Infrared detection device based on super-structure lens, design method and use method thereof

Abstract

The invention relates to the technical field of super-structure lens design, and provides an infrared detection device based on a super-structure lens, a design method and a use method thereof, which can be suitable for extreme environments such as high radiation and lay a solid foundation for research on thermal infrared imaging leakage detection equipment which is miniaturized, has high performance and is suitable for nuclear facility environments.

Description

Technical Field

[0001] The present invention relates to the technical field of metasurface lens design, and specifically, to an infrared detection device based on a metasurface lens, a design method thereof, and a usage method thereof. Background Art

[0002] The safety of nuclear facilities is of utmost importance. As the operating time of nuclear facilities increases, problems such as bursting, leakage, and damage may occur in system-related industrial equipment (such as pipelines), resulting in a decrease in the stability of the system and posing potential hazards to the safe production of nuclear facilities. Therefore, it is of great significance and value to develop an efficient, convenient, and reliable method for detecting the state of industrial equipment. Infrared thermal imaging is an imaging method that uses an infrared imaging detection device to display the temperature value distribution of the object or scene to be measured. Although the human eye cannot observe infrared radiation, the infrared imaging detection device can convert it into a visible thermal image, showing the temperature distribution of the object or scene to be measured. Therefore, the infrared imaging detection device can break through the human visual barrier, can detect in various complex environments, and has the advantages of good concealment, strong anti-interference ability, and strong target recognition ability, and has a wide range of applications in the fields of nuclear facility monitoring, high-voltage power grid detection, security monitoring, aerospace, military equipment, etc.

[0003] Infrared thermal imaging technology has the following characteristics: 1. Non-contact measurement; 2. Radiation energy detection; 3. The results are intuitive and easy to understand. Therefore, in addition to making the identification of high-temperature areas more accurate, thermal imagers can also help workers work more safely. It can not only reduce the time required for workers to diagnose situations in adverse areas, but also allow workers to analyze the operation of radioactive equipment from a distance, reducing or avoiding contact with radioactive substances. Therefore, infrared thermal imaging can be regarded as a powerful tool for monitoring the state of operating nuclear facilities. Currently, it has been used for the state monitoring of faults such as radioactive waste liquid tanks, cooling water pumps, slurry pumps, and heavy water leakage in the fuel channels of heavy water reactors; the method of using infrared thermal imaging can more accurately find the location of equipment faults and can monitor the equipment fault state in real time during equipment operation, reducing risks such as bursting, leakage, and damage caused by thinning due to corrosion and fatigue.

[0004] An infrared thermal imaging system mainly consists of an infrared optical system, a core unit, an intelligent processing circuit, a battery, a housing, a display screen, etc. Among them, the infrared lens is the core part of the infrared imaging system and is one of the research focuses of infrared thermal imaging. However, the structural volume of existing infrared thermal imaging systems is mainly occupied by the optical system. The lenses used in infrared optical systems generally consist of curved lenses with a certain thickness and weight, resulting in difficulties in miniaturizing, integrating, and arraying infrared thermal imaging devices.

[0005] In this context, with the increasing requirements of optical systems for imaging quality and system structure, the advantages of diffractive optical technology in the field of optical imaging systems have become more prominent. The metasurface lens designed based on the principle of metasurface can efficiently control the characteristics of light waves such as amplitude, phase, polarization, and frequency through sub-wavelength-sized microelement structures, and inherently has spherical aberration. In the past, the functions achieved by multiple traditional lenses often only require one metasurface lens to replace, greatly reducing the volume of the traditional lens group and having good imaging quality, becoming the key to solving the above problems. Summary of the Invention

[0006] One of the purposes of the present invention is to propose a design method for a metasurface lens. According to the characteristics of the infrared radiation parameters of the working facilities, infrared metasurface lenses working in different infrared bands are specially designed, and a backend target recognition algorithm is supplemented to detect the pipeline breakage and leakage in real time. It has the characteristics of being thin and light in structure, high diffraction efficiency, and a wide applicable spectral range, and can be applied to extreme environments such as high radiation, laying a solid foundation for the research of miniaturized, high-performance thermal infrared imaging leakage detection equipment applicable to the nuclear facility environment.

[0007] The technical solution of the present invention is as follows:

[0008] A design method for a metasurface lens includes the following steps:

[0009] S100: According to the wavelength range of the infrared electromagnetic waves radiated by industrial equipment due to breakage and leakage faults, confirm the working band of the metasurface lens and construct a microelement database according to the working band;

[0010] S200: Calculate and optimize the phase distribution of the metasurface lens, and match it with the actual microelements in the database to obtain the required metasurface lens;

[0011] S300: Use micro-nano processing technology to process and prepare the metasurface lens, and connect the prepared metasurface lens with an infrared detector to obtain an infrared imaging device;

[0012] S400: The infrared imaging device takes pictures to establish a data set, uses the YOLOv4 framework to establish a leakage detection model on the basis of the data set, and integrates the model and the infrared imaging device to construct a complete infrared monitoring system to monitor the facilities in real time.

[0013] Further, step S100 includes:

[0014] S110: Determine the working band according to the change of infrared electromagnetic waves at the breakage or leakage of the nuclear equipment, and select multiple discretely distributed wavelengths from the working band. Select suitable materials as the substrate and microelement structure, and calculate the phase and transmittance change of the scattering field of the microelement structure under different structural parameters at the selected wavelengths to obtain the microelement database;

[0015] S120: Select a suitable material as the substrate and the micro - element structure. At the selected wavelength, calculate the amplitude, phase, and transmittance change of the scattering field of the micro - element structure under different structural parameters to obtain a micro - element database.

[0016] Further, S210: After the incident light field U in (x, y) passes through the metasurface lens, the complex amplitude U out (x, y) of the actual light field in the focal plane can be obtained. The calculation formula for the distribution is:

[0017]

[0018] where is the phase modulation of the incident light by the metasurface lens;

[0019] F, F -1 denote the Fourier transform;

[0020] f is the focal length of the lens;

[0021] λ is the wavelength of the incident electromagnetic wave;

[0022] j is the imaginary unit;

[0023] μ, ν are the corresponding frequency - domain coordinates in the Fourier transform;

[0024] S220: After obtaining the light field U out (x, y) in the focal plane, based on the difference between the complex amplitude distribution of the actual light field in the focal plane and the complex amplitude U ideal (x, y) distribution of the ideal light field, and the difference between the actual phase that the micro - element can control and the phase arrangement of the simulated metasurface lens in the database, establish a loss function and calculate the system loss;

[0025] S230: Based on the obtained loss, use an optimization algorithm to inversely update the phase arrangement of the metasurface lens. Iterate for multiple rounds until the focusing efficiency is stable, and then end the optimization to obtain the phase arrangement of the metasurface lens with the optimal achromatic effect;

[0026] S240: Select a suitable micro - element structure from the database for spatial arrangement to precisely match the required spatial phase distribution of the metasurface lens.

[0027] Another object of the present invention is to propose an infrared detection device based on a metasurface lens, including a metasurface lens, an infrared detector, and a connecting device for connecting the metasurface lens and the infrared detector. The metasurface lens is designed by the method according to any one of claims 1 to 3;

[0028] The connecting device includes a front - end part and a rear - end part. The front - end part and the rear - end part are respectively placed with the metasurface lens and the infrared detector. The two parts are connected by threads.

[0029] Further, the front end part and the rear end part are threadedly connected through a threaded rotating device to adjust the distance between the metasurface lens and the infrared detector, so that the changing infrared electromagnetic waves emitted by the device due to leakage and breakage faults can act on the infrared detector to present a better and clearer image.

[0030] When designing the overall length of the connecting device, the focal length of the infrared lens should be considered to avoid the situation where the length of the device is too long and the infrared detector cannot obtain a clear thermal image.

[0031] Another object of the present invention is to propose a method for using an infrared detection device based on a metasurface lens, and the steps are as follows:

[0032] S500: Take pictures related to equipment breakage and leakage, and label the fault areas of the equipment to obtain a labeled infrared image dataset.

[0033] S600: Enhance the image data, divide the enhanced dataset according to a ratio of 6:2:2 to obtain a training set, a validation set and a test set; use a clustering algorithm to calculate the width and height of the equipment breakage area in the current dataset, so as to determine the size of the detection frame and improve the detection accuracy;

[0034] S700: Build a YOLOv4 model framework, and introduce an attention mechanism module to strengthen the attention to the areas where the infrared electromagnetic waves change due to faults such as leakage and breakage of the equipment, to obtain a detection model for faulty equipment to be trained;

[0035] And use the obtained training set, validation set and test set to train, validate and test the constructed detection model to obtain a trained real-time monitoring model for industrial equipment.

[0036] S800: Use the infrared monitoring device composed of the infrared imaging system and the monitoring model to monitor industrial equipment, and timely feedback the area to the staff when problems such as leakage, breakage and explosion occur.

[0037] Further, the image data enhancement method in step S600 is:

[0038] Perform data enhancement operations such as translation, rotation, cutting, and scaling on the initial dataset to obtain a complete enhanced dataset.

[0039] The working principle and beneficial effects of the present invention are:

[0040] Compared with existing industrial facility detection equipment and design methods, the present invention combines the design of metasurface lenses, the YOLOv4 target detection algorithm, etc., and optimally designs a thin, light, integratable, and infrared detection device with a suitable working band and its implementation method. The infrared detection device can not only break through the human visual barrier and achieve detection in various complex environments, with advantages such as good concealment, strong anti-interference ability, and strong target recognition ability, but also compared with the infrared optical system lens composed of curved lenses with a certain thickness and weight in the traditional sense, it has the characteristics of simple and thin structure, high diffraction efficiency, wide applicable spectral range, and easy processing. One lens can be used to replace the traditional lens group to achieve the same effect, and it can be applied to extreme environments such as high temperature and high radiation, and has strong practicability in the field of miniaturized and high-performance thermal infrared imaging detection equipment for nuclear industrial facilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The present invention will be further described in detail below in conjunction with the drawings and specific embodiments.

[0042] Figure 1 is the overall design process of the metasurface lens;

[0043] Figure 2 is the top view of four micro-elements;

[0044] Figure 3 is the partial GDS diagram of the metasurface lens;

[0045] Figure 4 is the schematic diagram of the working process of the infrared detection device;

[0046] Figure 5 is the front part of the integrated packaging device of the metasurface lens and the movable infrared detector;

[0047] Figure 6 is the rear part of the integrated packaging device of the metasurface lens and the movable infrared detector;

[0048] Figure 7 is the connection method between the connecting device and the infrared detector;

[0049] Figure 8 is an example of the model detecting pipeline leakage Figure 1 ;

[0050] Figure 9 is an example of the model detecting pipeline leakage Figure 2 . SPECIFIC EMBODIMENTS

[0051] The following will describe the technical solutions in the embodiments of the present invention clearly and completely in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0052] The following embodiments disclose an infrared detection device, a design method and a use method thereof based on a metasurface lens, which relates to the field of infrared thermal imaging and can be used for monitoring high-sensitive conditions in the nuclear industry, such as detecting leaks of hot fluids in pipelines. It mainly includes an infrared detection unit based on a metasurface lens and an intelligent infrared image recognition and detection unit. Its working principle is that when industrial facilities such as pipelines undergo thermal changes due to faults such as defects, bursts, and leaks, they radiate infrared electromagnetic waves of different wavelengths. The infrared electromagnetic waves act on the metasurface lens with multi-wavelength achromatic function and are focused on the infrared detector to obtain a clear infrared image. Then, the obtained infrared image is input into a target recognition model constructed with the YOLOv4 model framework for detection, and the position where the fault occurs can be obtained in real time. The advantages of this application are that it can not only locate the fault position relatively quickly, without being interfered by occlusion or insufficient illumination, but also can monitor the state of industrial equipment in real time during equipment operation to prevent faults from occurring. At the same time, a metasurface lens is used as the main part of the detection unit. Compared with the traditional optical system composed of curved lenses with a certain thickness and weight, this device has the characteristics of being thin and light in structure, high diffraction efficiency, and can independently design the working band according to requirements, and can be applied to extreme environments such as high temperature and high radiation, and can be applied to the field of thermal infrared imaging detection equipment for nuclear facilities.

[0053] Embodiment 1

[0054] As Figure 1 shown, a design method of a metasurface lens includes the following steps:

[0055] S100: According to the wavelength range of the infrared electromagnetic waves radiated by industrial equipment due to damage and leakage faults, confirm the working band of the metasurface lens and construct a microelement database according to the working band;

[0056] Among them, step S100 includes:

[0057] S110: Determine the working band according to the change of infrared electromagnetic waves at the damaged or leaked part of the nuclear equipment, and select a plurality of discretely distributed wavelengths from the working band. Select suitable materials as the substrate and microelement structure, and calculate the phase and transmittance change of the scattering field of the microelement structure under different structural parameters at the selected wavelengths to obtain the microelement database;

[0058] Specifically, the infrared radiation change caused by leakage is mainly in the long infrared band. Therefore, in this embodiment, seven discrete wavelengths of 9μm, 9.5μm, 10μm, 10.5μm, 11μm, 11.5μm, and 12μm are selected from 9μm to 12μm for simulation, and the lens focal length f and the lens size are determined. The microelement height h is set to 8μm, and a single microelement structural unit P x *P y = 6.2μm * 6.2um. In this embodiment, the substrate and the nanocolumns can be made of commonly used optical materials such as silicon or germanium. Both have high refractive index, transmittance, and phase modulation ability in the long infrared band, and are relatively ideal dielectric materials.

[0059] As Figure 2 shown, the microelement types are divided into four: square columns, circular columns, a single substrate carrying four rectangular columns, and a single substrate carrying four circular columns. The geometric parameters of the rectangular columns and circular columns are the side length L and the radius R respectively, and the geometric parameters of the toroidal columns are the inner diameter R_1 and the outer diameter R_2. The latter two are consistent with the geometric parameters of the individual rectangular columns and circular columns. The geometric parameters of the microelement structure are all adjustable variables and all have high rotational symmetry, making the metasurface lens polarization-independent.

[0060] At the seven selected discrete wavelengths, calculate the transmittance and phase change of the scattered field of the microelement structure with different structural parameters; specifically, the phase change of the microelement structure is calculated using the finite element method or the finite difference time domain method. Import the model of the structural element into the software of the finite element method or the finite difference time domain method to calculate the phase shift database and transmittance database of the five microelements at the seven wavelengths and different geometric parameters.

[0061] S200: Use the diffraction neural network algorithm to optimize the phase arrangement of the metasurface lens, calculate the optimized phase distribution of the metasurface lens, and match it with the actual microelements in the database to obtain the required metasurface lens, specifically as follows:

[0062] S210: When the incident light field U in (x, y) passes through the metasurface lens, the complex amplitude U out (x, y) distribution formula of the actual light field on the focal plane can be obtained:

[0063]

[0064] where is the phase modulation of the incident light by the metasurface lens;

[0065] F, F -1 represents the Fourier transform;

[0066] f is the lens focal length;

[0067] λ is the wavelength of the incident electromagnetic wave;

[0068] j is the imaginary unit;

[0069] μ and ν are the corresponding frequency domain coordinates in the Fourier transform;

[0070] S220: After obtaining the focused plane optical field U out (x, y), then based on the difference between the actual complex amplitude distribution of the optical field in the focused plane and the ideal optical field complex amplitude U ideal (x, y) distribution, and the difference between the phase that the micro - element can actually control and the phase arrangement of the simulated metasurface lens in the database, a loss function is established, and the system loss is calculated;

[0071] S230: Based on the obtained loss, use an optimization algorithm to inversely update the phase arrangement of the metasurface lens. Iterate multiple rounds until the focusing efficiency is stable, and then end the optimization to obtain the phase arrangement of the metasurface lens with the optimal achromatic effect;

[0072] S240: Select appropriate micro - element structures from the database for spatial arrangement to precisely match the required spatial phase distribution of the metasurface lens.

[0073] S300: Use micro - nano processing technology to fabricate and prepare the metasurface lens, connect the fabricated metasurface lens with an infrared detector to obtain an infrared imaging device, as Figure 3 shown in the layout of the metasurface lens;

[0074] S400: The infrared imaging device takes pictures to establish a dataset. Based on the dataset, use the YOLOv4 framework to establish a leakage detection model, and integrate the model and the infrared imaging device to construct a complete infrared monitoring system for real - time monitoring of facilities.

[0075] Selecting YOLOv4 can achieve a double breakthrough in accuracy and speed through innovations such as the CSPDarknet53 backbone network, Mish activation function, SPP - PANet feature fusion, Mosaic data augmentation, and CIoU loss function. Its modular design makes it show strong adaptability and practicality in fields such as industry, agriculture, and medical care, and it has become one of the benchmark models for real - time object detection.

[0076] Example Two

[0077] Use micro - nano processing technology to fabricate the designed metasurface lens. After obtaining the sample, integrate and package the sample with a movable infrared detector. As Figures 4 - 7 shown, an infrared detection device based on a metasurface lens includes a metasurface lens, an infrared detector, and a connection device (or packaging device) for connecting the metasurface lens and the infrared detector.

[0078] The connecting device includes a front end part and a rear end part. The front end part is used to place the metasurface lens, and after placing the lens in the groove of the front end part, it can be stabilized by a cover. The rear end part is used to place the infrared detector. The two parts are connected by threads, and the distance between the metasurface lens and the infrared detector can be adjusted by rotating the threaded device, so that it can focus better to obtain a clearer imaging effect, and the changing infrared electromagnetic waves emitted by the device due to leakage and breakage faults can act on the infrared detector to present a better-quality and clearer image.

[0079] To ensure that the imaged image is clear enough, the thickness of the two parts of the device is determined by the focal length f. When designing the overall length of the connecting device, the focal length of the infrared lens should be considered to avoid the situation where the length of the device is too long and the infrared detector cannot obtain a clear thermal image.

[0080] Embodiment 3

[0081] After connecting the metasurface lens and the infrared detector, build a YOLOv4 pipeline leakage detection model, and the steps are as follows:

[0082] S500: Take pictures to build a data set. When in use, first find the pipeline leakage point. The different infrared radiation changes generated by the liquid at the leakage point in the surrounding environment can simultaneously determine the position and the size of the leakage range of the liquid, that is, determine the liquid leakage area; use the above device to take pictures of the liquid leakage area image, and rotate the thread to adjust the lens position during the shooting process to obtain the clearest image. Take pictures of different angles, different scenes and different leakage amounts to obtain a batch of infrared pipeline leakage data, and use annotation tools to annotate the pipeline leakage areas in this batch of data to obtain a labeled infrared image data set of pipeline leakage.

[0083] S600: Enhance the image data. Specifically, perform data enhancement operations such as translation, rotation, cutting, and scaling on the initial data set to obtain an enlarged data set. In the obtained infrared image data set, the label is used to correspondingly mark the pipeline leakage area in the image.

[0084] Then improve the detection accuracy. Specifically, divide the enhanced data set according to the ratio of 6:2:2 to obtain a training set, a validation set and a test set; use the clustering algorithm to calculate the width and height of the damaged area of the device in the current data set, so as to adjust and determine the size of the detection frame, thereby improving the detection accuracy.

[0085] Specifically: S700: Build a YOLOv4 model framework, and introduce an attention mechanism module to strengthen the attention to the areas where the infrared electromagnetic waves change due to leakage, breakage and other faults of the device, to obtain a detection model of the faulty device to be trained;

[0086] By training to determine the feature fusion weights at different scales, the model pays more attention to important features, extracts features of leakage areas more accurately, improves the fine-grained detection effect of the model, and enhances the ability of the model to detect pipeline leaks. On the basis of the above adjustments, a pipeline leak detection model is trained.

[0087] And using the obtained training set, validation set and test set, the constructed detection model is trained, validated and tested to obtain a trained real-time monitoring model for industrial equipment.

[0088] S800: The infrared monitoring device composed of an infrared imaging system and a monitoring model is used to monitor industrial equipment: The designed infrared detection device and target detection algorithm are used to detect the pipeline in real time, and it is judged whether it is a real result based on the monitoring results of two consecutive frames. If the images obtained in two consecutive frames are both judged to be leaking, it is considered that there is a pipeline leak in the detected area. Figures 8 - 9 The figure shows an example diagram of using the established model to detect pipeline leaks with a commercial infrared device.

[0089] Finally, when problems such as leakage, breakage, and explosion occur, the area is timely feedback to the staff.

[0090] In this embodiment, aiming at the characteristics of infrared radiation parameters generated when industrial equipment fails such as leakage and breakage, a metasurface lens working in the corresponding infrared band is specially designed, and after obtaining the infrared image, it is sent into an image recognition model constructed with the YOLOv4 model framework for detection to obtain the specific fault location.

[0091] It aims to overcome the limitations of the existing infrared thermal imaging system, such as large volume and heavy weight. Among them, the infrared detection device (or infrared imaging device) cleverly integrates the advanced design concept of the infrared metasurface lens and the YOLOv4 target detection algorithm, and carefully optimizes and designs an infrared detection device that is lightweight, highly integrated and has a precisely matched working band. This device aims at the specific infrared radiation characteristics released by industrial facilities due to faults such as breakage, leakage and explosion. Through the precisely designed infrared metasurface lens, it realizes the efficient capture and focusing of infrared signals, and effectively solves the detection problems caused by obstacles or insufficient illumination in traditional detection methods. Its unique structural design endows the device with excellent concealment and strong anti-interference ability, ensuring stable and reliable operation in various complex environments. Compared with the traditional infrared optical system that relies on thick curved lenses, the infrared metasurface lens adopted in the present invention has significant advantages such as compact structure, light weight and simple processing, and can replace complex lens groups with a single lens unit to achieve the same or even better optical performance. In addition, this device is particularly suitable for extreme environments such as high temperature and high radiation, providing a new, miniaturized and high-performance thermal infrared imaging solution for the field of industrial equipment fault detection in nuclear environments.

[0092] In summary, this embodiment demonstrates important value in aspects such as technological innovation, performance improvement, and practical application, laying a solid foundation for the research of thermal infrared imaging leak detection equipment in the nuclear facility environment.

[0093] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A meta-lens design method, characterized in that: The steps include: S100: According to the wavelength range of infrared electromagnetic waves radiated by industrial equipment due to damage or leakage, the working band of the meta-lens is determined and a micro-element database is constructed according to the working band; S200: Calculate and optimize to obtain the phase distribution of the meta-lens, and use the actual micro-element in the database to match it to obtain the required meta-lens; S300: using micro-nano processing technology to process and prepare a meta-lens, and connecting the prepared meta-lens to an infrared detector to obtain an infrared imaging device; S400: Infrared imaging equipment is used to take pictures to establish a data set. Based on the data set, the YOLOv4 framework is used to establish a leak detection model. The model and infrared imaging equipment are then integrated to build a complete infrared monitoring system to monitor the facilities in real time.

2. A meta-lens design method as claimed in claim 1, characterized in that: Step S100 includes: S110: determining the working band according to the change of infrared electromagnetic waves at the damaged or leaked nuclear equipment, and selecting a plurality of discretely distributed wavelengths from the working band; S120: Select appropriate materials as a substrate and a micro-element structure, and calculate the phase and transmittance changes of the scattering field of the micro-element structure under different structural parameters at the selected wavelength to obtain a micro-element database.

3. A meta-lens design method according to claim 1, characterized in that: Step S200 includes: S210: In the incident light field U in After (x, y) passes through the meta-lens, the actual light field complex amplitude U of the focal plane can be obtained out (x, y) distribution calculation formula: in is the phase modulation of the incident light by the meta-lens; F.F -1 represents Fourier transform; f is the focal length of the lens; λ is the wavelength of the incident electromagnetic wave; j is the imaginary unit; μ and ν are the corresponding frequency domain coordinates in Fourier transform; S220: Get the focused plane light field U out (x, y), and then according to the actual light field complex amplitude distribution of the focusing plane and the ideal light field complex amplitude U ideal The difference between the (x, y) distribution, the difference between the actual adjustable phase of the microelement in the database and the phase arrangement of the simulated meta-lens is used to establish a loss function, and the system loss is calculated; S230: Based on the obtained loss, the meta-lens phase arrangement is reversely updated using an optimization algorithm, and the optimization is terminated after multiple iterations until the focusing efficiency is stable, thereby obtaining a meta-lens phase arrangement with the best achromatic effect; S240: Select appropriate micro-element structures from the database for spatial arrangement to accurately match the spatial phase distribution required by the meta-lens.

4. An infrared detection device based on a meta-lens, characterized in that: It comprises a meta-lens and an infrared detector, and a connecting device for connecting the meta-lens and the infrared detector, wherein the meta-lens is designed by the method according to any one of claims 1 to 3; The connecting device comprises a front end part and a rear end part, wherein the meta-lens and the infrared detector are placed in the front end part and the rear end part respectively, and the two parts are connected by threads.

5. The infrared detection device based on metalens according to claim 4, characterized in that: The front end portion and the rear end portion are threadedly connected via a threaded rotating device to adjust the distance between the meta-lens and the infrared detector, so that the changed infrared electromagnetic waves emitted by the device due to leakage or damage can act on the infrared detector to present better quality and clearer imaging. The overall length of the connecting device should be designed taking into account the focal length of the infrared lens to avoid the situation where the device is too long and the infrared detector cannot obtain a clear thermal image.

6. A method for using the infrared detection device of the metalens according to any one of claims 4 to 5, characterized in that: Here are the steps: S500: Take pictures of equipment damage and leakage, and label the faulty area of ​​the equipment to obtain a labeled infrared image dataset. S600: enhancing the image data, and dividing the enhanced data set into a ratio of 6:2:2 to obtain a training set, a validation set, and a test set; The clustering algorithm is used to calculate the width and height of the damaged area of ​​the device in the current data set, thereby determining the size of the detection frame to improve the detection accuracy; S700: Build the YOLOv4 model framework and introduce the attention mechanism module to strengthen the attention to the area where the infrared electromagnetic wave changes due to leakage, damage and other faults of the equipment, and obtain the fault equipment detection model to be trained; The obtained training set, validation set and test set are used to train, validate and test the constructed detection model to obtain a trained real-time monitoring model for industrial equipment. S800: The infrared monitoring device consisting of an infrared imaging system and a monitoring model is used to monitor industrial equipment and provide timely feedback to staff in the event of leakage, damage, explosion or other problems.

7. The method for using the infrared detection device of a metalens according to claim 6, characterized in that: The image data enhancement method in step S600 is: The initial data set is subjected to data enhancement operations such as translation, rotation, cutting, and scaling to obtain an enhanced complete data set.

Images