Phase unwrapping method, device, equipment, medium and product

Abstract

The invention discloses a phase unwrapping method, device and equipment, a medium and a product. The method comprises the following steps: acquiring a to-be-processed interference fringe image corresponding to the surface of a wafer and unwrapping operation prompt information; wherein the unwrapping operation prompt information is used for prompting a logical reasoning process of phase extraction and phase unwrapping, an intermediate reasoning result and a discrimination mode of abnormal phase jump; and processing the to-be-processed interference fringe image and the unwrapping operation prompt information according to an unwrapping image generation model obtained by pre-training, and obtaining an unwrapping image corresponding to the to-be-processed interference fringe image, so as to determine surface shape information corresponding to the wafer surface based on the unwrapping image. According to the technical scheme provided by the embodiment of the invention, in the process of performing phase unwrapping on the phase image, the generated abnormal phase jump is effectively processed, and the unwrapping image with higher precision is generated.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a phase unwrapping method, device, equipment, medium and product. Background Art

[0002] In order to accurately restore the surface shape information of the wafer, phase extraction and unwrapping constitute the key steps in the interferometric measurement process. Due to the limitations of the interference fringe denoising method, the residual noise may cause abnormal phase jumps during the phase extraction and unwrapping process, resulting in errors in the phase unwrapping results.

[0003] In the related technology, in the current interference fringe phase extraction and unwrapping method, there may be a problem of phase unwrapping discontinuity caused by improper handling of abnormal phase jumps, and the abnormal phase jumps cannot be effectively handled, which in turn causes the calculated wafer surface shape to deviate from the actual situation, affecting the accuracy of the wafer surface shape solution. Summary of the invention

[0004] The present invention provides a phase unwrapping method, device, equipment, medium and product to achieve the effect of effectively processing abnormal phase jumps generated during phase unwrapping of a phase image and generating an unwrapped image with higher accuracy.

[0005] According to one aspect of the present invention, a phase unwrapping method is provided, the method comprising:

[0006] Obtaining the interference fringe image to be processed corresponding to the wafer surface and unwrapping operation prompt information; wherein the unwrapping operation prompt information is used to prompt the logical reasoning process of phase extraction and phase unwrapping, the intermediate reasoning results, and the judgment method of abnormal phase jump;

[0007] Processing the interference fringe image to be processed and the unwrapping operation prompt information according to a pre-trained unwrapping image generation model, and obtaining an unwrapping image corresponding to the interference fringe image to be processed, so as to determine the surface shape information corresponding to the wafer surface based on the unwrapping image;

[0008] Among them, the unwrapping image generation model at least includes a feature extraction sub-model, an unwrapping feature generation sub-model, a phase unwrapping feature generation sub-model and a feature conversion sub-model; the feature extraction sub-model is used to extract the image features of the interference fringe image to be processed; the unwrapping feature generation sub-model is used to generate an unwrapping operation prompt vector according to the unwrapping operation prompt information; the phase unwrapping feature generation sub-model is used to generate a phase unwrapping image feature corresponding to the interference fringe image to be processed based on the image features and the unwrapping operation prompt vector; the feature conversion sub-model is used to convert the phase unwrapping image features into an unwrapping image corresponding to the interference fringe image to be processed.

[0009] According to another aspect of the present invention, a phase unwrapping device is provided, the device comprising:

[0010] An image acquisition module, used to acquire the interference fringe image to be processed corresponding to the wafer surface and unwrapping operation prompt information; wherein the unwrapping operation prompt information is used to prompt the logical reasoning process of phase extraction and phase unwrapping, the intermediate reasoning results and the judgment method of abnormal phase jump;

[0011] A phase unwrapping module is used to process the interference fringe image to be processed and the unwrapping operation prompt information according to a pre-trained unwrapping image generation model, and obtain an unwrapping image corresponding to the interference fringe image to be processed, so as to determine the surface information corresponding to the wafer surface based on the unwrapping image; wherein the unwrapping image generation model at least includes a feature extraction submodel, an unwrapping feature generation submodel, a phase unwrapping feature generation submodel and a feature conversion submodel; the feature extraction submodel is used to extract the image features of the interference fringe image to be processed; the unwrapping feature generation submodel is used to generate an unwrapping operation prompt vector according to the unwrapping operation prompt information; the phase unwrapping feature generation submodel is used to generate a phase unwrapping image feature corresponding to the interference fringe image to be processed based on the image features and the unwrapping operation prompt vector; the feature conversion submodel is used to convert the phase unwrapping image feature into an unwrapping image corresponding to the interference fringe image to be processed.

[0012] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0013] at least one processor; and

[0014] a memory communicatively connected to the at least one processor; wherein,

[0015] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the phase unwrapping method described in any embodiment of the present invention.

[0016] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the phase unwrapping method described in any embodiment of the present invention when executed.

[0017] According to another aspect of the present invention, a computer program product is provided. The computer program product comprises a computer program. When the computer program is executed by a processor, the phase unwrapping method according to any embodiment of the present invention is implemented.

[0018] The technical solution of the embodiment of the present invention provides a data basis and operation guidance for subsequent phase unwrapping by obtaining the interference fringe image to be processed and the unwrapping operation prompt information corresponding to the wafer surface, and the unwrapping operation prompt information is at least used to prompt the logical reasoning process of phase extraction and phase unwrapping, the intermediate reasoning results and the judgment method of abnormal phase jumps, and provides operation guidance for judging and processing abnormal phase jumps in the phase unwrapping process; further, the interference fringe image to be processed and the unwrapping operation prompt information are processed according to the pre-trained unwrapping image generation model, and the unwrapping image corresponding to the interference fringe image to be processed is obtained, so as to determine the surface information corresponding to the wafer surface based on the unwrapping image, solves the problem in the related art that the abnormal phase jump cannot be effectively processed during the phase extraction and phase unwrapping process, thereby causing discontinuous phase unwrapping, and realizes the effect of effectively processing the abnormal phase jump generated in the process of phase unwrapping of the phase image, and generating a more accurate unwrapping image, and by using the unwrapping operation prompt information as the model input object, it provides operation guidance for the model to perform abnormal phase jump judgment and processing, ensuring the accuracy and continuity of the phase information in the wrapping image.

[0019] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 is a flow chart of a phase unwrapping method provided according to Embodiment 1 of the present invention;

[0022] Figure 2 is a flow chart of a phase unwrapping method provided according to Embodiment 2 of the present invention;

[0023] Figure 3 is a flow chart of a phase unwrapping method provided according to Embodiment 2 of the present invention;

[0024] Figure 4 is a flow chart of a phase unwrapping method provided according to Embodiment 3 of the present invention;

[0025] Figure 5 is a structural schematic diagram of a phase unwrapping device provided according to Embodiment 4 of the present invention;

[0026] Figure 6 It is a schematic diagram of the structure of an electronic device for implementing the phase unwrapping method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] Embodiment 1

[0030] Figure 1 1 is a flow chart of a phase unwrapping method provided in the first embodiment of the present invention. This embodiment is applicable to the case where phase extraction and phase unwrapping are performed on the interference fringe image after denoising. The method can be performed by a phase unwrapping device. The phase unwrapping device can be implemented in the form of hardware and / or software. The phase unwrapping device can be configured in a terminal and / or a server. Figure 1 As shown, the method includes:

[0031] S110, obtaining the interference fringe image to be processed corresponding to the wafer surface and unpacking operation prompt information.

[0032] Among them, the wafer refers to the silicon wafer used in the manufacture of semiconductor integrated circuits. Because of its circular shape, it is called a wafer. In the manufacturing process of high-end computing chips, the wafer is a key material, and the shape accuracy of its surface has a significant impact on the implementation of manufacturing technology. Therefore, accurate measurement of the wafer surface is an indispensable step in semiconductor manufacturing, which is not only crucial to improving the yield of the wafer, but also the basis for ensuring the reliability of the subsequent chip manufacturing process. In general, the surface structure of the wafer surface is analyzed, usually including four main steps: image preprocessing, phase extraction, phase unwrapping and wavefront fitting. In this embodiment, the interference fringe image obtained after image preprocessing is mainly phase extracted and phase unwrapped to obtain an unwrapped image. The wafer surface refers to the outermost layer of the wafer. The interference fringe image to be processed can be an interference fringe image to be subjected to phase extraction and phase unwrapping processing. The unwrapping operation prompt information can be understood as information that explains the phase extraction and phase unwrapping operations. The unwrapping operation prompt information is at least used to describe the logical reasoning process of phase extraction and phase unwrapping, the intermediate reasoning results, and the judgment method of abnormal phase jump. The logical reasoning process of phase extraction and phase unwrapping can be used to characterize the solution steps required for phase extraction and phase unwrapping of the interference fringe image to be processed. It can be understood that when analyzing the surface structure of the wafer surface, phase extraction usually processes the pre-processed interference fringe image to solve the phase difference caused by the interference of the reference light and the measurement light. Phase unwrapping, also known as phase unwrapping, is to unfold the phase in sequence according to the phase expansion series to form a smooth and continuous phase unwrapping diagram. In addition, in the process of phase unwrapping, the phase jump is processed to solve the phase overlap. The intermediate reasoning result can be used to characterize the intermediate solution result obtained in the solution process. The judgment method of abnormal phase jump can be understood as a method for judging the phase jump to be excluded. Abnormal phase jump can be a phase jump that needs to be excluded. Abnormal phase jump usually refers to a non-real phase jump caused by various reasons (such as noise, interference, equipment failure, etc.) during the signal processing process. The abnormal phase jump has nothing to do with the real phase change of the signal itself, but is caused by external factors. Phase jump refers to the phenomenon that the signal phase suddenly jumps from one value to another in a short period of time.

[0033] In this embodiment, the interference fringe image to be processed may be an interference fringe image obtained after image denoising. Furthermore, the interference fringe image to be processed corresponding to the wafer surface may be obtained by performing image denoising on the interference fringe image corresponding to the wafer surface that has been obtained.

[0034] Based on this, the above-mentioned technical solutions also include: obtaining an original interference fringe image corresponding to the wafer surface; denoising the original interference fringe image according to a preset image denoising method to obtain a processed interference fringe image corresponding to the wafer surface.

[0035] Among them, the original interference fringe image can be an interference fringe image obtained by measuring the wafer surface through an interferometer. Generally, the wafer is placed on the measuring platform of the interferometer to ensure that the wafer surface is perpendicular to the measuring optical path of the interferometer. Further, the light is divided into reference light and measurement light by a beam splitter inside the interferometer. Among them, the reference light passes through a fixed path of the interferometer, and the measurement light is reflected back through the wafer surface. The reference light and the measurement light meet inside the interferometer and produce interference phenomenon to form interference fringes. Further, a shooting device can be used to collect interference fringe images, and the collected interference fringe images can be used as the original interference fringe images corresponding to the wafer surface. Among them, the interferometer used can be a measuring instrument suitable for wafer surface measurement, optionally, a Fizeau interferometer, a Michelson interferometer, etc. The shooting device can be any device capable of image acquisition, optionally, a charge coupled device camera. The image denoising method can be a method of correcting and enhancing the interference fringe image to reduce noise. Optionally, the image denoising method includes frequency domain low-pass filtering, spin filtering, homomorphic filtering, partial differential equation method and wavelet transform, etc.

[0036] As an optional implementation of this embodiment, the wafer can be placed on the measuring platform of the Fizeau interferometer, and further, the light can be divided into reference light and measurement light by a beam splitter inside the interferometer. The reference light and the measurement light meet inside the interferometer and produce interference to form interference fringes. Furthermore, a charge coupled device camera can be used to capture the interference fringe image, and the captured interference fringe image can be used as the original interference fringe image corresponding to the wafer surface. Furthermore, the original interference fringe image can be denoised by a preset denoising method, and the denoised image can be used as the interference fringe image to be processed.

[0037] It should be noted that, in order to clearly and intuitively describe the process of phase extraction and phase unwrapping of interference fringe images, the intermediate reasoning results and the judgment method of abnormal phase jumps can be represented in the form of images. Furthermore, in order to reduce the amount of data processing in the subsequent data processing process and improve data processing efficiency, features can be extracted from the image, and unwrapping operation prompt information can be constructed based on the extracted image features.

[0038] Based on this, on the basis of the above-mentioned technical solutions, it also includes: obtaining unpacking process description information, wherein the unpacking process description information includes step description information of multiple unpacking operation steps, intermediate result images associated with each unpacking operation step, and phase jump discrimination images of abnormal phase jumps; performing feature extraction on the intermediate result image associated with at least one unpacking operation step to obtain intermediate result features associated with at least one unpacking operation step, and performing feature extraction on the phase jump discrimination image to obtain abnormal phase jump discrimination features corresponding to the abnormal phase jump; constructing unpacking operation prompt information based on the step description information of multiple unpacking operation steps, the intermediate result features associated with at least one unpacking operation step, and the abnormal phase jump discrimination features.

[0039] Among them, the unwrapping process description information can be understood as information that explains the phase extraction and phase unwrapping process. The unwrapping operation steps can be the steps performed for phase extraction and phase unwrapping of the interference fringe image. The step description information can be text information that explains the unwrapping operation steps. Exemplarily, the unwrapping process may include 5 unwrapping operation steps, and the step description information of each unwrapping operation step is: step one, interference fringe image modeling; step two, bringing in pixel intensity; step three, least squares method to solve the equation group; step four, inverse tangent phase extraction; step five, phase unwrapping. The intermediate result image can be a reference image that characterizes the execution status and reasoning results of the unwrapping operation steps. The phase jump discrimination image of the abnormal phase jump is used to characterize the discrimination method of the abnormal phase jump. In other words, the abnormal phase jump can be judged and identified based on the phase jump discrimination image.

[0040] As an optional implementation of this embodiment, the unwrapping process description information can be obtained based on the prior indication of the interference fringe phase unwrapping process, and the obtained unwrapping process description information can include step description information of multiple unwrapping operation steps, an intermediate result image associated with at least one unwrapping operation step, and a phase jump discrimination image of an abnormal phase jump. Further, for the intermediate result image associated with at least one unwrapping operation step, feature extraction is performed on the intermediate result image according to a preset feature extraction algorithm, and the extracted image features are used as the intermediate result image associated with the unwrapping operation step. Then, the intermediate result features associated with at least one unwrapping operation step can be obtained. And, feature extraction is performed on the phase jump discrimination image according to a preset feature extraction algorithm, and the extracted image features are used as abnormal phase jump discrimination features corresponding to the abnormal phase jump. Further, the unwrapping operation prompt information can be constructed based on the step description information of multiple unwrapping operation steps, the intermediate result features associated with at least one unwrapping operation step, and the abnormal phase jump discrimination features.

[0041] S120, processing the interference fringe image to be processed and the unwrapping operation prompt information according to the pre-trained unwrapping image generation model, and obtaining an unwrapping image corresponding to the interference fringe image to be processed, so as to determine the surface shape information corresponding to the wafer surface based on the unwrapping image.

[0042] Among them, the unwrapped image generation model can be understood as a deep learning model that generates an unwrapped image that meets the phase unwrapping requirements according to the input object. The unwrapped image generation model can be a deep learning model that takes the interference fringe image to be processed and the unwrapping operation prompt information as input objects, and then performs phase extraction and phase unwrapping processing on the interference fringe image to be processed according to the unwrapping operation prompt information to generate an unwrapped image. Inputting the unwrapping operation prompt information into the unwrapping generation model can be used to guide the unwrapping generation model to perform a specific task to generate a corresponding output. The unwrapped image generation model can be a deep deep learning model including at least one sub-model. The unwrapped image generation model includes at least a feature extraction sub-model, an unwrapping feature generation sub-model, a phase unwrapping feature generation sub-model and a feature conversion sub-model. The unwrapped image can be an image containing the original, continuous phase information. It can be understood that the phase map is a visual representation that shows the phase information distributed in space or time. When the phase change is greater than a certain threshold (such as 2π), the phase information will be "wrapped" or "folded", resulting in the absolute value of the phase value being unable to accurately represent the original information. The goal of unwrapping the image is to eliminate this wrapping effect and restore the true phase information. The surface information can be used to characterize the surface state of the wafer surface. Optionally, the surface information can include information such as surface flatness, smoothness, cleanliness, and morphological characteristics.

[0043] The feature extraction submodel is used to extract image features of the interference fringe image to be processed. The feature extraction submodel can be a neural network model that takes the interference fringe to be processed as an input object to extract features from the input object. The feature extraction submodel can be a neural network model of any model structure. Optionally, the feature extraction submodel can include an encoder and a linear layer. The encoder can be used to extract features from the interference fringe image to be processed. The linear layer can be used to align the extracted features.

[0044] The unwrapping feature generation submodel is used to generate an unwrapping operation prompt vector based on the unwrapping operation prompt information. The unwrapping feature generation submodel can be a deep learning model that takes the unwrapping operation prompt information as an input object and encodes the input object to generate an operation prompt vector. The unwrapping operation prompt vector can be understood as a feature vector that characterizes the phase extraction operation, the phase unwrapping operation, and the discrimination operation of the abnormal phase jump. The unwrapping operation prompt vector can be used as an operation guide for the phase unwrapping feature generation submodel to perform phase extraction, phase unwrapping, and abnormal phase jump discrimination and elimination. The unwrapping feature generation submodel can be a neural network model of any model structure. Optionally, the unwrapping feature generation submodel may include an unwrapping feature generation module and a first feature alignment module. The unwrapping feature generation module may be used to generate an operation prompt feature vector based on the unwrapping operation prompt information. The first feature alignment module may be used to align the generated operation prompt feature vector. The first feature alignment module may include a linear layer. It should be noted that the unwrapping feature generation model may be a neural network model that can process information that fuses visual information and language information. In other words, the unwrapped feature generation module is able to process and understand information from different modalities, i.e., images and texts, and achieve cross-modal interaction and reasoning.

[0045] Among them, the phase unwrapping feature generation submodel is used to generate phase unwrapping image features corresponding to the interference fringe image to be processed based on image features and unwrapping operation prompt vectors. The phase unwrapping feature generation submodel can be a deep learning model that takes image features and unwrapping operation prompt vectors as input objects, and performs phase extraction, phase unwrapping, and abnormal phase jump discrimination and elimination on image features according to the unwrapping operation prompt vector to generate phase unwrapping image features. The phase unwrapping feature generation submodel may include a phase unwrapping feature generation module and a second feature alignment module. The phase unwrapping feature generation module can be used to perform phase extraction, phase unwrapping, and abnormal phase jump discrimination and elimination on image features according to the unwrapping operation prompt vector to generate phase unwrapping image features. The second feature alignment module can be used to align the generated operation prompt feature vectors. The second feature alignment module may include a linear layer.

[0046] The feature conversion submodel can be used to convert the phase unwrapping image features into an unwrapped image corresponding to the interference fringe image to be processed. The feature conversion submodel can be a neural network model that uses the phase unwrapping features as input objects to perform feature conversion on the input objects. The feature conversion submodel can be a neural network model of any model structure. Optionally, the feature conversion submodel can include a decoder.

[0047] In practical applications, the limitations of the image denoising method used in the image preprocessing stage can lead to abnormal phase jumps caused by phase extraction and solution errors. In turn, the phase unwrapping diagram obtained in the phase unwrapping stage shows obvious discontinuities or errors, which have a serious impact on the final result of the subsequent wavefront fitting, thereby affecting the accuracy of the circular surface shape solution.

[0048] In view of the above situation, in this embodiment, the unwrapping operation prompt information and the interference fringe image to be processed after denoising can be obtained. Further, an unwrapping image generation model for performing phase extraction, phase unwrapping, and discrimination and elimination of abnormal phase jumps on the input image to generate an unwrapped image can be pre-trained, and the interference fringe image to be processed and the unwrapping operation prompt information are input into the unwrapping image generation model, and the unwrapping operation prompt information is used as a guide for the unwrapping operation process, and the phase extraction, phase unwrapping, and discrimination and elimination of abnormal phase jumps are performed on the image to be processed according to the unwrapping operation prompt information. Then, the unwrapping image corresponding to the interference fringe image to be processed can be output.

[0049] In this embodiment, after obtaining the interference fringe image to be processed and the unwrapping operation prompt information, the interference fringe image to be processed and the unwrapping operation prompt information can be input into the unwrapping image generation model. Further, the interference fringe image to be processed can be input into the feature extraction sub-model, so as to extract the features of the interference fringe image to be processed based on the feature extraction sub-model to obtain the image features. And, the unwrapping operation prompt information is input into the unwrapping feature generation sub-model, so as to encode the unwrapping operation prompt information based on the unwrapping generation sub-model to obtain the unwrapping operation prompt vector. Further, the image features and the unwrapping operation prompt vector can be input into the phase unwrapping feature generation sub-model, so as to process the image features according to the unwrapping operation prompt vector based on the phase unwrapping feature generation sub-model, and output the unwrapping image corresponding to the interference fringe image to be processed. Further, the unwrapping image can be wavefront fitted to obtain the surface shape information corresponding to the wafer surface.

[0050] The technical solution of the embodiment of the present invention provides a data basis and operation guidance for subsequent phase unwrapping by obtaining the interference fringe image to be processed and the unwrapping operation prompt information corresponding to the wafer surface, and the unwrapping operation prompt information is at least used to prompt the logical reasoning process of phase extraction and phase unwrapping, the intermediate reasoning results and the judgment method of abnormal phase jumps, and provides operation guidance for judging and processing abnormal phase jumps in the phase unwrapping process; further, the interference fringe image to be processed and the unwrapping operation prompt information are processed according to the pre-trained unwrapping image generation model, and the unwrapping image corresponding to the interference fringe image to be processed is obtained, so as to determine the surface information corresponding to the wafer surface based on the unwrapping image, solves the problem in the related art that the abnormal phase jump cannot be effectively processed during the phase extraction and phase unwrapping process, thereby causing discontinuous phase unwrapping, and realizes the effect of effectively processing the abnormal phase jump generated in the process of phase unwrapping of the phase image, and generating a more accurate unwrapping image, and by using the unwrapping operation prompt information as the model input object, it provides operation guidance for the model to perform abnormal phase jump judgment and processing, ensuring the accuracy and continuity of the phase information in the wrapping image.

[0051] Embodiment 2

[0052] Figure 2 It is a flow chart of a phase unwrapping method provided in the second embodiment of the present invention. On the basis of the above-mentioned embodiment, the method of obtaining the unwrapped image is further refined. Optionally, the interference fringe image to be processed and the unwrapping operation prompt information are processed according to the unwrapping image generation model obtained by pre-training, and the unwrapping image corresponding to the interference fringe image to be processed is obtained, including: inputting the interference fringe image to be processed into the feature extraction submodel for feature extraction to obtain the image features corresponding to the interference fringe image to be processed, and inputting the unwrapping operation prompt information into the unwrapping feature generation submodel for information encoding to generate the unwrapping operation prompt vector; inputting the image features and the unwrapping operation prompt vector into the phase unwrapping feature generation submodel to generate the phase unwrapping image features corresponding to the interference fringe image to be processed; inputting the phase unwrapping image features into the feature conversion submodel for feature conversion to obtain the unwrapping image corresponding to the interference fringe image to be processed. The specific implementation method can refer to the technical solution of this embodiment. Among them, the technical terms that are the same or similar to those in the above-mentioned embodiments are not repeated here.

[0053] like Figure 2 As shown, the method includes:

[0054] S210, obtaining the interference fringe image to be processed corresponding to the wafer surface and unpacking operation prompt information.

[0055] S220, inputting the interference fringe image to be processed into the feature extraction sub-model for feature extraction to obtain image features corresponding to the interference fringe image to be processed, and inputting the unpacking operation prompt information into the unpacking feature generation sub-model for information encoding to generate an unpacking operation prompt vector.

[0056] Among them, the image feature can be understood as a feature that represents the deep meaning of the image, and the feature is easy to be understood and processed by the model. The image feature can be represented based on any form of expression, and optionally, the image feature is represented based on a vector form. The unwrapping operation hint vector can be understood as a feature vector that represents the phase extraction operation, the phase unwrapping operation, and the discrimination operation of the abnormal phase jump, which is easy to understand and process the model. The unwrapping operation hint vector can be a text vector that represents the unwrapping operation process.

[0057] As an optional implementation manner of the first embodiment of the present invention, a feature extraction submodel includes an encoder and a linear layer; the interference fringe image to be processed is input into the feature extraction submodel for feature extraction to obtain image features corresponding to the interference fringe image to be processed, including: performing feature extraction on the interference fringe image to be processed based on the encoder to obtain features of the image to be processed; performing feature alignment on the features of the image to be processed based on the linear layer to obtain image features corresponding to the interference fringe image to be processed.

[0058] In this embodiment, the unwrapping feature generation sub-model may include an unwrapping feature generation module and a first feature alignment module. When the unwrapping operation prompt information is input into the unwrapping feature generation sub-model, the unwrapping operation prompt information may be processed based on the unwrapping feature generation model and the first feature alignment module in sequence to obtain an unwrapping operation prompt vector.

[0059] Optionally, the unpacking feature generation sub-model includes an unpacking feature generation module and a first feature alignment module; the unpacking operation prompt information is input into the unpacking feature generation sub-model for information encoding to generate an unpacking operation prompt vector, including: encoding the unpacking operation prompt information based on the unpacking feature generation module to obtain a prompt vector to be processed; aligning the prompt vector to be processed based on the first feature alignment module to obtain an unpacking operation prompt vector.

[0060] The prompt vector to be processed may be a feature vector obtained by encoding the unpacking operation prompt information.

[0061] In this embodiment, when the unpacking prompt information is input into the unpacking feature generation sub-model, the unpacking prompt information can be encoded based on the unpacking feature generation module to obtain a prompt vector to be processed. Further, feature alignment can be performed on the prompt vector to be processed based on the first feature alignment module to obtain an unpacking operation prompt vector.

[0062] S230, inputting the image features and the unwrapping operation prompt vector into the phase unwrapping feature generation sub-model to generate phase unwrapping image features corresponding to the interference fringe image to be processed.

[0063] In this embodiment, after obtaining the image features and the unwrapping operation hint vector, the image features and the unwrapping operation hint vector can be input into the phase unwrapping feature generation sub-model to process the image features according to the unwrapping operation hint vector based on the phase unwrapping feature sub-model and generate the phase unwrapping image features.

[0064] Optionally, the phase unwrapping feature generation submodel includes a phase unwrapping feature generation module and a second feature alignment module; the image features and the unwrapping operation prompt vector are input into the phase unwrapping feature generation submodel to generate phase unwrapping image features corresponding to the interference fringe image to be processed, including: processing the image features and the unwrapping operation prompt vector based on the phase unwrapping feature generation module to generate the unwrapped image features to be processed; and performing feature alignment on the unwrapped image features to be processed based on the second feature alignment module to obtain the phase unwrapping image features corresponding to the interference fringe image to be processed.

[0065] As an optional implementation of this embodiment, when the image features and the unwrapping operation prompt vector are input into the phase unwrapping feature generation submodel, the image features and the unwrapping operation prompt vector can be processed based on the phase unwrapping feature generation module, and the unwrapping image features to be processed can be generated. Further, the unwrapping image features to be processed can be aligned based on the second feature alignment module, and the features obtained after the feature alignment are used as the phase unwrapping image features corresponding to the interference fringe image to be processed.

[0066] S240, inputting the phase unwrapping image features into the feature conversion sub-model for feature conversion to obtain an unwrapped image corresponding to the interference fringe image to be processed, so as to determine the surface shape information corresponding to the wafer surface based on the unwrapped image.

[0067] Among them, the feature conversion sub-model may include a decoder.

[0068] In this embodiment, after obtaining the phase unwrapped image features, the phase unwrapped image features can be input into the feature conversion sub-model. Further, the phase unwrapped image features can be feature converted based on the decoder in the feature conversion sub-model, and an unwrapped image corresponding to the interference fringe image to be processed can be obtained. Then, the surface shape information corresponding to the wafer surface can be determined based on the unwrapped image.

[0069] For example, Figure 3 FIG. 1 is a flow chart of a phase unwrapping method provided according to an embodiment of the present invention. Figure 3 As shown, the interference fringe image to be processed and the unwrapping operation prompt information are input into the unwrapping image generation model. Further, the interference fringe image to be processed can be extracted based on the encoder in the feature extraction submodel to obtain the image features to be processed; the image features to be processed can be aligned based on the linear layer in the feature extraction submodel to obtain the image features corresponding to the interference fringe image to be processed; and the unwrapping prompt information can be encoded based on the unwrapping feature generation module in the unwrapping feature generation submodel to obtain the prompt vector to be processed. Further, the prompt vector to be processed can be aligned based on the first feature alignment module in the unwrapping feature generation submodel to obtain the unwrapping operation prompt vector. Further, the image features and the unwrapping operation prompt vector can be input into the phase unwrapping feature generation submodel, and the image features and the unwrapping operation prompt vector can be processed based on the phase unwrapping feature generation module in the phase unwrapping feature generation submodel to generate the unwrapping image features to be processed. Further, the unwrapping image features to be processed can be aligned based on the second feature alignment module in the phase unwrapping feature generation submodel, and the features obtained after the feature alignment are used as the phase unwrapping image features corresponding to the interference fringe image to be processed. Furthermore, the phase unwrapped image features are input into the feature conversion sub-model. Furthermore, the phase unwrapped image features can be feature converted based on the decoder in the feature conversion sub-model, and an unwrapped image corresponding to the interference fringe image to be processed is obtained.

[0070] The technical solution of the embodiment of the present invention is as follows: by inputting the interference fringe image to be processed into a feature extraction submodel for feature extraction, image features corresponding to the interference fringe image to be processed are obtained; and by inputting unwrapping operation prompt information into an unwrapping feature generation submodel for information encoding to generate an unwrapping operation prompt vector; further, the image features and the unwrapping operation prompt vector are input into a phase unwrapping feature generation submodel to generate phase unwrapping image features corresponding to the interference fringe image to be processed; the phase unwrapping image features are input into a feature conversion submodel for feature conversion to obtain an unwrapped image corresponding to the interference fringe image to be processed, so that in the process of phase unwrapping of the phase image, the abnormal phase jump generated is effectively processed and an unwrapped image with higher accuracy is generated. By adopting the unwrapping operation prompt information as the model input object, operational guidance is provided for the model to perform abnormal phase jump discrimination and processing, thereby ensuring the accuracy and continuity of the phase information in the wrapped image.

[0071] Embodiment 3

[0072] Figure 4 This is a flow chart of a phase unwrapping method provided in the third embodiment of the present invention. Based on the above embodiment, before applying the unwrapped image generation model, the pre-built deep learning model can be trained first, and then the unwrapped image can be generated based on the trained unwrapped image generation model. The specific implementation method can refer to the technical solution of this embodiment. Among them, the technical terms that are the same or similar to the above embodiments are not repeated here.

[0073] like Figure 4 As shown, the method includes:

[0074] S310, obtaining a plurality of training samples; wherein the training samples include a sample interference fringe image corresponding to a sample wafer surface, unwrapping operation prompt information, and a real unwrapping image corresponding to the sample interference fringe image.

[0075] It should be noted that before applying the unwrapped image generation model provided in this embodiment, the pre-constructed deep learning model can be trained in a supervised or unsupervised manner. Before training the deep learning model, multiple training samples can be constructed to train the model based on the multiple training samples. In order to improve the generation accuracy of the unwrapped image generation model, as many and rich training samples as possible can be constructed.

[0076] Among them, the sample interference fringe image can be an interference fringe image that has been denoised and is to be subjected to phase extraction and phase unwrapping processing. The sample interference fringe image can be an image obtained by photographing the interference fringes formed on the surface of the sample wafer through a camera device, or an image reconstructed by an image reconstruction model, or an image pre-stored in a storage space, etc. The unwrapping operation prompt information can be understood as information that explanatoryally describes the phase extraction and phase unwrapping operations. The unwrapping operation prompt information is at least used to describe the logical reasoning process of phase extraction and phase unwrapping, the intermediate reasoning results, and the judgment method of abnormal phase jumps. The real unwrapping image can be an image in which the phase information in the sample interference fringes is phase-unwrapped to restore the original, continuous phase information.

[0077] In this embodiment, in order to construct a large number of training samples in a rich manner, sample interference fringe images corresponding to the surfaces of multiple sample wafers may be obtained. Furthermore, each sample interference fringe image may be processed separately to obtain a real unwrapping image corresponding to each sample interference fringe image. Furthermore, unwrapping operation prompt information may be obtained, and training samples may be constructed based on the sample interference fringe image corresponding to the surface of the sample wafer, the real unwrapping image corresponding to the sample interference fringe image, and the unwrapping operation prompt information, thereby obtaining multiple training samples. Furthermore, a pre-constructed deep learning model may be trained based on multiple training samples to obtain a trained deep learning model, and the trained deep learning model may be used as an unwrapping image generation model.

[0078] S320: For multiple training samples, the sample interference fringe images and unwrapping operation prompt information in the training samples are input into a pre-built deep learning model, and the actual unwrapping image is output.

[0079] The deep learning model includes a feature extraction sub-model, an unwrapping feature generation sub-model, a phase unwrapping feature generation sub-model and a feature conversion sub-model. The actual unwrapped image is the unwrapped image output after the sample interference fringe image and the unwrapping operation prompt information are input into the deep learning model.

[0080] In this embodiment, for multiple training samples, the sample interference fringe image and the unwrapping operation prompt information in the training samples can be input into a pre-built deep learning model. Further, the sample interference fringe image can be input into a feature extraction sub-model to extract features of the sample interference fringe image based on the feature extraction sub-model to obtain sample image features. And, the unwrapping operation prompt information can be input into an unwrapping feature generation sub-model to encode the unwrapping operation prompt information based on the unwrapping generation sub-model to obtain an unwrapping operation prompt vector. Further, the sample image features and the unwrapping operation prompt vector can be input into a phase unwrapping feature generation sub-model to process the sample image features according to the unwrapping operation prompt vector based on the phase unwrapping feature generation sub-model, and output an actual unwrapping image corresponding to the sample interference fringe image.

[0081] S330 , performing loss processing on the actual unwrapped image and the true unwrapped image according to the target loss function to obtain a target loss value.

[0082] Among them, the target loss function is determined based on the error loss function and the cross entropy loss function.

[0083] Exemplarily, the joint loss function can be determined based on the following formula:

[0084] L uni =αL1+(1-α)L cct

[0085] Among them, L uni represents the target loss function; α represents the weight factor; L1 represents the error loss function; L cct represents the cross entropy loss function.

[0086] In this embodiment, in order to effectively train the unwrapped image generation model to generate a noise-free phase unwrapped image, loss processing can be performed on the actual unwrapped image and the true unwrapped image based on a target loss function obtained by weighted fusion of two loss functions.

[0087] Optionally, the actual unwrapped image and the real unwrapped image are subjected to loss processing according to a target loss function to obtain a target loss value, including: performing loss processing on the actual unwrapped image and the real unwrapped image according to an error loss function to obtain a first loss value; performing loss processing on the actual unwrapped image and the real unwrapped image according to a cross entropy loss function to obtain a second loss value; and performing weighted summation of the first loss value and the second loss value to obtain a target loss value.

[0088] The error loss function may be a function that characterizes the degree of difference between the image content of the actual unwrapped image and the true unwrapped image. The error loss function may be an L1 loss function. The cross entropy loss function may be a function that characterizes the degree of difference between the feature distribution of the actual unwrapped image and the true unwrapped image.

[0089] As an optional implementation of this embodiment, the actual unwrapped image and the real unwrapped image may be subjected to loss processing according to the error loss function to obtain a first loss value. Also, the actual unwrapped image and the real unwrapped image may be subjected to loss processing according to the cross entropy loss function to obtain a second loss value. Further, a first weight corresponding to the first loss value may be determined, a second weight corresponding to the second loss value may be determined, and the product between the first loss value and the first weight may be determined to obtain a first loss value to be superimposed, and the product between the second loss value and the second weight may be determined to obtain a second loss value to be superimposed. Further, the first loss value to be superimposed and the second loss value to be superimposed may be added to obtain a target loss value.

[0090] S340, based on the target loss value, the model parameters in the deep learning model are modified, and the convergence of the target loss function in the deep learning is used as the training goal to obtain the unwrapped image generation model.

[0091] Optionally, the unwrapping feature generation submodel includes an unwrapping feature generation module and a first feature alignment module; the phase unwrapping feature generation submodel includes a phase unwrapping feature generation module and a second feature alignment module; and the model parameters in the deep learning model are corrected based on the target loss value, including: correcting the model parameters of the feature extraction submodel, the first feature alignment module, the second feature alignment module and the feature conversion submodel based on the target loss value.

[0092] In this embodiment, during the model training process, the model parameters of the unwrapping feature generation module and the phase unwrapping feature generation module in the deep learning model can be frozen, and the model parameters of the feature extraction submodel, the first feature alignment module, the second feature alignment module, and the feature conversion submodel can be corrected based on the target loss value. After that, the training error of the target loss function in the deep learning model, that is, the loss parameter, can be used as a condition for detecting whether the current target loss function has reached convergence, for example, whether the training error is less than the preset error or whether the error change trend tends to be stable, or whether the current model iteration number is equal to the preset number of times. If the convergence condition is reached, for example, the training error of the target loss function is less than the preset error or the error change tends to be stable, it indicates that the deep learning model training is completed, and the iterative training can be stopped at this time. If it is detected that the convergence condition is not reached at present, other training samples can be further obtained to train the deep learning model until the training error of the target loss function is within the preset range. When the training error of the loss function reaches convergence, the trained deep learning model can be used as the unwrapping image generation model.

[0093] S350, obtaining the interference fringe image to be processed corresponding to the wafer surface and unpacking operation prompt information.

[0094] S360, processing the interference fringe image to be processed and the unwrapping operation prompt information according to the pre-trained unwrapping image generation model, and obtaining an unwrapping image corresponding to the interference fringe image to be processed, so as to determine the surface shape information corresponding to the wafer surface based on the unwrapping image.

[0095] The technical solution of the embodiment of the present invention is to obtain multiple training samples; wherein the training samples include sample interference fringe images, unwrapping operation description information and real unwrapped images corresponding to the sample interference fringe images; further, for multiple training samples, the sample interference fringe images and unwrapping operation description information in the training samples are input into a pre-constructed deep learning model, and the actual unwrapped images are output; further, the actual unwrapped images and the real unwrapped images are subjected to loss processing according to the target loss function to obtain the target loss value; further, based on the target loss value, the model parameters in the deep learning model are corrected, and the convergence of the target loss function in the deep learning is used as the training target to obtain the unwrapped image generation model, thereby achieving the effect of training an unwrapped image generation model based on the constructed training samples, which can perform phase unwrapping processing on the phase image and effectively process the generated abnormal phase jump to generate the unwrapped image, thereby effectively improving the intelligence, accuracy and robustness of the model.

[0096] Embodiment 4

[0097] Figure 5Schematic diagram of the structure of a phase unwrapping device provided by the fourth embodiment of the present invention. Figure 5 As shown, the device includes: an image acquisition module 410 and a phase unwrapping module 420.

[0098] Among them, the image acquisition module 410 is used to obtain the interference fringe image to be processed and the unwrapping operation prompt information corresponding to the wafer surface; wherein the unwrapping operation prompt information is at least used to describe the logical reasoning process of phase extraction and phase unwrapping, the intermediate reasoning results and the judgment method of abnormal phase jump; the phase unwrapping module 420 is used to process the interference fringe image to be processed and the unwrapping operation prompt information according to the unwrapping image generation model obtained by pre-training, and obtain the unwrapping image corresponding to the interference fringe image to be processed, so as to determine the surface shape information corresponding to the wafer surface based on the unwrapping image; wherein the unwrapping The image generation model includes at least a feature extraction sub-model, an unwrapping feature generation sub-model, a phase unwrapping feature generation sub-model and a feature conversion sub-model; the feature extraction sub-model is used to extract the image features of the interference fringe image to be processed; the unwrapping feature generation sub-model is used to generate an unwrapping operation prompt vector according to the unwrapping operation prompt information; the phase unwrapping feature generation sub-model is used to generate a phase unwrapping image feature corresponding to the interference fringe image to be processed based on the image features and the unwrapping operation prompt vector; the feature conversion sub-model is used to convert the phase unwrapping image features into an unwrapping image corresponding to the interference fringe image to be processed.

[0099] The technical solution of the embodiment of the present invention provides a data basis and operation guidance for subsequent phase unwrapping by obtaining the interference fringe image to be processed and the unwrapping operation prompt information corresponding to the wafer surface, and the unwrapping operation prompt information is at least used to prompt the logical reasoning process of phase extraction and phase unwrapping, the intermediate reasoning results and the judgment method of abnormal phase jumps, and provides operation guidance for judging and processing abnormal phase jumps in the phase unwrapping process; further, the interference fringe image to be processed and the unwrapping operation prompt information are processed according to the pre-trained unwrapping image generation model, and the unwrapping image corresponding to the interference fringe image to be processed is obtained, so as to determine the surface information corresponding to the wafer surface based on the unwrapping image, solves the problem in the related art that the abnormal phase jump cannot be effectively processed during the phase extraction and phase unwrapping process, thereby causing discontinuous phase unwrapping, and realizes the effect of effectively processing the abnormal phase jump generated in the process of phase unwrapping of the phase image, and generating a more accurate unwrapping image, and by using the unwrapping operation prompt information as the model input object, it provides operation guidance for the model to perform abnormal phase jump judgment and processing, ensuring the accuracy and continuity of the phase information in the wrapping image.

[0100] Optionally, the device further includes: an original image acquisition module and an image denoising module.

[0101] An original image acquisition module is used to acquire an original interference fringe image corresponding to the wafer surface;

[0102] The image denoising module is used to denoise the original interference fringe image according to a preset image denoising method to obtain a to-be-processed interference fringe image corresponding to the wafer surface.

[0103] Optionally, the device further includes: a process description information acquisition module, a feature extraction module and an operation description information determination module.

[0104] A process description information acquisition module, used to acquire unpacking process description information, wherein the unpacking process description information includes step description information of multiple unpacking operation steps, an intermediate result image associated with each of the unpacking operation steps, and a phase jump discrimination image of an abnormal phase jump;

[0105] a feature extraction module, configured to perform feature extraction on an intermediate result image associated with at least one of the unwrapping operation steps to obtain an intermediate result feature associated with at least one of the unwrapping operation steps, and to perform feature extraction on the phase jump discrimination image to obtain an abnormal phase jump discrimination feature corresponding to the abnormal phase jump;

[0106] An operation prompt information determination module is used to construct unpacking operation prompt information based on step description information of multiple unpacking operation steps, intermediate result features associated with at least one unpacking operation step, and the abnormal phase jump discrimination feature.

[0107] Optionally, the phase unwrapping module 420 includes: a feature extraction unit, a feature generation unit and a feature conversion unit.

[0108] a feature extraction unit, configured to input the interference fringe image to be processed into the feature extraction sub-model for feature extraction to obtain image features corresponding to the interference fringe image to be processed, and to input the unwrapping operation prompt information into the unwrapping feature generation sub-model for information encoding to generate an unwrapping operation prompt vector;

[0109] A feature generation unit, used for inputting the image feature and the unwrapping operation prompt vector into the phase unwrapping feature generation sub-model to generate a phase unwrapping image feature corresponding to the interference fringe image to be processed;

[0110] The feature conversion unit is used to input the phase unwrapping image features into the feature conversion sub-model for feature conversion to obtain an unwrapped image corresponding to the interference fringe image to be processed.

[0111] Optionally, the unwrapping feature generation submodel includes an unwrapping feature generation module and a first feature alignment module; the feature extraction unit includes: an information encoding subunit and an information alignment subunit;

[0112] An information encoding subunit, configured to encode the unpacking operation prompt information based on the unpacking feature generation module to obtain a prompt vector to be processed;

[0113] An information alignment subunit is used to align the prompt vector to be processed based on the first feature alignment module to obtain an unpacking operation prompt vector

[0114] Optionally, the phase unwrapping generation sub-model includes a phase unwrapping feature generation module and a second feature alignment module; the feature generation unit includes: a feature generation sub-unit and a feature alignment sub-unit.

[0115] A feature generation subunit, configured to process the image feature and the unwrapping operation prompt vector based on the phase unwrapping feature generation module to generate an unwrapped image feature to be processed;

[0116] The feature alignment subunit is used to perform feature alignment on the features of the unwrapped image to be processed based on the second feature alignment module to obtain phase unwrapped image features corresponding to the interference fringe image to be processed.

[0117] Optionally, the device further includes: a sample acquisition module, an actual unwrapped image determination module, a loss determination module and a model parameter correction module.

[0118] A sample acquisition module, used to acquire a plurality of training samples; wherein the training samples include a sample interference fringe image of a sample wafer surface, unwrapping operation description information, and a real unwrapping image corresponding to the sample interference fringe image;

[0119] An actual unwrapped image determination module is used for inputting the sample interference fringe images in the training samples and the unwrapping operation prompt information into a pre-built deep learning model for a plurality of training samples, and outputting an actual unwrapped image;

[0120] A loss determination module, used for performing loss processing on the actual unwrapped image and the true unwrapped image according to a target loss function to obtain a target loss value; wherein the target loss function is determined based on an error loss function and a cross entropy loss function;

[0121] A model parameter correction module is used to correct the model parameters in the deep learning model based on the target loss value, and take the convergence of the target loss function in the deep learning as the training goal to obtain an unwrapped image generation model.

[0122] Optionally, the deep learning model includes a feature extraction sub-model, an unwrapping feature generation sub-model, a phase unwrapping feature generation sub-model and a feature conversion sub-model; the unwrapping feature generation sub-model includes an unwrapping feature generation module and a first feature alignment module; the phase unwrapping feature generation sub-model includes a phase unwrapping feature generation module and a second feature alignment module;

[0123] The model parameter correction module is specifically used to correct the model parameters of the feature extraction sub-model, the first feature alignment module, the second feature alignment module and the feature conversion sub-model based on the target loss value.

[0124] The phase unwrapping device provided in the embodiment of the present invention can execute the phase unwrapping method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0125] Embodiment 5

[0126] Figure 6 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0127] like Figure 6As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0128] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0129] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 performs the various methods and processes described above, such as the phase unwrapping method.

[0130] In some embodiments, the phase unwrapping method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the phase unwrapping method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the phase unwrapping method in any other suitable manner (e.g., by means of firmware).

[0131] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0132] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0133] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0134] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0135] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a target blockchain network, and the Internet.

[0136] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0137] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0138] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A phase unwrapping method, characterized in that: include: Obtaining the interference fringe image to be processed corresponding to the wafer surface and unwrapping operation prompt information; wherein the unwrapping operation prompt information is used to prompt the logical reasoning process of phase extraction and phase unwrapping, the intermediate reasoning results, and the judgment method of abnormal phase jump; Processing the interference fringe image to be processed and the unwrapping operation prompt information according to a pre-trained unwrapping image generation model, and obtaining an unwrapping image corresponding to the interference fringe image to be processed, so as to determine the surface shape information corresponding to the wafer surface based on the unwrapping image; Among them, the unwrapping image generation model at least includes a feature extraction sub-model, an unwrapping feature generation sub-model, a phase unwrapping feature generation sub-model and a feature conversion sub-model; the feature extraction sub-model is used to extract the image features of the interference fringe image to be processed; the unwrapping feature generation sub-model is used to generate an unwrapping operation prompt vector according to the unwrapping operation prompt information; the phase unwrapping feature generation sub-model is used to generate a phase unwrapping image feature corresponding to the interference fringe image to be processed based on the image features and the unwrapping operation prompt vector; the feature conversion sub-model is used to convert the phase unwrapping image features into an unwrapping image corresponding to the interference fringe image to be processed.

2. The phase unwrapping method according to claim 1, characterized in that: Also includes: Acquire an original interference fringe image corresponding to the wafer surface; The original interference fringe image is denoised according to a preset image denoising method to obtain a to-be-processed interference fringe image corresponding to the wafer surface.

3. The phase unwrapping method according to claim 1, characterized in that: Also includes: Acquire unpacking process description information, wherein the unpacking process description information includes step description information of a plurality of unpacking operation steps, an intermediate result image associated with each of the unpacking operation steps, and a phase jump discrimination image of an abnormal phase jump; Performing feature extraction on at least one of the intermediate result images associated with the unwrapping operation step to obtain an intermediate result feature associated with at least one of the unwrapping operation steps, and performing feature extraction on the phase jump discrimination image to obtain an abnormal phase jump discrimination feature corresponding to the abnormal phase jump; The unpacking operation prompt information is constructed based on the step description information of the plurality of unpacking operation steps, the intermediate result feature associated with at least one of the unpacking operation steps, and the abnormal phase jump discrimination feature.

4. The phase unwrapping method according to claim 1, characterized in that: The unwrapping image generation model obtained by pre-training processes the interference fringe image to be processed and the unwrapping operation prompt information, and obtains an unwrapping image corresponding to the interference fringe image to be processed, including: Inputting the interference fringe image to be processed into the feature extraction sub-model to perform feature extraction, thereby obtaining image features corresponding to the interference fringe image to be processed; and inputting the unwrapping operation prompt information into the unwrapping feature generation sub-model to perform information encoding, thereby generating an unwrapping operation prompt vector; Inputting the image feature and the unwrapping operation prompt vector into the phase unwrapping feature generation submodel to generate a phase unwrapping image feature corresponding to the interference fringe image to be processed; The phase unwrapping image features are input into the feature conversion submodel for feature conversion to obtain an unwrapped image corresponding to the interference fringe image to be processed.

5. The phase unwrapping method according to claim 4, characterized in that: The unwrapping feature generation sub-model includes an unwrapping feature generation module and a first feature alignment module; the step of inputting the unwrapping operation prompt information into the unwrapping feature generation sub-model for information encoding to generate an unwrapping operation prompt vector includes: Encoding the unpacking operation prompt information based on the unpacking feature generation module to obtain a prompt vector to be processed; The prompt vectors to be processed are aligned based on the first feature alignment module to obtain an unpacking operation prompt vector.

6. The phase unwrapping method according to claim 4, characterized in that: The phase unwrapping generation sub-model includes a phase unwrapping feature generation module and a second feature alignment module; The step of inputting the image feature and the unwrapping operation hint vector into the phase unwrapping feature generation submodel to generate a phase unwrapping image feature corresponding to the interference fringe image to be processed includes: Processing the image feature and the unwrapping operation prompt vector based on the phase unwrapping feature generation module to generate an unwrapped image feature to be processed; Based on the second feature alignment module, feature alignment is performed on the features of the unwrapped image to be processed to obtain phase unwrapped image features corresponding to the interference fringe image to be processed.

7. The phase unwrapping method according to claim 1, characterized in that: Also includes: Acquire a plurality of training samples; wherein the training samples include a sample interference fringe image of a sample wafer surface, unwrapping operation description information, and a real unwrapping image corresponding to the sample interference fringe image; For a plurality of training samples, the sample interference fringe images in the training samples and the unwrapping operation description information are input into a pre-built deep learning model, and an actual unwrapping image is output; Performing loss processing on the actual unwrapped image and the true unwrapped image according to a target loss function to obtain a target loss value; wherein the target loss function is determined based on an error loss function and a cross entropy loss function; The model parameters in the deep learning model are modified based on the target loss value, and the convergence of the target loss function in the deep learning is used as a training goal to obtain an unwrapped image generation model.

8. The phase unwrapping method according to claim 7, characterized in that: The deep learning model includes a feature extraction sub-model, an unwrapping feature generation sub-model, a phase unwrapping feature generation sub-model and a feature conversion sub-model; the unwrapping feature generation sub-model includes an unwrapping feature generation module and a first feature alignment module; the phase unwrapping feature generation sub-model includes a phase unwrapping feature generation module and a second feature alignment module; The modifying of the model parameters in the deep learning model based on the target loss value includes: Model parameters of the feature extraction sub-model, the first feature alignment module, the second feature alignment module and the feature conversion sub-model are modified based on the target loss value.

9. A phase unwrapping device, characterized in that: include: An image acquisition module, used to acquire the interference fringe image to be processed corresponding to the wafer surface and unwrapping operation prompt information; wherein the unwrapping operation prompt information is used to prompt the logical reasoning process of phase extraction and phase unwrapping, the intermediate reasoning results and the judgment method of abnormal phase jump; A phase unwrapping module is used to process the interference fringe image to be processed and the unwrapping operation prompt information according to a pre-trained unwrapping image generation model, and obtain an unwrapping image corresponding to the interference fringe image to be processed, so as to determine the surface information corresponding to the wafer surface based on the unwrapping image; wherein the unwrapping image generation model at least includes a feature extraction submodel, an unwrapping feature generation submodel, a phase unwrapping feature generation submodel and a feature conversion submodel; the feature extraction submodel is used to extract the image features of the interference fringe image to be processed; the unwrapping feature generation submodel is used to generate an unwrapping operation prompt vector according to the unwrapping operation prompt information; the phase unwrapping feature generation submodel is used to generate a phase unwrapping image feature corresponding to the interference fringe image to be processed based on the image features and the unwrapping operation prompt vector; the feature conversion submodel is used to convert the phase unwrapping image feature into an unwrapping image corresponding to the interference fringe image to be processed.

10. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so as to enable the at least one processor to perform the phase unwrapping method according to any one of claims 1 to 8.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the phase unwrapping method according to any one of claims 1 to 8 when executed.

12. A computer program product, characterized in that The computer program product comprises a computer program which, when executed by a processor, implements the phase unwrapping method according to claims 1-8.

Images