A calibration method and device for the incident angle offset of the central wavelength of spectral imaging

Abstract

The present invention mainly provides a calibration method and device for the incident angle offset of the central wavelength of spectral imaging, belonging to the technical field of spectral imaging image calibration. The calibration method of the present invention is based on the established calibration model. First, spectral image data of uniformly distributed specific samples and a large reference whiteboard are collected, and the spectral reflectance and reference reflectance at different incident angles of the samples are calculated through normalization, forming a data set and establishing a corresponding mapping relationship to form a training sample set, which is input into the calibration model for training to obtain the final calibration model. After the model is built, the samples to be detected are collected and a new spectral data set is calculated and input into the calibration model to calibrate the spectral data. The present invention improves the recognition accuracy of the Fabry-Perot interferometric cavity spectral camera, eliminates the influence of the central wavelength offset, has a clever method design and is easy to implement; at the same time, there is no need to re-establish the spectral application model, and the implementation cost is lower and the efficiency is higher.

Description

Technical Field

[0001] The present invention belongs to the technical field of spectral imaging image calibration, and particularly relates to a calibration method and device for the offset of the central wavelength incident angle in spectral imaging. Background Art

[0002] The offset of the central wavelength is an inherent physical property of the Fabry-Perot interferometric cavity (Fabry-Perot cavity). As long as the incident light cannot enter the Fabry-Perot interferometric cavity collinearly, any multi-spectral or hyperspectral imaging system based on the Fabry-Perot interferometric cavity will have the problem of central wavelength offset, and it is independent of the technical implementation method of the Fabry-Perot interferometric cavity (electrostatic / piezoelectric / other methods). Currently, existing multi-spectral or hyperspectral imaging systems based on the Fabry-Perot interferometric cavity cannot make the incident light enter the Fabry-Perot interferometric cavity completely collinearly, so there will inevitably be a problem of central wavelength offset.

[0003] In existing technical solutions, more efforts are made to design and modify the optical system of spectral imaging to make the light enter the Fabry-Perot interferometric cavity as collinearly as possible. However, in currently known optical systems for spectral imaging based on the Fabry-Perot, it is impossible to make 100% of all light enter the Fabry-Perot interferometric cavity collinearly, and this will lead to problems such as a smaller FOV of the spectral camera and a lower light transmittance. At the same time, the technical difficulty and implementation cost of modifying the optical system are also higher.

[0004] There are also technical solutions that attempt to add spectral spatial features to the spectral image to eliminate the influence of the central wavelength offset, so that the later spectral application model can view the spectral reflectance data at the correct wavelength. Although it can solve the influence of the central wavelength offset on the spectral application model, it is necessary to re-establish the spectral application model, adding new problems. Summary of the Invention

[0005] In view of the above problems, the present invention patent proposes a method for building a calibration model for the offset of the central wavelength incident angle in spectral imaging, including the following steps:

[0006] Step 1, using a spectral camera of a Fabry-Perot interferometric cavity to capture and collect spectral image data of a specific sample uniformly distributed; the uniformly distributed specific sample means that within the spectral image range, each incident angle θ is distributed with a specific sample substance; the spectral image imaging data is a three-dimensional spectral cube, including two-dimensional image information of m rows * n columns and one-dimensional spectral information of k spectral bands;

[0007] Use a spectral camera with a Fabry - Perot interferometric cavity to capture spectral image data of a large standard diffuse reflection reference whiteboard. The spectral image data of the large standard diffuse reflection reference whiteboard is also a spectral cube in three - dimensional form, containing two - dimensional image information and one - dimensional spectral information in k spectral bands;

[0008] Step 2: Use the spectral cube of the obtained large standard diffuse reflection reference whiteboard to normalize the spectral cube of a specific sample, and obtain the reflectance of the specific sample substance at each incident angle θ, as well as the reference spectral reflectance of the specific sample substance when vertically incident at 0 degrees at the image center position;

[0009] Step 3: Establish a data set collection X for each incident angle θ and its corresponding reflectance of the specific sample substance θ , and at the same time, establish a set Y of reference spectral reflectances in k spectral bands;

[0010] X θ =[Reflection_1, Reflection_2,..., Reflection_k, incident angle θ]

[0011] Reflection_k represents the reflectance at the angle θ in the k - th spectral band;

[0012] Y = [Reflection_0_1, Reflection_0_2,..., Reflection_0_k] Reflection_0_k represents the reference spectral reflectance when the incident angle is 0 in the k - th spectral band;

[0013] Step 4: Establish a mapping relationship between X θ and Y, and form a training sample data set;

[0014] Step 5: Initialize the parameters of the calibration model, input the training sample data set, and after multiple iterations, stop the iteration when the prediction error is reduced to the set threshold to obtain the final calibration model.

[0015] Preferably, when using the spectral cube of the obtained large standard diffuse reflection reference whiteboard to normalize the spectral cube of a specific sample, the normalization formula is as follows:

[0016]

[0017] Cube_norm represents a three-dimensional matrix of size m*n*k, representing the normalized spectral cube of the target object, from which the reflectance value Reflection of each pixel in the two-dimensional image of the spectral cube in k spectral bands can be obtained; Cube_target represents a three-dimensional matrix of size m*n*k, representing the spectral cube of a specific sample before normalization; Cube_white represents a three-dimensional matrix of size m*n*k, representing the spectral cube of a large standard diffuse reflection reference white board.

[0018] Preferably, the influence of the black reference is also considered. In step 1, the spectral cube of the large black reference taken is also obtained, that is, the camera lens cap is closed to ensure that there is no ambient light around, and then the shooting is carried out; the normalization formula is further deformed as:

[0019]

[0020] Cube_dark represents a three-dimensional matrix of size m*n*k, representing the spectral cube of the large black reference. Preferably, the calculation formula for the incident angle θ is:

[0021] θ = arctan(w / h)

[0022] The incident angle θ is calculated based on the horizontal distance w deviating from the optical center axis of the camera lens and the vertical distance h between the specific sample and the lens center.

[0023] Preferably, the calibration model adopts any one of a neural network model, a support vector machine model, a logistic regression model or a Bayesian classification model.

[0024] The second aspect of the present invention provides a calibration method for the offset of the central wavelength incident angle of spectral imaging, including the following processes:

[0025] Obtain the spectral image data of the specific sample to be detected taken by the spectral camera passing through the Fabry-Perot interferometer cavity, that is, the spectral cube;

[0026] Use the spectral cube of the large standard diffuse reflection reference white board obtained to normalize the spectral cube of the specific sample to be detected, obtain its reflectance in k spectral bands and the corresponding incident angle θ, and establish a data set;

[0027] Input the established data set into the calibration model built by the model building method as described in the first aspect for calibration;

[0028] Output the calibrated reflectance data set of the correct specific sample to be detected for spectral curve plotting.

[0029] In a third aspect of the present invention, there is also provided a calibration device for the incident angle offset of the central wavelength of spectral imaging. The device includes at least one processor and at least one memory; a computer execution program is stored in the memory; when the processor executes the execution program stored in the memory, the processor can execute the calibration method for the incident angle offset of the central wavelength of spectral imaging as described in the second aspect.

[0030] In a fourth aspect of the present invention, there is provided a computer-readable storage medium in which a computer execution program is stored. When the computer execution program is executed by a processor, it is used to implement the calibration method for the incident angle offset of the central wavelength of spectral imaging as described in the second aspect.

[0031] Advantages of the present invention: In existing technical solutions, more efforts are made to design and modify the optical system of spectral imaging to make the light enter the Fabry-Perot interferometer cavity as collimated as possible. However, this will lead to problems such as a smaller FOV of the spectral camera and a lower light transmittance. At the same time, the technical difficulty and implementation cost of modifying the optical system are also higher. The calibration method of the present invention does not require changing the existing spectral imaging system. By obtaining uniformly distributed specific sample spectral image data and normalizing it with a large standard diffuse reflection reference whiteboard, the actual reflectance at each angle corresponding to each two-dimensional spatial coordinate in each spectral segment of the spectral image and the reference reflectance of each spectral segment are obtained. A mapping relationship is established to obtain a dataset training sample. By training, a machine learning model for mapping the spectral reflectance at different incident angles and 0-degree incident angle is built, which can calibrate the reflectance of the corresponding spectral segment of the sample to be detected in the re-shot image, so as to accurately draw a reflectance spectral curve, improve the recognition accuracy, eliminate the influence of the central wavelength offset, with a clever method design and easy implementation; at the same time, there is no need to re-establish a spectral application model, with a lower implementation cost, higher efficiency, and stronger universality and generality. Description of the Drawings

[0032] Figure 1 It is a schematic diagram of the principle of the Fabry-Perot interferometer cavity.

[0033] Figure 2 It is a schematic diagram of the central wavelength offset actually observed in the experiment.

[0034] Figure 3 It is a schematic diagram of the phenomenon of the spectral curve offset caused by the central wavelength offset.

[0035] Figure 4 It is a flow chart of the calibration method of the present invention.

[0036] Figure 5 It is a schematic diagram of a spectral cube.

[0037] Figure 6It is an image of the leather sample in the 920nm spectral band.

[0038] Figure 7 It is an image of the large standard diffuse reflection reference white board in the 920nm spectral band.

[0039] Figure 8 It is the normalized reflectance image of the leather sample.

[0040] Figure 9 It is a schematic diagram of the incident angle of the camera lens offset.

[0041] Figure 10 It is the reflectance curve at different incident angles in the 920nm spectral band.

[0042] Figure 11 It is the reflectance curves of 10 spectral bands at 0 and 11 angles.

[0043] Figure 12 It is a schematic diagram of the structure of the neural network multi-layer perceptron classification model.

[0044] Figure 13 It is a schematic diagram of the reflectance curve during the verification of the test sample.

[0045] Figure 14 It is a comparison chart of the classification of genuine and fake leather before the calibration of the reflectance curve.

[0046] Figure 15 It is a comparison chart of the classification of genuine and fake leather after the calibration of the reflectance curve.

[0047] Figure 16 It is a simple schematic diagram of the structure of the calibration device in Embodiment 2 of the present invention. Detailed implementation mode

[0048] The invention will be further described below in conjunction with specific embodiments.

[0049] Embodiment 1:

[0050] The transmittance formula of the Fabry-Perot interferometer cavity is:

[0051]

[0052]

[0053] Where R represents the reflectance.

[0054] Such as Figure 1As shown, n represents the refractive index of the material; θ represents the refraction angle; T represents the transmittance, T1 represents the first transmittance, T2 represents the second transmittance; R represents the reflectivity, R0 represents the reflectivity of the first reflection, R1 represents the reflectivity of the second reflection, and R2 represents the reflectivity of the third reflection; l represents the thickness of the Fabry-Perot cavity. Figure 1 From the principle diagram of the Fabry-Perot interferometer cavity and the derived transmittance formula of the Fabry-Perot interferometer cavity, it can be seen that within the first-order range, the central wavelength λ with maximum transmittance through the Fabry-Perot interferometer cavity is inversely proportional to the incident angle θ of the light. When θ changes in the range of [0,90°], as θ increases, cosθ decreases, and the central wavelength λ with maximum transmittance also decreases.

[0055] Figure 2 This is the center wavelength shift phenomenon observed in the experiment. The horizontal axis represents the image width and the vertical axis represents the image height. The unit is nanometer. It can be seen that from the middle of the image to the periphery of the image, the center wavelength shifts from 720nm to 690nm. The center wavelength of the image center is larger than the center wavelength of the periphery of the image. This phenomenon is consistent with the conclusion derived from the above formula. In other words, the larger the θ is, the more the corresponding image area is biased toward the periphery of the image, and the center wavelength of the area that obtains the maximum reflectivity also decreases.

[0056] Figure 3 It is a phenomenon observed in the experiment that the spectral curve is shifted due to the central wavelength shift. The horizontal axis of the spectral curve represents the wavelength, and the vertical axis represents the reflectivity. The same target is placed in different spatial positions and spectral images are taken. By comparing the multi-spectral curve and the hyperspectral curve, it can be seen that due to the influence of the central wavelength shift, the band range of the spectral characteristic curve of point A is from 725nm to 920nm; the band range of the spectral characteristic curve of point B is from 705nm to 885nm.

[0057] In response to the above problems, the patent of the present invention proposes a brand-new method, which obtains uniformly distributed spectral image data of specific samples and normalizes it through a large standard diffuse reflectance reference whiteboard, obtains the actual reflectivity of each angle corresponding to each spectral segment in the two-dimensional spatial coordinates of each spectral segment of the spectral image and the benchmark reflectivity of each spectral segment, establishes a mapping relationship to obtain data set training samples, and builds a spectral reflectivity mapping machine learning model for different incident angles and 0 degree incident angle through training. The reflectivity of the corresponding spectral segment of the sample to be tested that is photographed again can be calibrated, so that an accurate reflectivity spectral curve can be drawn, eliminating the influence of the central wavelength offset.

[0058] The most important thing about the calibration method of the present invention is how to build such a calibration model. After the calibration model is built, calibration can be performed based on the calibration model. Figure 4The model building method and calibration method of the present invention are described in detail.

[0059] The model building method in the present invention mainly includes the following steps:

[0060] Step 1, using a spectral camera with a Fabry - Perot interferometer cavity to capture and collect spectral image data of uniformly distributed specific samples; the uniformly distributed specific samples are such that for each incident angle θ within the spectral image range, there are specific sample substances distributed; the spectral image imaging data is a three - dimensional spectral cube, containing two - dimensional image information of m rows * n columns and one - dimensional spectral information of k spectral bands. The structure of the spectral cube is as Figure 5 shown, with two - dimensional image information (spatial information) of m rows * n columns (for example: 1024 rows x 1280 columns) and one - dimensional spectral information of k spectral bands (for example: k = 10 spectral bands).

[0061] Using a spectral camera with a Fabry - Perot interferometer cavity to obtain spectral image data of a large standard diffuse reflection reference whiteboard, the spectral image data of the large standard diffuse reflection reference whiteboard is also a three - dimensional spectral cube, containing two - dimensional image information and one - dimensional spectral information of k spectral bands;

[0062] Step 2, using the spectral cube of the obtained large standard diffuse reflection reference whiteboard to normalize the spectral cube of the specific samples, obtaining the reflectivity of the specific sample substances at each incident angle θ, and the reference spectral reflectivity of the specific sample substances when vertically incident at 0 degrees at the image center position;

[0063] Step 3, establishing a data set collection X of each incident angle θ and its corresponding reflectivity of the specific sample substances θ , and at the same time, establishing a set Y of reference spectral reflectivities of k spectral bands;

[0064] X θ = [Reflection_1, Reflection_2,..., Reflection_k, incident angle θ]

[0065] Reflection_k represents the reflectivity at the angle θ in the k - th spectral band;

[0066] Y = [Reflection_0_1, Reflection_0_2,..., Reflection_0_k] Reflection_0_k represents the reference spectral reflectivity when the incident angle is 0 in the k - th spectral band;

[0067] Step 4, establishing a mapping relationship between X θ and Y, and forming a training sample data set;

[0068] Step 5: Initialize the parameters of the calibration model, input the training sample dataset, and after multiple iterations, stop the iteration when the prediction error is reduced to the set threshold to obtain the final calibration model.

[0069] In this embodiment, the above method will be elaborated by taking leather samples as an example.

[0070] 1. Collect spectral image data of uniformly distributed leather samples and spectral image data of a large standard diffuse reflection reference whiteboard.

[0071] Use a 10-channel (i.e., 10 spectral bands, channel range: 713nm - 920nm) multispectral camera to observe uniform genuine and fake leather samples. As Figure 6 shown, it is the original spectral image of the 10th channel (920nm) of a leather sample, where the leather image covers most of the image range on the right (also applicable to samples covering the entire image range, here only for illustration), ensuring that leather samples are distributed at each incident angle θ.

[0072] Similarly, use a 10-channel multispectral camera to observe a uniform large standard diffuse reflection reference whiteboard. As Figure 7 shown, it is the original spectral image of the 10th channel (920nm) of the large standard diffuse reflection reference whiteboard, which covers the entire image range.

[0073] 2. Perform normalization processing to obtain the reflectance and reference spectral reflectance of the leather sample at each incident angle θ.

[0074] Use the spectral cube of the obtained large standard diffuse reflection reference whiteboard to perform normalization processing on the spectral cube of the leather sample. The normalization formula is as follows:

[0075]

[0076] Cube_norm represents a three-dimensional matrix of size m*n*k, representing the spectral cube of the target object after normalization. The reflectance value Reflection of each pixel point in the two-dimensional image of the spectral cube in k spectral bands can be obtained; Cube_target represents a three-dimensional matrix of size m*n*k, representing the spectral cube of a specific sample before normalization; Cube_white represents a three-dimensional matrix of size m*n*k, representing the spectral cube of the large standard diffuse reflection reference whiteboard. The reference spectral reflectance is obtained according to the above formula at the position of 0-degree vertical incidence of the leather sample at the center of the image.

[0077] If the influence of the black reference is considered, the spectral cube of the large black reference taken needs to be obtained, that is, close the camera lens cap to ensure that there is no ambient light around, and then take a picture; the normalization formula is further deformed as:

[0078]

[0079] Cube_dark represents a three-dimensional matrix of size m*n*k, representing the large black reference spectral cube.

[0080] As Figure 8 shown is the normalized reflectance image of the leather sample in the 10th channel (920 nm).

[0081] The calculation formula for the incident angle θ is:

[0082] θ = arctan(w / h)

[0083] The incident angle θ is calculated based on the horizontal distance w deviating from the optical center axis of the camera lens and the vertical distance h between a specific sample and the lens center, as Figure 9 shown.

[0084] Comparing the reflectances at different incident angles of 920 nm, as Figure 10 shown, the reflectance at an incident angle of 0 degrees is used as the reference spectral reflectance without angular offset. As the incident angle increases, the reflectance shows an overall decreasing trend.

[0085] Comparing the reflectance curves of 10 channels at different incident angles, as Figure 11 shown, the solid line represents the reference spectral curve at an incident angle of 0 degrees, and the dashed line represents the spectral curve at an incident angle of 11 degrees, with a certain error compared to the reference spectral curve.

[0086] 3. Establish a data set set X for each incident angle θ and the reflectance of its corresponding leather sample θ , and at the same time, establish a set Y of reference spectral reflectances for k spectral bands, and establish a mapping relationship.

[0087] X θ = [Reflection_1, Reflection_2,..., Reflection_k, incident angle θ]

[0088] Reflection_k represents the reflectance at the angle θ in the kth spectral band;

[0089] Y = [Reflection_0_1, Reflection_0_2,..., Reflection_0_k] Reflection_0_k represents the reference spectral reflectance at an incident angle of 0 in the kth spectral band.

[0090] At the same time, establish a mapping relationship between X θ and Y, and form a training sample data set.

[0091] Y = NeuralNet(X θ )

[0092] 4. Initialize and train the calibration machine model to obtain the final calibration model.

[0093] Machine learning is an artificial intelligence technology. Its basic idea is to let the computer automatically find patterns and make decisions by providing the computer with a large amount of data and guiding information. In machine learning, the algorithm observes the data and learns the relationships between the data to predict the output. Machine learning can be divided into three types: supervised learning, unsupervised learning, and semi-supervised learning. Supervised learning means that the algorithm learns through known input / output pairs during the training process, while unsupervised learning learns without known outputs. Semi-supervised learning means learning is carried out when part of the training data has labels and part has no labels. The calibration model in this embodiment adopts a neural network multi-layer perceptron classification model, and the structure diagram is as Figure 12 shown. It can also adopt any one of the support vector machine model, logistic regression model, or Bayesian classification model.

[0094] Initialize the neural network model parameters, input the training samples, and after multiple iterations, stop the iteration when the prediction error is reduced to the set threshold to obtain the final neural network model. Use the test samples for verification. As Figure 13 shown, the solid circle solid line represents the reference reflectance curve at the 0-degree incident angle, the triangle dashed line represents the reflectance curve at the 11-degree incident angle before calibration, and the square solid line represents the reflectance curve after neural network calibration. It can be seen that it is closer to the reference reflectance curve.

[0095] The calibration method for the spectral imaging center wavelength incident angle offset of the present invention is based on the calibration model built through the above process. The process and effect verification of applying the calibration model for calibration are as follows:

[0096] Obtain the spectral image data of the leather sample to be detected captured by the spectral camera passing through the Fabry-Perot interferometer cavity, that is, the spectral cube;

[0097] Use the spectral cube of the large standard diffuse reflection reference whiteboard obtained to normalize the spectral cube of the leather sample to be detected, obtain its reflectance at k spectral bands and the corresponding incident angle θ, and establish a data set;

[0098] Input the established data set into the calibration model built by the model building method for calibration;

[0099] Output the calibrated reflectance data set of the correct specific sample to be detected for spectral curve plotting.

[0100] As Figure 14 andFigure 15 As shown in the figure, new genuine and fake leather sample blocks are cut, and the classification effects before and after reflectance calibration are compared. The left figure shows the original spectral data of the leather sample in the 828nm channel, and the right figure shows the predicted label image of the three-classification (genuine leather, fake leather, background) for each pixel. The prediction accuracy rate is marked in the center of the image based on the proportion of the predicted label results of each leather image block. The average accuracy rate before reflectance calibration is 92%, and the average accuracy rate after reflectance calibration is increased to 97%, verifying the effectiveness of the calibration method of the present invention.

[0101] Example 2:

[0102] As Figure 16 shown, the present invention also provides a calibration device for the incident angle offset of the spectral imaging center wavelength. The device includes at least one processor, at least one memory, and an internal bus; a computer execution program is stored in the memory; when the processor executes the execution program stored in the memory, the processor can execute the calibration method for the incident angle offset of the spectral imaging center wavelength as described in Example 1. The internal bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, the bus in the attached drawings of the present application is not limited to only one bus or one type of bus. The memory may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk memory, and may also be a USB flash drive, a mobile hard disk, a read-only memory, a disk, or an optical disc, etc. The device can be provided as a terminal, a server, or other forms of devices.

[0103] Figure 16 It is a block diagram of an exemplary device. The device may include one or more of the following components: a processing component, a memory, a power component, a multimedia component, an audio component, an input / output (I / O) interface, a sensor component, and a communication component. The processing component generally controls the overall operation of the electronic device, such as operations associated with display, telephone call, data communication, camera operation, and recording operation. The processing component may include one or more processors to execute instructions to complete all or part of the steps of the above method. In addition, the processing component may include one or more modules to facilitate the interaction between the processing component and other components. For example, the processing component may include a multimedia module to facilitate the interaction between the multimedia component and the processing component.

[0104] The memory is configured to store various types of data to support the operation of the electronic device. Examples of such data include instructions for any application or method operating on the electronic device, contact data, phone book data, messages, pictures, videos, and the like. The memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disk.

[0105] The power supply component provides power to various components of the electronic device. The power supply component may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the electronic device. The multimedia component includes a screen that provides an output interface between the electronic device and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors can not only sense the boundaries of the touch or swipe actions but also detect the duration and pressure associated with the touch or swipe operation. In some embodiments, the multimedia component includes a front camera and / or a rear camera. When the electronic device is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera can receive external multimedia data. Each of the front camera and the rear camera can be a fixed optical lens system or have a focal length and optical zoom capabilities.

[0106] The audio component is configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC) that is configured to receive external audio signals when the electronic device is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals can be further stored in the memory or transmitted via the communication component. In some embodiments, the audio component further includes a speaker for outputting audio signals. The I / O interface provides an interface between the processing component and the peripheral interface module, and the peripheral interface module can be a keyboard, a click wheel, buttons, etc. These buttons can include, but are not limited to: a home button, a volume button, a power-on button, and a lock button.

[0107] The sensor assembly includes one or more sensors for providing status assessment of various aspects for the electronic device. For example, the sensor assembly can detect the on / off state of the electronic device, the relative positioning of components, such as the display and keypad of the electronic device, the sensor assembly can also detect a change in the position of the electronic device or a component of the electronic device, the presence or absence of user contact with the electronic device, the orientation or acceleration / deceleration of the electronic device, and the temperature change of the electronic device. The sensor assembly can include a proximity sensor configured to detect the presence of nearby objects without any physical contact. The sensor assembly can also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly can further include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0108] The communication component is configured to facilitate communication between the electronic device and other devices in a wired or wireless manner. The electronic device can access a wireless network based on communication standards, such as WiFi, 2G, or 3G, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component further includes a near field communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0109] In an exemplary embodiment, the electronic device can be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components for performing the above method.

[0110] Embodiment 3:

[0111] The present invention also provides a computer-readable storage medium storing a computer execution program, which when executed by a processor, is used to implement the calibration method for the incident angle offset of the spectral imaging center wavelength as described in Embodiment 1.

[0112] Specifically, a system, apparatus or device equipped with a readable storage medium can be provided. On this readable storage medium, software program code for implementing the functions of any one of the above embodiments is stored, and the computer or processor of the system, apparatus or device is made to read and execute the instructions stored in the readable storage medium. In this case, the program code read from the readable medium itself can implement the functions of any one of the above embodiments. Therefore, the machine-readable code and the readable storage medium storing the machine-readable code constitute a part of the present invention.

[0113] The above storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk (such as CD-ROM, CD-R, CD-RW, DVD-20ROM, DVD-RAM, DVD-RW, DVD-RW), magnetic tape, etc. The storage medium can be any available medium accessible by a general or special-purpose computer.

[0114] It should be understood that the above processor can be a central processing unit (English: Central Processing Unit, abbreviated: CPU), and can also be other general-purpose processors, digital signal processors (English: Digital Signal Processor, abbreviated: DSP), application specific integrated circuits (English: Application Specific Integrated Circuit, abbreviated: ASIC), etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in combination with the invention can be directly embodied as being executed and completed by a hardware processor, or can be executed and completed by a combination of hardware and software modules in the processor.

[0115] It should be understood that the storage medium is coupled to the processor, so that the processor can read information from the storage medium and can write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an application specific integrated circuit (Application Specific Integrated Circuits, abbreviated: ASIC). Of course, the processor and the storage medium can also exist as discrete components in a terminal or a server.

[0116] The computer-readable program instructions described herein can be downloaded to various computing / processing devices from a computer-readable storage medium or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing / processing device.

[0117] The computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state-setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as the "C" language or similar programming languages. The computer-readable program instructions may be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider). In some embodiments, by using the status information of the computer-readable program instructions to customize an electronic circuit, such as a programmable logic circuit, a field-programmable gate array (FPGA), or a programmable logic array (PLA), the electronic circuit can execute the computer-readable program instructions to implement various aspects of the present disclosure.

[0118] The above are only the preferred embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the protection scope of the present application.

[0119] Although the specific implementation manners of the present invention have been described above, they are not limitations on the protection scope of the present invention. Those skilled in the art should understand that, based on the technical solutions of the present invention, various modifications or deformations that can be made without creative efforts by those skilled in the art are still within the protection scope of the present invention.

Claims

1. A method for establishing a calibration model for the incident angle offset of the central wavelength of spectral imaging, characterized in that, Including the following steps: Step 1, use a spectral camera with a Fabry-Perot interferometer cavity to capture and collect spectral image data of a specific sample with a uniform distribution; the specific sample with a uniform distribution is such that for each incident angle θ within the spectral image range, there is a specific sample substance distributed; the spectral image imaging data is a spectral cube in a three-dimensional form, containing two-dimensional image information of m rows * n columns and one-dimensional spectral information of k spectral bands; Use a spectral camera with a Fabry-Perot interferometer cavity to capture spectral image data of a large standard diffuse reflection reference whiteboard, and the spectral image data of the large standard diffuse reflection reference whiteboard is also a spectral cube in a three-dimensional form, containing two-dimensional image information and one-dimensional spectral information of k spectral bands; Step 2, use the spectral cube of the obtained large standard diffuse reflection reference whiteboard to perform normalization processing on the spectral cube of the specific sample, to obtain the reflectance of the specific sample substance at each incident angle θ, and the reference spectral reflectance of the specific sample substance when vertically incident at 0 degrees at the image center position; Step 3: Establish a dataset set X by associating each incident angle θ with the reflectance of its corresponding specific sample substance θ , and at the same time, establish a set Y of reference spectral reflectances for k spectral bands; X θ = [Reflection_1, Reflection_2,..., Reflection_k, incident angle θ] Reflection_k represents the reflectance at the angle θ in the k-th spectral band; Y = [Reflection_0_1, Reflection_0_2,..., Reflection_0_k] Reflection_0_k represents the reference spectral reflectance when the incident angle is 0 in the k-th spectral band; Step 4, establish the mapping relationship between X θ and Y, and form a training sample data set; Step 5, initialize the parameters of the calibration model, input the training sample data set, and after multiple iterations, stop the iteration when the prediction error is reduced to the set threshold to obtain the final calibration model.

2. The calibration model building method for the incident angle offset of the central wavelength of spectral imaging according to claim 1, wherein: The normalization processing of using the spectral cube of the obtained large standard diffuse reflection reference whiteboard on the spectral cube of the specific sample is as follows: Cube_norm represents a three-dimensional matrix of size m * n * k, representing the spectral cube of the target object after normalization, and the reflectance value Reflection of each pixel point in the two-dimensional image of the spectral cube can be obtained in k spectral bands; Cube_target represents a three-dimensional matrix of size m * n * k, representing the spectral cube of the specific sample before normalization; Cube_white represents a three-dimensional matrix of size m * n * k, representing the spectral cube of the large standard diffuse reflection reference whiteboard.

3. A method for establishing a calibration model for the incident angle offset of the central wavelength of spectral imaging, as described in claim 2, wherein: Also consider the influence of the black reference. In step 1, the spectral cube of the large black reference captured needs to be obtained, that is, close the camera lens cap to ensure there is no ambient light around, and then perform the shooting; the normalization formula is further deformed as: Cube_dark represents a three-dimensional matrix of size m * n * k, representing the spectral cube of the large black reference.

4. The calibration model building method for the incident angle offset of the central wavelength of spectral imaging according to claim 1, wherein, The calculation formula for the incident angle θ is: θ = arctan(w / h) The incident angle θ is calculated based on the horizontal distance w deviating from the optical central axis of the camera lens and the vertical distance h between the specific sample and the lens center.

5. The calibration model building method for the incident angle offset of the central wavelength of spectral imaging according to claim 1, characterized in that: The calibration model adopts any one of a neural network model, a support vector machine model, a logistic regression model, or a Bayesian classification model.

6. A calibration method for the incident angle offset of the central wavelength of spectral imaging, characterized in that, Including the following process: Obtain the spectral image data of a specific sample to be detected captured by a spectral camera passing through a Fabry-Perot interferometer cavity, i.e., an image spectral cube; Use the spectral cube of a large standard diffuse reflection reference whiteboard obtained to normalize the spectral cube of the specific sample to be detected, obtain its reflectance at k spectral bands and the corresponding incident angle θ, and establish a data set; Input the established data set into the calibration model built by the model building method described in any one of claims 1 to 5 for calibration; Output the calibrated reflectance data set of the correct specific sample to be detected for spectral curve plotting.

7. A calibration device for the incident angle offset of the central wavelength of spectral imaging, characterized in that, The device includes at least one processor and at least one memory; a computer execution program is stored in the memory; when the processor executes the execution program stored in the memory, the processor can execute the calibration method for the spectral imaging central wavelength incident angle offset described in claim 6.

8. A computer-readable storage medium, characterized in that: A computer execution program is stored in the computer-readable storage medium, and when the computer execution program is executed by a processor, it is used to implement the calibration method for the spectral imaging central wavelength incident angle offset described in claim 6.

Images