In-situ encryption design method based on rare earth doped lead halide perovskite thin film

Abstract

The invention discloses an in-situ encryption design method based on a rare earth doped lead halide perovskite thin film, and belongs to the technical field of quantum dot imaging. In order to solve the problems that a traditional digital image encryption technology is low in encryption efficiency, poor in real-time performance and difficult to integrate with intelligent equipment, the invention provides an in-situ encryption design method based on a rare earth doped lead halide perovskite thin film. In the image information acquisition process, the spatial distribution of RGB values of an image acquired by a CCD sensor is directly regulated and controlled, the correlation of adjacent pixels of the acquired image is reduced, the information entropy of the image is increased, and simple and efficient in-situ encryption of image information is realized; according to the method, decryption is realized through an inverse process of in-situ encryption, the quality of a decrypted image is improved by using a convolutional neural network algorithm, image acquisition and encryption are synchronously completed, and a simple and efficient technical path is provided for real-time secure transmission of image data in a mobile terminal scene.

Description

Technical Field

[0001] The present invention relates to the field of quantum dot imaging technology, and in particular to an in-situ encryption design method based on rare earth-doped lead halide perovskite thin films. Background Art

[0002] Current mainstream digital image encryption technologies can be categorized into two types: computer cryptography and optical image encryption. Computer cryptography, based on permutation and diffusion mechanisms, uses mathematical methods to encrypt image information, but its efficiency is limited by the algorithm's computational complexity. Optical image encryption, while able to achieve efficient parallel encryption of digital images using optical methods, relies on sophisticated optical systems or micro-nano manufacturing techniques, resulting in complex fabrication processes and difficulties in integration with smart portable devices such as mobile phones. Summary of the Invention

[0003] The purpose of the present invention is to provide an in-situ encryption design method based on rare earth doped lead halide perovskite thin films to solve the problems raised in the above background technology.

[0004] To achieve the above objectives, the present invention provides the following technical solution: an in-situ encryption design method based on rare earth-doped lead halide perovskite thin films, comprising the following steps:

[0005] Prepare a rare earth-doped lead halide perovskite film, wherein the rare earth-doped lead halide perovskite film includes a fluorescent film and a filter film, and construct an in-situ digital image encryption camera system integrated with the rare earth-doped lead halide perovskite film;

[0006] In the in-situ digital image encryption camera system, the target to be captured is input as input information, and the original image is captured through the intensity and color recorded by the CCD sensor;

[0007] Perform in-situ image encryption. Turn on the UV365nm ultraviolet light source to excite the fluorescent film to emit spatially distributed visible light. At the same time, the filter film adjusts the filter bandpass range. The fluorescent key image and filter key image information of the rare earth-doped lead halide perovskite film are integrated into the original image acquisition process to output the encrypted image.

[0008] Regulate the spatial distribution of RGB values of images collected by the CCD sensor to reduce the correlation between adjacent pixels and increase information entropy;

[0009] Implement encrypted image decryption, perform the inverse operation of the encryption process and reconstruct the original image, and use the convolutional neural network algorithm as the decryption network to decrypt the encrypted image.

[0010] Furthermore, preparing a rare earth-doped lead halide perovskite film specifically includes:

[0011] The multi-color fluorescent film and filter film were prepared by spin coating and ion exchange methods;

[0012] Spin coating of a low concentration of rare earth ions and PbX2 in N,N-dimethylformamide solution;

[0013] Add CsX alcohol solution and obtain monochrome CsPbX3:RE by thermal annealing 3+ fluorescent film;

[0014] The PDMS film is used as a mask to cover the designated area of the monochrome fluorescent film. The halogen vapor generated by the volatilization of HBr in acetic acid solution or hydrochloric acid solution is used to perform controllable Br / Cl or I / Br ion exchange in the unprotected area. After multiple mask iterations, multi-color CsPbCl 3-x Br x :RE 3+ and CsPbBr 3-x I x :RE 3+ Fluorescent film system.

[0015] Furthermore, preparing a rare earth-doped lead halide perovskite film specifically includes:

[0016] The preparation method of multicolor fluorescent film is adopted, and the exciton recombination effect of CsPbX3 material is suppressed by high concentration rare earth ion doping to prepare CsPbCl 3-x Br x :RE 3+ and CsPbBr 3-x I x :RE 3+ Filter membrane system.

[0017] Furthermore, an in-situ digital image encryption camera system is constructed, including:

[0018] The in-situ digital image encryption camera system includes a CCD camera, a UV365nm ultraviolet light source, a rare earth-doped lead halide perovskite fluorescent film and a filter film;

[0019] The prepared rare earth-doped lead halide perovskite fluorescent film and filter film are integrated in front of the CCD camera lens, and the fluorescent film is excited by a UV365nm ultraviolet light source to form an in-situ digital image encryption camera system.

[0020] Furthermore, the fluorescence key image and the filter key image information of the rare earth-doped lead halide perovskite film are integrated into the acquisition process of the original image, specifically including:

[0021] The fluorescent film and filter film are integrated in front of the camera lens, and the fluorescent film is excited by a UV365nm ultraviolet light source. The original image, encrypted image and key image are collected separately through the image acquisition function of the CCD camera.

[0022] By changing the pattern of the rare earth-doped lead halide perovskite film or the angle at which the UV365nm ultraviolet light source excites the fluorescent film, the fluorescent information or filter information of the film is modulated to obtain different fluorescent key images and filter key images, thereby changing the encrypted image information collected by the camera.

[0023] The original image is captured without a film, the encrypted image is captured with a UV365nm ultraviolet light source on, and the key image is a superposition of a fluorescent key image and a filtered key image, captured at the same shooting position as the original image with the UV365nm light source on.

[0024] The original image, encrypted image and key image are collected and encrypted through saturated addition and multiplication operations.

[0025] Furthermore, the acquisition and encryption of the original image, the encrypted image, and the key image by saturated addition and multiplication operations are specifically completed by the following formula:

[0026]

[0027] Among them, S P is the pixel value of a given pixel in the original image; K PL is the pixel value of a given pixel in the fluorescent key image; K PT is the pixel value of a given pixel in the filtered key image; C P is the pixel value of a given pixel in the encrypted image; T is the maximum value of the pixel value range. T is determined based on the storage bit width of the CCD sensor. For an 8-bit image, the pixel value range of T is 0-255.

[0028] Furthermore, the method of regulating the spatial distribution of RGB values of an image collected by a CCD sensor to reduce the correlation between adjacent pixels and increase information entropy specifically includes:

[0029] After integrating the fluorescent key image and the filtered key image information, if the pixel value does not exceed the maximum value of the 8-bit image, the pixel difference within the thin film pattern changes, reducing the correlation between adjacent pixels within the thin film pattern and expanding the information entropy;

[0030] By increasing the variety of fluorescent film colors, the pixel value range is broadened and the correlation between patterns is weakened;

[0031] Using wide half-width fluorescent materials or broadband filter materials to increase the effective wavelength integration range, increase the pixel value fluctuation amplitude, reduce the correlation between adjacent blocks and expand information entropy;

[0032] After integrating the fluorescent key image and the filtered key image information, the pixel value change recorded at a certain pixel point on the CCD sensor is expressed by the following formula:

[0033]

[0034] Among them, D PLλ (x, y) is the pixel value of a point on the fluorescent film, D PTλ (x, y) is the pixel value of a point on the filter film, S λ (x,y) and D λ (x, y) are the spectrum of a point on the object and the camera response curve corresponding to the point on the object.

[0035] Furthermore, when decrypting the encrypted image, performing the inverse operation of the encryption process and reconstructing the original image, the information of the encrypted image and the information of the rare earth-doped lead halide perovskite film are separated through saturated division and subtraction operations, which is specifically accomplished by the following formula:

[0036]

[0037] in, Represents the pixel value of a given pixel in the decrypted image.

[0038] Furthermore, the decryption of the encrypted image using a convolutional neural network as a decryption network specifically includes:

[0039] The convolutional neural network takes the encrypted image and the key image as input and outputs the decrypted original image. The network structure used by the convolutional neural network is modified based on Unet. The network structure includes 4 encoder stages and 4 decoder stages.

[0040] At the end of each encoder stage, the feature map is downsampled to 1 / 2 size using convolution with a kernel size of 4×4 and a stride of 2;

[0041] Before each decoder stage, the feature map is upsampled to 2 times the scale using bilinear interpolation;

[0042] Pass low-level feature maps from each encoder stage to the corresponding decoder stage via skip connections;

[0043] Introducing residual learning in convolutional neural networks to simplify training.

[0044] Furthermore, the introduction of residual learning in the convolutional neural network to simplify training specifically includes:

[0045] Use the residual module as the basic module of the encoder and decoder, which consists of two 3×3 convolutional layers, a ReLU activation layer and a 1×1 convolutional layer;

[0046] A dataset containing paired encrypted images, key images, and original images is collected for decryption network training. 256×256 spatial image patches are randomly cropped from the collected dataset and optimized using the Adam optimizer.

[0047] The dataset is divided into a training set and a test set with a ratio of 8:2. The initial learning rate and batch size parameters are optimized.

[0048] Based on the data set and parameter settings, the decryption network is trained and the network parameters are optimized.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] The present invention introduces the fluorescence and filtering information of rare earth-doped lead halide perovskite materials into the digital image acquisition process. By designing an in-situ digital image encryption camera based on a rare earth-doped lead halide perovskite film, the tunable fluorescence and filtering properties of the rare earth-doped lead halide perovskite film are utilized to directly regulate the spatial distribution of the RGB values of the image captured by the CCD sensor, achieving simultaneous image acquisition and encryption. This destroys the correlation between adjacent pixels of the captured image (<0.1) and increases its information entropy (>7.8), thereby achieving simple and efficient in-situ encryption of image information. Decryption is achieved through the inverse process of in-situ encryption, and a convolutional neural network is used to optimize the decrypted image quality (peak signal-to-noise ratio >35dB). Compared with traditional digital image encryption technologies, this technology has the advantages of simple operation, fast encryption speed, and ease of miniaturization. This technology provides a simple and efficient in-situ image encryption technology, offering a new solution for the secure transmission of digital images and a simple and efficient technical path for the real-time secure transmission of image data in mobile terminal scenarios (such as mobile phones). The technology has potential application value in fields such as personal privacy protection, medical imaging, and military communications. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a working principle diagram of the in-situ digital image encryption camera of the rare earth-doped lead halide perovskite film of the present invention;

[0052] Figure 2 This is a schematic diagram of the fluorescent film preparation process of the present invention;

[0053] Figure 3 This is a schematic diagram of a principle prototype of the in-situ digital image encryption of the present invention;

[0054] Figure 4 This is a schematic diagram of the specific process of decryption using the convolutional neural network algorithm in the present invention. DETAILED DESCRIPTION

[0055] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0056] See also Figure 1-4 , the present invention provides the following technical solutions:

[0057] The in-situ encryption design method based on rare earth doped lead halide perovskite film includes the following steps:

[0058] Prepare rare earth-doped lead halide perovskite thin films, including fluorescent films and filter films, and construct an in-situ digital image encryption camera system integrated with the rare earth-doped lead halide perovskite thin films;

[0059] In the in-situ digital image encryption camera system, the target to be captured is input as input information, and the original image is captured through the intensity and color recorded by the CCD sensor;

[0060] Perform in-situ image encryption. Turn on the UV365nm ultraviolet light source to excite the fluorescent film to emit spatially distributed visible light. At the same time, the filter film adjusts the filter bandpass range. The fluorescent key image and filter key image information of the rare earth-doped lead halide perovskite film are integrated into the original image acquisition process to output the encrypted image.

[0061] Regulate the spatial distribution of RGB values of images collected by the CCD sensor to reduce the correlation between adjacent pixels and increase information entropy;

[0062] Implement encrypted image decryption, perform the inverse operation of the encryption process and reconstruct the original image, and use the convolutional neural network algorithm as the decryption network to decrypt the encrypted image.

[0063] In the above embodiment, a new in-situ digital image encryption method based on the tunable optical properties of rare earth-doped lead halide perovskite thin films directly regulates the spatial distribution of the RGB values of the CCD sensor through the fluorescence and filtering characteristics of the thin film, thereby achieving simultaneous completion of image acquisition and encryption. This solves the problems of low efficiency and poor real-time performance of traditional encryption technologies, and provides a simple and efficient technical path for the real-time and secure transmission of image data in mobile terminal scenarios (such as mobile phones). It has potential application value in fields such as personal privacy protection, medical imaging, and military communications.

[0064] Preparation of rare earth doped lead halide perovskite thin film, specifically comprising:

[0065] The multi-color fluorescent film and filter film were prepared by spin coating and ion exchange methods;

[0066] Spin coating of a low concentration of rare earth ions and PbX2 in N,N-dimethylformamide solution;

[0067] Add CsX alcohol solution and obtain monochrome CsPbX3:RE by thermal annealing 3+ fluorescent film;

[0068] The PDMS film is used as a mask to cover the designated area of the monochrome fluorescent film. The halogen vapor generated by the volatilization of HBr in acetic acid solution or hydrochloric acid solution is used to perform controllable Br / Cl or I / Br ion exchange in the unprotected area. After multiple mask iterations, multi-color CsPbCl 3-x Br x :RE 3+ and CsPbBr 3-x I x :RE 3+ Fluorescent film system;

[0069] The preparation method of multicolor fluorescent film is adopted, and the exciton recombination effect of CsPbX3 material is suppressed by high concentration rare earth ion doping to prepare CsPbCl 3-x Br x :RE 3+ and CsPbBr 3-x I x :RE 3+ Filter membrane system.

[0070] In the above-mentioned embodiment, a combination of spin coating and ion exchange methods enables the controllable preparation of multi-color fluorescent and filter films. Rare earth ion doping and halogen ratio control are employed to impart tunable fluorescence wavelength, intensity, and filter bandpass to the film, resulting in a film transmittance exceeding 80%, providing a highly stable optical control medium for subsequent image encryption. Furthermore, the multi-color film system, constructed through multiple mask iterations, systematically reduces the correlation between adjacent pixels in the captured image at the material level by reducing the film unit size and increasing the number of fluorescent colors, laying the foundation for improving the information entropy of encrypted information.

[0071] Build an in-situ digital image encryption camera system, including:

[0072] The in-situ digital image encryption camera system includes a CCD camera, a UV365nm ultraviolet light source, a rare earth-doped lead halide perovskite fluorescent film, and a filter film;

[0073] The prepared rare earth-doped lead halide perovskite fluorescent film and filter film are integrated in front of the CCD camera lens, and the fluorescent film is excited by a UV365nm ultraviolet light source to form an in-situ digital image encryption camera system.

[0074] In the above embodiment, the design of integrating the CCD camera, UV365nm ultraviolet light source and rare earth-doped thin film breaks through the limitation of traditional optical encryption relying on precision systems. Through the modular integration of the film and the lens, the hardware-level fusion of image acquisition and encryption is realized. It has the characteristics of small size and easy deployment, and can be directly adapted to the imaging module of mobile terminals such as mobile phones. The synergistic excitation mechanism of the UV light source and the film eliminates the need for additional algorithm calculations to generate fluorescent keys and filter keys, significantly improving the real-time performance of the encryption process and solving the pain point that traditional technologies are difficult to integrate with smart devices.

[0075] The fluorescence key image and filter key image information of the rare earth doped lead halide perovskite film are integrated into the acquisition process of the original image, specifically including:

[0076] The fluorescent film and filter film are integrated in front of the camera lens, and the fluorescent film is excited by a UV365nm ultraviolet light source. The original image, encrypted image and key image are collected separately through the image acquisition function of the CCD camera.

[0077] By changing the pattern of the rare earth-doped lead halide perovskite film or the angle at which the UV365nm ultraviolet light source excites the fluorescent film, the fluorescent information or filter information of the film is modulated to obtain different fluorescent key images and filter key images, thereby changing the encrypted image information collected by the camera.

[0078] The original image is captured without a film, the encrypted image is captured with the UV365nm light source on, and the key image is a superposition of the fluorescent key image and the filtered key image, captured at the same shooting position as the original image with the UV365nm light source on.

[0079] The acquisition and encryption of the original image, encrypted image, and key image are performed through saturated addition and multiplication operations, which are specifically completed by the following formula:

[0080]

[0081] Among them, S P is the pixel value of a given pixel in the original image; K PL is the pixel value of a given pixel in the fluorescent key image; K PT is the pixel value of a given pixel in the filtered key image; C P is the pixel value of a given pixel in the encrypted image; T is the maximum value of the pixel value range. T is determined based on the storage bit width of the CCD sensor. For an 8-bit image, the pixel value range of T is 0-255.

[0082] In the above-described embodiment, the optical properties of the film are converted into key information for image encryption through hardware-level integration of saturated addition and multiplication operations. By varying the film pattern or the light source excitation angle modulation method, the same hardware system can generate an unlimited number of key combinations, significantly expanding the encryption system's key space. The synchronized acquisition mechanism for the original image, encrypted image, and key image ensures the integrity of the encryption process, while pixel value calculations based on the CCD storage bit width avoid data overflow and ensure the reliability and decryptability of the encrypted image.

[0083] Regulate the spatial distribution of RGB values of images collected by the CCD sensor to reduce the correlation between adjacent pixels and increase information entropy. Specifically, this includes:

[0084] After integrating the fluorescent key image and the filtered key image information, if the pixel value does not exceed the maximum value of the 8-bit image, the pixel difference within the thin film pattern changes, reducing the correlation between adjacent pixels within the thin film pattern and expanding the information entropy;

[0085] By increasing the variety of fluorescent film colors, the pixel value range is broadened and the correlation between patterns is weakened;

[0086] Using wide half-width fluorescent materials or broadband filter materials to increase the effective wavelength integration range, increase the pixel value fluctuation amplitude, reduce the correlation between adjacent blocks and expand information entropy;

[0087] After integrating the fluorescent key image and the filtered key image information, the pixel value change recorded at a certain pixel point on the CCD sensor is expressed by the following formula:

[0088]

[0089] Among them, D PLλ (x, y) is the pixel value of a point on the fluorescent film, D PTλ (x, y) is the pixel value of a point on the filter film, S λ (x,y) and D λ (x, y) are the spectrum of a point on the object and the camera response curve corresponding to the point on the object.

[0090] In the above-described embodiment, the coordinated modulation of fluorescence and filter information achieves deep destruction of image statistical features. The computational logic of superimposing fluorescence film information and multiplying it with filter film information significantly increases the amplitude of pixel value fluctuations, reducing the correlation between adjacent pixels to below 0.1 and increasing information entropy to above 7.8, effectively eliminating spatial redundancy in the image. The use of wide half-width-at-half-maximum fluorescent materials and broadband filter materials further enhances the randomness of pixel values by expanding the wavelength integration range, bringing the encrypted image closer to an ideal noise distribution and improving encryption strength from an information-theoretic perspective.

[0091] When decrypting the encrypted image, performing the inverse operation of the encryption process and reconstructing the original image, the encrypted image information and the rare earth-doped lead halide perovskite film information are separated through saturated division and subtraction operations, which is specifically accomplished by the following formula:

[0092]

[0093] in, Represents the pixel value of a given pixel in the decrypted image.

[0094] Use convolutional neural network as the decryption network to decrypt the encrypted image, including:

[0095] The convolutional neural network takes the encrypted image and the key image as input and outputs the decrypted original image. The network structure used by the convolutional neural network is modified based on Unet. The network structure includes 4 encoder stages and 4 decoder stages.

[0096] At the end of each encoder stage, the feature map is downsampled to 1 / 2 size using convolution with a kernel size of 4×4 and a stride of 2;

[0097] Before each decoder stage, the feature map is upsampled to 2 times the scale using bilinear interpolation;

[0098] Pass low-level feature maps from each encoder stage to the corresponding decoder stage via skip connections;

[0099] Introducing residual learning into convolutional neural networks to simplify training;

[0100] Use the residual module as the basic module of the encoder and decoder. The residual module consists of two 3×3 convolutional layers, a ReLU activation layer, and a 1×1 convolutional layer.

[0101] A dataset containing paired encrypted images, key images, and original images is collected for decryption network training. 256×256 spatial image patches are randomly cropped from the collected dataset and optimized using the Adam optimizer.

[0102] The dataset is divided into a training set and a test set with a ratio of 8:2. The initial learning rate and batch size parameters are optimized.

[0103] Based on the data set and parameter settings, train the decryption network and optimize the network parameters

[0104] In the above embodiment, the introduction of the Unet architecture and the CNN decryption network with residual learning solves the problem of information loss that may occur in the actual encryption process. The network realizes multi-level extraction and recovery of encrypted image features through the encoder-decoder structure and skip connections. After 100 rounds of training and optimization, the peak signal-to-noise ratio exceeds 35dB, ensuring high-quality reconstruction of the decrypted image. The Adam optimizer and the scientific division of the data set improve the generalization ability of the network, making the decryption process both efficient and robust, and adapting to the encrypted image decryption needs in different scenarios.

[0105] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. An in-situ densification design method based on rare earth-doped lead halide perovskite thin films, characterized in that: The following steps are involved: Prepare a rare earth-doped lead halide perovskite film, wherein the rare earth-doped lead halide perovskite film includes a fluorescent film and a filter film, and construct an in-situ digital image encryption camera system integrated with the rare earth-doped lead halide perovskite film; In the in-situ digital image encryption camera system, the target to be captured is input as input information, and the original image is captured through the intensity and color recorded by the CCD sensor; Perform in-situ image encryption. Turn on the UV365nm ultraviolet light source to excite the fluorescent film to emit spatially distributed visible light. At the same time, the filter film adjusts the filter bandpass range. The fluorescent key image and filter key image information of the rare earth-doped lead halide perovskite film are integrated into the original image acquisition process to output the encrypted image. Regulate the spatial distribution of RGB values of images collected by the CCD sensor to reduce the correlation between adjacent pixels and increase information entropy; Implement encrypted image decryption, perform the inverse operation of the encryption process and reconstruct the original image, and use the convolutional neural network algorithm as the decryption network to decrypt the encrypted image.

2. The in-situ encryption design method based on rare earth doped lead halide perovskite film according to claim 1, characterized in that: Preparation of rare earth doped lead halide perovskite thin film, specifically comprising: The multi-color fluorescent film and filter film were prepared by spin coating and ion exchange methods; Spin coating of a low concentration of rare earth ions and PbX2 in N,N-dimethylformamide solution; Add CsX alcohol solution and obtain monochrome CsPbX3:RE by thermal annealing 3+ fluorescent film; The PDMS film is used as a mask to cover the designated area of the monochrome fluorescent film. The halogen vapor generated by the volatilization of HBr in acetic acid solution or hydrochloric acid solution is used to perform controllable Br / Cl or I / Br ion exchange in the unprotected area. After multiple mask iterations, multi-color CsPbCl 3-x Br x :RE 3+ and CsPbBr 3-x I x :RE 3+ Fluorescent film system.

3. The in-situ encryption design method based on rare earth doped lead halide perovskite film according to claim 2, characterized in that: Preparation of rare earth doped lead halide perovskite thin film, specifically also including: The preparation method of multicolor fluorescent film is adopted, and the exciton recombination effect of CsPbX3 material is suppressed by high concentration rare earth ion doping to prepare CsPbCl 3-x Br x :RE 3+ and CsPbBr 3-x I x :RE 3+ Filter membrane system.

4. The in-situ encryption design method based on rare earth doped lead halide perovskite film according to claim 1, characterized in that: Build an in-situ digital image encryption camera system, including: The in-situ digital image encryption camera system includes a CCD camera, a UV365nm ultraviolet light source, a rare earth-doped lead halide perovskite fluorescent film and a filter film; The prepared rare earth-doped lead halide perovskite fluorescent film and filter film are integrated in front of the CCD camera lens, and the fluorescent film is excited by a UV365nm ultraviolet light source to form an in-situ digital image encryption camera system.

5. The in-situ encryption design method based on rare earth doped lead halide perovskite film according to claim 1, characterized in that: The fluorescence key image and filter key image information of the rare earth doped lead halide perovskite film are integrated into the acquisition process of the original image, specifically including: The fluorescent film and filter film are integrated in front of the camera lens, and the fluorescent film is excited by a UV365nm ultraviolet light source. The original image, encrypted image and key image are collected separately through the image acquisition function of the CCD camera. By changing the pattern of the rare earth-doped lead halide perovskite film or the angle at which the UV365nm ultraviolet light source excites the fluorescent film, the fluorescent information or filter information of the film is modulated to obtain different fluorescent key images and filter key images, thereby changing the encrypted image information collected by the camera. The original image is captured without a film, the encrypted image is captured with a UV365nm ultraviolet light source on, and the key image is a superposition of a fluorescent key image and a filtered key image, captured at the same shooting position as the original image with the UV365nm light source on. The original image, encrypted image and key image are collected and encrypted through saturated addition and multiplication operations.

6. The in-situ encryption design method based on rare earth doped lead halide perovskite film according to claim 5, characterized in that: The acquisition and encryption of the original image, the encrypted image, and the key image by saturated addition and multiplication operations are specifically completed by the following formula: Among them, S P is the pixel value of a given pixel in the original image; K PL is the pixel value of a given pixel in the fluorescent key image; K PT is the pixel value of a given pixel in the filtered key image; C P is the pixel value of a given pixel in the encrypted image; T is the maximum value of the pixel value range. T is determined based on the storage bit width of the CCD sensor. For an 8-bit image, the pixel value range of T is 0-255.

7. The in-situ encryption design method based on rare earth doped lead halide perovskite film according to claim 1, characterized in that: The method of regulating the spatial distribution of RGB values of an image collected by a CCD sensor, reducing the correlation between adjacent pixels, and increasing information entropy specifically includes: After integrating the fluorescent key image and the filtered key image information, if the pixel value does not exceed the maximum value of the 8-bit image, the pixel difference within the thin film pattern changes, reducing the correlation between adjacent pixels within the thin film pattern and expanding the information entropy; By increasing the variety of fluorescent film colors, the pixel value range is broadened and the correlation between patterns is weakened; Using wide half-width fluorescent materials or broadband filter materials to increase the effective wavelength integration range, increase the pixel value fluctuation amplitude, reduce the correlation between adjacent blocks and expand information entropy; After integrating the fluorescent key image and the filtered key image information, the pixel value change recorded at a certain pixel point on the CCD sensor is expressed by the following formula: Among them, D PLλ (x, y) is the pixel value of a point on the fluorescent film, D PTλ (x, y) is the pixel value of a point on the filter film, S λ (x,y) and D λ (x, y) are the spectrum of a point on the object and the camera response curve corresponding to the point on the object.

8. The in-situ encryption design method based on rare earth doped lead halide perovskite film according to claim 1, characterized in that: When decrypting the encrypted image, performing the inverse operation of the encryption process and reconstructing the original image, the encrypted image information and the rare earth-doped lead halide perovskite film information are separated through saturated division and subtraction operations, which is specifically accomplished by the following formula: in, Represents the pixel value of a given pixel in the decrypted image.

9. The in-situ encryption design method based on rare earth doped lead halide perovskite film according to claim 1, characterized in that: The method of using a convolutional neural network as a decryption network to decrypt an encrypted image specifically includes: The convolutional neural network takes the encrypted image and the key image as input and outputs the decrypted original image. The network structure used by the convolutional neural network is modified based on Unet. The network structure includes 4 encoder stages and 4 decoder stages. At the end of each encoder stage, the feature map is downsampled to 1 / 2 size using convolution with a kernel size of 4×4 and a stride of 2; Before each decoder stage, the feature map is upsampled to 2 times the scale using bilinear interpolation; Pass low-level feature maps from each encoder stage to the corresponding decoder stage via skip connections; Introducing residual learning in convolutional neural networks to simplify training.

10. The in-situ encryption design method based on rare earth doped lead halide perovskite film according to claim 9, characterized in that: The introduction of residual learning to simplify training in convolutional neural networks specifically includes: Use the residual module as the basic module of the encoder and decoder, which consists of two 3×3 convolutional layers, a ReLU activation layer and a 1×1 convolutional layer; A dataset containing paired encrypted images, key images, and original images is collected for decryption network training. 256×256 spatial image patches are randomly cropped from the collected dataset and optimized using the Adam optimizer. The dataset is divided into a training set and a test set with a ratio of 8:

2. The initial learning rate and batch size parameters are optimized. Based on the data set and parameter settings, the decryption network is trained and the network parameters are optimized.

Images