obtaining the target value of the test object based on the detection information of the reference area of ​​the alternative size includes: compressing the detected parameters at each reference position of the second reference group to obtain a second compression value; repeating the compression process for each second reference group to obtain the second compression value of each second reference group, and obtaining a second correspondence between the position of each second reference group along the test direction and the second compression value, wherein the test direction is perpendicular to the second compression direction; and obtaining the target value of the test object along the test direction based on the second correspondence.

8. The detection method according to claim 7, characterized in that, The second compression direction is parallel to or has an acute angle with the direction to be optimized; If the target value to be measured includes the size of the target along the direction to be measured, the target value to be measured along the direction to be measured is obtained according to the second correspondence, including: obtaining the gradient value of the second compression value at each reference position in the second correspondence to obtain the gradient relationship; obtaining the reference positions corresponding to the two maxima of the gradient relationship in the reference region to obtain the two extreme positions; obtaining the distance between the two extreme positions to obtain the size of the target along the direction to be measured.

9. The detection method according to claim 6, characterized in that, The detection information of the reference area consists of multiple second reference groups, each of which includes multiple reference positions having the same position in the second compression direction. The reference positions in different second reference groups are located in the second compression direction; Obtaining the target value of the test object based on the detection information of the reference area of ​​the alternative size includes: compressing the detected parameters at some or all reference positions of the second reference group to obtain a second compressed value; repeating the compression process for each second reference group to obtain the second compressed value of each second reference group, and obtaining a second correspondence between the position of each second reference group along the test direction and the second compressed value, wherein the test direction is perpendicular to the second compression direction; performing autocorrelation on the second correspondence within each reference position of the reference area to obtain an autocorrelation function; obtaining the time delay variable that maximizes the value of the autocorrelation function as a position variable; and obtaining the position information of the center of the test object based on the position variable.

10. The detection method according to claim 9, characterized in that, The target value to be measured includes: alignment error between different targets to be measured; obtaining the target value of the object to be measured based on the detection information of the reference area of ​​the candidate size, further includes: after obtaining the position information of the center of the target to be measured based on the position variable, performing difference processing on the position information of different targets to be measured to obtain the alignment error between the centers of the targets to be measured; or... The target value to be measured includes the position information of the center of the target along the direction to be measured; the reference area includes at least two sub-reference areas; the at least two sub-reference areas include a first sub-reference area and a second sub-reference area set at intervals, the first sub-reference area includes at least a portion of the edge contour of one end of the target along the direction to be measured, and the second sub-reference area includes at least a portion of the edge contour of the other end of the target along the direction to be measured.

11. The detection method according to claim 6, characterized in that, The one or more directions to be optimized include a first direction and a third direction; The optimization operation for each direction to be optimized includes: optimizing the size of the reference area along the first direction; The dimensions of the reference region along the third direction are optimized. The target to be tested includes at least two marker units arranged symmetrically at the center, and the marker units are rectangular; the reference area is rectangular, the first direction is the length direction of the marker unit, the third direction is the width direction of the marker unit, and the third direction is perpendicular to the first direction.

12. The detection method according to claim 6, characterized in that, The discrete value of the target value to be measured is: multiple backup values. The variance, root mean square, or standard deviation of the target value to be measured is selected; the median includes the mean or median.

13. The detection method according to claim 1, characterized in that, The shape of the reference area is the same as the shape of the target to be measured.

14. The detection method according to claim 1, characterized in that, The test object includes multiple test targets, which are located in different layers of the test object or formed in different processes. The target to be tested includes multiple detection marks symmetrically arranged about the center of symmetry of the target to be tested; the target value includes: the alignment error between the centers of symmetry of different targets to be tested; Setting a reference area for the detection information, such that the center of the reference area coincides with the center of the detection information of the target to be tested, includes: setting a sub-reference area group for the detection information of each detection mark, wherein the shape of the sub-reference area group is the same as the shape of the detection mark, and the reference area includes one or more of the sub-reference area groups; and ensuring that the center of symmetry of each sub-reference area group coincides with the center of symmetry of the detection information of the detection mark.

15. The detection method according to claim 14, characterized in that, The detection information of the reference area consists of multiple third reference groups, which include multiple reference positions having the same position in the third compression direction; The reference positions in different third reference groups are located in different directions of the third compression. To ensure that the center of symmetry of each sub-reference group coincides with the center of symmetry of the detection information of the detection mark, the following steps are taken: compressing the detection parameters at all reference positions of the third reference group to obtain a third compressed value; repeating the compression process for each third reference group to obtain the third compressed value for each third reference group, thereby obtaining a third correspondence between the position of each third reference group along the test direction and the third compressed value, wherein the test direction is perpendicular to the third compression direction. Obtain the autocorrelation function of the third correspondence, and obtain the time delay variable that maximizes the value of the autocorrelation function as the position variable; obtain the reference position of the center of the reference area based on the position variable; Alternatively, the detection mark may include multiple mark units; the sub-reference area group may include multiple sub-reference areas, such that the center of symmetry of each sub-reference area group coincides with the center of symmetry of the detection information of the detection mark; including: aligning the center of each sub-reference area with the center of each mark unit.

16. The detection method according to claim 7, 9, or 15, characterized in that, The compression process includes: weighting the detection information, squaring the weights, or squaring the weights; the weighting is to sum the parameters after multiplying them by a weighting coefficient.

17. The detection method according to claim 1, characterized in that, The detection information is the detection image, and the detection parameter is the pixel grayscale or light intensity value.

18. A detection system, characterized in that, include: A processor for performing the detection method according to any one of claims 1 to 17.

19. A storage medium, characterized in that, The storage medium stores one or more computer instructions, which are used to implement the detection method for the analyte as described in any one of claims 1 to 16.

Images