Image recognition and correction system based on optical model and environmental perception

Abstract

This invention discloses an image recognition and correction system based on an optical model and environmental perception, belonging to the field of image processing technology. The system includes an environmental perception module: This module acquires multi-medium refractive index, interface 3D morphology, and initial image data through a refractive index sensor, a 3D laser scanner, and a vision camera. This data is then dynamically corrected and fused from multiple sources to form an environmental perception dataset. The invention constructs an optical model based on the environmental perception data and combines it with an interference compensation sub-model to calculate the light propagation path. This interference compensation sub-model can quantify the interference effects of impurities, temperature gradients, and other interference factors in the medium by analyzing relevant data acquired by the environmental perception module, generating compensation terms that are integrated into the light propagation equation. This corrects the light propagation path, ensuring the accuracy of the light propagation path calculation and providing solid and reliable data support for subsequent image recognition and correction, thus improving the accuracy and quality of image recognition and correction.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, specifically to an image recognition and correction system based on optical models and environmental perception. Background Technology

[0002] In the field of image processing technology, accurate image acquisition and processing are fundamental to many applications, playing a crucial role in subsequent image recognition, analysis, and decision-making. However, in real-world environments, light propagation is often affected by the medium. When light passes through interfaces of different media, refraction occurs at these interfaces due to the different refractive indices of the media. This refraction leads to distortion in the acquired image, causing deviations in the shape and position information of the target object, thus severely impacting the accuracy and reliability of subsequent image recognition and analysis processes.

[0003] Existing image recognition and correction methods have significant shortcomings. They fail to fully consider the complex propagation characteristics of light in multi-media environments. Most of these methods only address single-media or simple imaging distortions, lacking effective solutions to the technical problems caused by refraction at multi-media interfaces. In a single-media environment, light propagation is relatively simple, and traditional image correction methods may adjust based on some simplified optical models or fixed distortion parameters. However, in a multi-media environment, the light propagation path is complex and variable, and optical phenomena such as refraction, reflection, and scattering at different media interfaces are intertwined, making it difficult for traditional methods to accurately simulate the light propagation process and effectively eliminate image distortion caused by multi-media refraction. This makes it difficult to meet the application requirements of high-precision image recognition. For example, in image acquisition scenarios involving multi-media environments such as underwater-air interfaces and glass-air interfaces, images processed by traditional methods still have obvious morphological and positional deviations, failing to provide accurate and reliable data support for subsequent image recognition. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide an image recognition and correction system based on optical models and environmental perception. This system can acquire key parameters of a multi-media environment through an environmental perception module, including the refractive index of different media, the three-dimensional morphology of the multi-media interface, and environmental texture, forming an environmental perception dataset. Based on this data, a multi-media light field refraction model construction module builds an adapted light field refraction model, adjusting the dimensions according to the media type and incorporating interface morphology parameters. Ray tracing is used to simulate the light propagation process. The light propagation path calculation module, based on the environmental perception dataset and the multi-media light field refraction model, uses a numerical discretization method to solve for the light propagation path. Furthermore, an interference compensation sub-model is established to correct interference factors such as impurities and temperature gradients in the medium. The target object light field reconstruction module reconstructs the target object light field distribution based on the light propagation path calculation results, and optimizes it through signal optimization processing methods. Finally, the image recognition and correction module uses a deep learning-based recognition model to recognize and process the reconstructed image, and corrects the original distorted image by combining the reconstructed image. The correction results are verified through relevant indicators, which improves the system's adaptability to multi-media environments, ensures the accuracy of light propagation path calculation, provides reliable data support for image recognition and correction, and has wide application adaptability.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an image recognition and correction system based on optical models and environmental perception, the system comprising:

[0006] Environmental perception module: Collects multi-media refractive index, interface 3D morphology and initial image data through refractive index sensor, 3D laser scanner and vision camera, and forms environmental perception dataset through dynamic correction and multi-source fusion;

[0007] Multi-medium light field refraction model construction module: Based on the environmental perception dataset, an adaptive model is constructed by combining Maxwell's equations and Fresnel's law. The model simulates the propagation direction, intensity changes and random optical phenomena of light through ray tracing, and describes the propagation law of light in multiple media.

[0008] Light propagation path calculation module: Based on the environmental perception dataset and the light field refraction model, it uses a numerical discretization method to solve the light propagation trajectory, and combines the interference compensation sub-model to correct the effects of impurities and temperature gradients, so that the path calculation is accurate and continuous.

[0009] Target object light field reconstruction module: By tracing the light path in reverse, integrating light parameters to construct the target's three-dimensional light field distribution, optimizing the light field data through energy attenuation and scattering interference compensation, and converting it into a standard image format;

[0010] Image recognition and correction module: It uses a deep learning-based recognition model to identify targets in standard light field images. At the same time, it combines the deviation between the standard image and the original distorted image, eliminates refraction distortion through geometric transformation and grayscale correction, and verifies and outputs the correction results.

[0011] Furthermore, the environmental perception module includes a refractive index sensor, a 3D laser scanner, and a vision camera, used to collect basic data of the multi-media environment. The refractive index sensor, based on the principles of interference or surface plasmon resonance, measures the refractive index of different media in real time and transmits the measurement data. It can also correct the measurement results according to changes in ambient temperature and pressure. The 3D laser scanner scans the multi-media interface to obtain the 3D point cloud data of the interface. By performing denoising, registration, and modeling operations on the 3D point cloud data, a 3D morphological model containing interface shape, curvature, and flatness parameters is generated. The vision camera collects environmental texture and initial image information of target objects, and integrates it with the data obtained by the refractive index sensor and the 3D laser scanner to form an environmental perception dataset.

[0012] Furthermore, the multi-medium optical field refraction model construction module constructs a multi-medium optical field refraction model based on the acquired media refractive index data and interface morphology data, combined with Maxwell's equations and Fresnel's law. The model dimensions are adjusted according to the type of medium, and the interface morphology parameters are incorporated as boundary conditions into the model construction process. At the same time, ray tracing is used to simulate the light propagation process within the model. The light is discretized into several light segments, and the propagation direction of each light segment in different media is calculated based on the multi-medium interface light refraction direction formula. The light intensity change is calculated using the light propagation intensity attenuation formula, and a probability model of light reflection, refraction, and scattering at the interface is established to describe the random optical phenomena in the light propagation process through probabilistic simulation.

[0013] Furthermore, the multi-medium light field refraction model construction module calculates the propagation direction of each light segment in different media based on the multi-medium interface light refraction direction formula. The calculation formula is as follows: ,in It is the angle of refraction of light in the second medium. It is the refractive index of the first medium, extracted from the environmental perception dataset. It is the refractive index of the second medium, extracted from the environmental perception dataset. It is the angle of incidence of light in the first medium. It is the distance from the point of incidence of the light ray to the center of the curvature of the interface. It is the radius of curvature of the medium interface. This is the interface curvature correction term, and its calculation formula is: When the interface is flat The change in light intensity is calculated using the light propagation intensity attenuation formula, which is as follows: ,in, It is the final intensity of light after propagation. It is the initial intensity of the light. It is the number of media layers that light passes through. It is the first The absorption coefficient of the layer medium, It is the first The scattering coefficient of the layer medium, Is the light in the first... Propagation distance in layered media It is the first The reflectivity of a medium interface is calculated using Fresnel's law and the refractive index of the medium. The formula is as follows: ,in , The first The refractive index of the medium on both sides of the interface.

[0014] Furthermore, the light propagation path calculation module takes the environmental perception dataset as input and the multi-medium light field refraction model as its basis. It uses a numerical discretization method to solve for the light propagation path. Specifically, the multi-medium space is divided into tiny grid cells with a preset precision. The light propagation equation is discretized within each grid cell. After substituting the refractive index of the medium and the interface morphology parameters, the propagation trajectory of the light in the cell is determined by iterative calculation using the discretization formula of the light propagation trajectory in the multi-medium space. When the light crosses the medium interface, the refraction parameters in the multi-medium light field refraction model are called to update the light propagation direction and speed. At the same time, an interference compensation sub-model is established for interference factors such as impurities and temperature gradients in the medium. By analyzing the impurity distribution data and temperature field data obtained by the environmental perception module, the influence of interference factors on light propagation is quantified, a compensation term is generated and integrated into the light propagation equation, and the light propagation path is corrected.

[0015] Furthermore, the light propagation path calculation module determines the light propagation trajectory within the unit through iterative calculation using the discretization formula for the light propagation trajectory in multi-media space. The formula is as follows: ,in, Is the light in the first... Position vectors of discrete units Is the light in the first... Position vectors of discrete units Is the light in the first... The propagation speed of each discrete unit in the medium is calculated using the following formula: ,in The speed of light in a vacuum The refractive index of the medium in which this unit is located. It is the discrete time step. Is the light in the first... The propagation direction unit vector in a discrete unit, This is the vector of light position shift caused by interference factors; simultaneously, an interference compensation sub-model is established to address interference factors such as impurities and temperature gradients in the medium, and its model formula is: ,in, It is the vector of light position shift caused by interference. It is the unit vector of the interference offset direction. It is the impurity interference coefficient. It is the concentration of impurities in the medium. It is the temperature gradient disturbance coefficient. It is the temperature gradient of the medium. It is the distance that light travels within the current medium unit.

[0016] Furthermore, the target object light field reconstruction module, based on the calculation results of the light propagation path, adopts a reverse ray tracing method starting from the imaging plane of the image acquisition device and tracing the propagation path of each imaging ray in reverse until it reaches the surface of the target object. By integrating the coordinates, superimposing the intensity, and calibrating the phase of the reverse-traced rays, a three-dimensional light field distribution of the target object containing light field intensity, phase, and direction parameters is constructed. At the same time, in response to energy attenuation and scattering interference during the light propagation process, a signal optimization method is used to optimize the reconstructed three-dimensional light field distribution. By constructing an attenuation function and a scattering kernel, the light energy loss is compensated and the phase deviation is corrected. The light field data is then denoised and converted into a standard image data format.

[0017] Furthermore, the target object light field reconstruction module compensates for light energy loss and corrects phase deviation by constructing an attenuation function and a scattering kernel. The phase deviation correction formula is as follows: ,in, It is the corrected phase of the optical field. It is the raw phase value extracted from the light field data. It is the number of media layers that light rays pass through from the target object to the imaging plane. It is the first The refractive index of the layered medium, It refers to the wavelength of light, preset to the center wavelength of visible light. Is the light in the first... Propagation distance in layered media It is the first Phase deviation caused by scattering from the layer medium.

[0018] Furthermore, the image recognition and correction module employs a deep learning-based recognition model to process the standard-format image data output by the target object light field reconstruction module. It is trained on a dataset containing multi-media environmental distortion images and corrected images. By extracting target object features and achieving target classification, localization, and attribute recognition, the module outputs the recognition results. Simultaneously, it corrects the original distorted image by combining the generated standard-format image data. Image registration aligns the original distorted image with the standard-format image, calculating their geometric and grayscale deviations. Geometric transformation and grayscale correction are used to adjust the geometric parameters and pixel values ​​of the original distorted image. After correction, the results are verified using peak signal-to-noise ratio and structural similarity index.

[0019] Compared with existing technologies, this image recognition and correction system based on optical models and environmental perception has the following advantages:

[0020] I. This invention constructs an optical model based on environmental perception data and combines it with an interference compensation sub-model to calculate the light propagation path. The interference compensation sub-model can quantify the interference effects of impurities, temperature gradients, and other interference factors in the medium by analyzing relevant data obtained by the environmental perception module, generating compensation terms that are incorporated into the light propagation equation, thereby correcting the light propagation path. This ensures the accuracy of the light propagation path calculation and provides solid and reliable data support for subsequent image recognition and correction, helping to improve the accuracy and quality of image recognition and correction.

[0021] Second, this invention prioritizes the acquisition of multi-media environmental data through an environmental perception module, and constructs an adapted multi-media light field refraction model based on this data. It fully considers the complexity and diversity of actual multi-media environments, avoids the problem of model decoupling from actual environment in traditional methods, and enables the system to better adapt to different multi-media scenarios.

[0022] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0024] Figure 1This is a block diagram of an image recognition and correction system based on optical models and environmental perception.

[0025] Figure 2 This is a flowchart of an image recognition and correction system based on optical models and environmental perception. Detailed Implementation

[0026] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0027] This invention provides an image recognition and correction system based on optical models and environmental perception, such as... Figure 1 As shown, the system includes an environment perception module, a multi-medium light field refraction model construction module, a light propagation path calculation module, a target object light field reconstruction module, and an image recognition and correction module. The environment perception module acquires key parameters of the multi-medium environment, including the refractive index of different media, the 3D morphology of the multi-medium interface, and environmental texture, forming an environment perception dataset. Based on this data, the multi-medium light field refraction model construction module builds an adapted light field refraction model, adjusting the dimensions according to the media type and incorporating interface morphology parameters. It then uses ray tracing to simulate the light propagation process. The light propagation path calculation module, based on the environment perception dataset and the multi-medium light field refraction model, uses a numerical discretization method to solve for the light propagation path. The system calculates the light propagation path and establishes an interference compensation sub-model to correct for interference factors such as impurities and temperature gradients in the medium. The target object light field reconstruction module reconstructs the light field distribution of the target object based on the light propagation path calculation results using inverse ray tracing and optimizes it through signal optimization processing methods. Finally, the image recognition and correction module uses a deep learning-based recognition model to recognize and process the reconstructed image, and corrects the original distorted image by combining the reconstructed image. The correction results are verified through relevant indicators, which improves the system's adaptability to multi-media environments, ensures the accuracy of light propagation path calculation, provides reliable data support for image recognition and correction, and has wide application adaptability.

[0028] Example 1

[0029] This embodiment is applied to the testing of precision equipment protected by glass covers in industrial production workshops, solving the problem of image distortion caused by refraction at the air-glass-air three-medium interface.

[0030] like Figure 2As shown, the environmental perception module is activated, and the refractive index sensor is calibrated to adapt to the measurement requirements of both air and glass media, ensuring measurement accuracy. The scanning resolution of the 3D laser scanner is configured to be 0.05mm, and the scanning range covers the glass cover and the internal equipment area. The focal length of the vision camera is adjusted to clearly capture equipment details, and the exposure time is set to 1 / 200s to avoid interference from workshop light. Standard refractive index data of air and industrial glass are loaded, the core processing unit parameters are initialized, the mesh division accuracy is set to 0.3mm, the reverse ray tracing iteration termination condition is error ≤0.008mm, and the recognition model inference threshold is 0.85.

[0031] A refractive index sensor measures the refractive index of the workshop air and the glass enclosure at a frequency of 0.5 Hz, generating a dual-medium refractive index distribution map. A 3D laser scanner scans the inner and outer surfaces of the glass enclosure (the air-glass and glass-air interfaces) to obtain 3D point cloud data of the interfaces. After processing, a 3D morphological model containing interface curvature and flatness is generated. A vision camera acquires initial images of the equipment inside the glass enclosure at a frame rate of 25 fps, extracts features such as surface texture and contour of the equipment, and integrates the three types of data through data fusion processing to form an environmental perception dataset and store it.

[0032] The system reads environmental perception datasets from its database, extracts refractive index distribution data for air and glass, and morphological parameters of the glass cover's inner and outer interfaces. Combining Maxwell's equations and Fresnel's law, a light field refraction model adapted to the air-glass-air three-medium environment is constructed. The model dimensions are adjusted according to the three-medium structure, and the morphological parameters of the glass cover's inner and outer interfaces are incorporated as boundary conditions into the model construction process. Initial light parameters (simulating the light emission angle and initial intensity of a visual camera lens) are input, and ray tracing is used to simulate the propagation of light in the three media: based on the formula for the direction of light refraction at multi-medium interfaces. Calculate the angle of refraction of light at the air-glass and glass-air interfaces to determine the direction of light propagation; use the formula for light intensity attenuation. ( The final intensity of light after propagation. The initial intensity of the light. The number of media layers that light passes through. For the first The absorption coefficient of the layer medium, It is the first The scattering coefficient of the layer medium, For the light in the first Propagation distance in layered media For the first The intensity variation of light propagating in each medium is calculated by using the reflectivity of the interface. At the same time, a probability model of light reflection, refraction and scattering at the interface is established. The random optical phenomena in the light propagation process are described by probabilistic simulation. A simulation dataset containing the coordinates of light segments, propagation direction and intensity is generated to verify the adaptability of the model to the current industrial testing environment.

[0033] The environmental perception dataset and the constructed three-medium light field refraction model are imported into the light propagation path calculation module. A numerical discretization method is used to divide the air-glass-air multi-medium space into tiny grid cells with a preset precision. Within each grid cell, the light propagation equation is discretized. After substituting the media refractive index and interface morphology parameters, iterative calculations determine the light propagation trajectory within the cell. This process is based on the discretization formula for the light propagation trajectory in the multi-medium space. , The light rays at the 1st , Position vectors of discrete units For the light in the first The propagation speed of each discrete unit in the medium For discrete time steps, For the light in the first The propagation direction unit vector in a discrete unit, The system updates the light position (to account for the light position offset vector caused by interference factors). When the light passes through the air-glass and glass-air interfaces, it calls the refraction parameters in the multi-medium light field refraction model to update the light propagation direction and speed in real time. Simultaneously, considering the interference factors of the distribution of minute impurities inside the glass and the workshop temperature gradient, it analyzes the impurity distribution data and temperature field data obtained by the environmental sensing module, and applies the offset formula of the interference compensation sub-model. ( The unit vector representing the direction of interference offset. The impurity interference coefficient is... The concentration of impurities in the medium. The temperature gradient disturbance coefficient is... For the medium temperature gradient, The influence of interference factors on light propagation (the distance of light propagation in the current medium unit) is quantified, a compensation term is generated and incorporated into the light propagation equation, the light propagation path is corrected, and through iterative calculation, the complete light propagation path from the device surface to the camera imaging plane is obtained, and the coordinate trajectory and optical parameters of each light ray are output.

[0034] The reverse ray tracing method starts from the imaging plane of the image acquisition device and traces the propagation path of each imaging ray backward until it reaches the surface of the target object. By integrating the coordinates, superimposing the intensity, and calibrating the phase of the traced rays, a three-dimensional light field distribution of the target object is constructed, including light field intensity, phase, and direction parameters. Simultaneously, to address energy attenuation and scattering interference during light propagation, a signal optimization method is used to optimize the reconstructed three-dimensional light field distribution. By constructing an attenuation function and a scattering kernel, energy loss is compensated and phase deviation is corrected. The phase deviation correction formula is as follows: ,in, It is the corrected phase of the optical field. It is the raw phase value extracted from the light field data. It is the number of media layers that light rays pass through from the target object to the imaging plane. It is the first The refractive index of the layered medium, It refers to the wavelength of light, preset to the center wavelength of visible light. Is the light in the first... Propagation distance in layered media It is the first The phase deviation caused by scattering from the layer medium is also addressed by denoising the light field data, converting the optimized light field data into a standard image data format, and clearly presenting the details of the device surface.

[0035] The standard image is input into the recognition model to identify defects such as scratches (2mm in length) and local deformation on the equipment surface, and output the defect category, location and confidence level; the original distorted image (blurred defect edges and positional offset) is registered and corrected, and the distortion is eliminated by geometric transformation and grayscale adjustment. After verification by PSNR (35dB) and SSIM (0.95), the corrected image and defect recognition results are stored to complete the industrial equipment inspection and correction.

[0036] Example 2

[0037] This embodiment is applied to the underwater biological detection scenario in freshwater lakes. It is necessary to solve the image distortion problem caused by refraction at the air-water dual-medium interface and achieve accurate identification and imaging correction of underwater fish targets.

[0038] First, the environmental perception module is activated to calibrate the refractive index sensor, ensuring that its measurement accuracy error in freshwater environment is less than 0.0001. The scanning resolution of the 3D laser scanner is configured to be 0.1mm, and the scanning range covers an underwater area of ​​1-5m. The focal length of the vision camera is adjusted to the optimal value for underwater imaging, and the exposure time is set to 1 / 100s with the white balance set to underwater mode. At the same time, standard refractive index data of freshwater at 25℃ and standard atmospheric pressure is loaded, the parameters of the system's core processing unit are initialized, the multi-media spatial grid division accuracy is set to 0.5mm, the iteration termination condition for reverse ray tracing is ray tracing error ≤0.01mm, and the inference confidence threshold of the image recognition model is 0.8.

[0039] A refractive index sensor continuously measures the refractive index of freshwater in the detection area at a frequency of 1 Hz, generating a refractive index distribution map. Simultaneously, the measured values ​​are dynamically corrected based on the lake's on-site temperature (22℃) and atmospheric pressure (100.8 kPa). A 3D laser scanner performs a full-range scan of the lake surface (air-water interface), acquiring 3D point cloud data of the interface. After denoising and registration processing, a 3D morphological model of the lake surface is generated. A vision camera acquires initial images of the underwater environment and fish at a frame rate of 30 fps, extracting initial features such as texture and color from the images. Through data fusion processing, the corrected refractive index data, the lake surface morphological model, and the initial image data are integrated to form an environmental perception dataset, which is then stored in the system database.

[0040] The system reads environmental perception datasets from its database, extracts freshwater refractive index distribution data and lake surface morphology parameters, and constructs an air-water dual-medium light field refraction model by combining Maxwell's equations and Fresnel's law. Initial ray parameters (simulating the light emission angle and initial intensity of a visual camera lens) are input, and ray tracing is used to simulate the propagation process of light in the air-water medium, generating a simulation dataset containing ray segment coordinates, propagation direction, and intensity to verify the model's adaptability to the actual underwater environment.

[0041] The environmental perception dataset and the constructed dual-medium light field refraction model are imported into the light propagation path calculation module. The air-water space is divided into tiny grid cells using a numerical discretization method, and the light propagation equation is solved within each cell. When light travels from the air to the water medium interface, the refraction parameters in the model are called to update the propagation direction and velocity. At the same time, for the distribution of tiny impurities (such as plankton) in the water, the influence of impurities on light propagation is quantified by analyzing the environmental perception data, and a compensation term is generated and incorporated into the light propagation equation to correct the light trajectory. Through iterative calculation, the complete light propagation path from the underwater fish target to the camera imaging plane is obtained, and the coordinate trajectory and optical parameters of each light ray are output.

[0042] Based on the light propagation path data, a reverse ray tracing method is used to trace the light rays from each pixel on the camera's imaging plane back to the surface of the underwater fish, recording the intensity and phase information of each tracing ray. All tracing rays are integrated to construct a three-dimensional light field matrix of the fish target. To address the energy attenuation of light propagating in water and interference from impurities, a signal optimization processing method is used to optimize the three-dimensional light field matrix, compensating for energy loss and correcting phase deviation. At the same time, the light field data is denoised, and the optimized light field data is converted into a standard RGB format image.

[0043] The standard image is input into the trained deep learning-based recognition model. The model extracts features and outputs the fish category (e.g., crucian carp, carp), location coordinates, and confidence score. If the confidence score of a certain fish target is 0.75 (below the preset threshold of 0.8), the model returns to the light field reconstruction module to adjust the denoising parameters and regenerate the standard image. After the second recognition, the confidence score increases to 0.86. The image registration process is then called to align the original underwater distorted image (with fish body stretching and blurred edges) with the standard image. The geometric deviation and grayscale deviation between the two are calculated. Geometric transformation is used to correct the fish body stretching distortion, and grayscale mapping is used to adjust the image brightness deviation to generate a clear image after correction. Finally, the correction effect is verified by PSNR (32dB) and SSIM (0.92). The corrected image and recognition results are stored in the system database, completing the underwater target detection and correction.

[0044] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. An image recognition and correction system based on optical models and environmental perception, characterized in that, The system includes: Environmental perception module: Collects multi-media refractive index, interface 3D morphology and initial image data through refractive index sensor, 3D laser scanner and vision camera, and forms environmental perception dataset through dynamic correction and multi-source fusion; Multi-medium optical field refraction model construction module: Based on the environmental perception dataset, a multi-medium optical field refraction model is constructed by combining Maxwell's equations and Fresnel's law. The model simulates the propagation direction, intensity changes and random optical phenomena of light through ray tracing, and describes the propagation law of light in multiple media. Light propagation path calculation module: Based on the environmental perception dataset and the multi-medium light field refraction model, it uses a numerical discretization method to solve the light propagation path, and combines the interference compensation sub-model to correct the effects of impurities and temperature gradients, so that the path calculation is accurate and continuous. Target object light field reconstruction module: By tracing the light path in reverse, integrating light parameters to construct the target's three-dimensional light field distribution, optimizing the light field data through energy attenuation and scattering interference compensation, and converting it into a standard image data format; Image recognition and correction module: It uses a deep learning-based recognition model to identify targets in standard image data. At the same time, it combines the deviation between the standard image and the original distorted image, eliminates refractive distortion through geometric transformation and grayscale correction, and verifies and outputs the correction results.

2. The image recognition and correction system based on optical model and environmental perception according to claim 1, characterized in that, The environmental perception module includes a refractive index sensor, a 3D laser scanner, and a vision camera. It is used to collect basic data of the multi-media environment. The refractive index sensor, based on the principles of interference or surface plasmon resonance, measures the refractive index of different media in real time and transmits the measurement data. It can also correct the measurement results according to changes in ambient temperature and pressure. The 3D laser scanner scans the multi-media interface to obtain the 3D point cloud data of the interface. By performing denoising, registration, and modeling operations on the 3D point cloud data, a 3D morphological model containing interface shape, curvature, and flatness parameters is generated. The vision camera collects environmental texture and initial image data of target objects and integrates them with the data obtained by the refractive index sensor and the 3D laser scanner to form an environmental perception dataset.

3. The image recognition and correction system based on optical model and environmental perception according to claim 1, characterized in that, The multi-medium optical field refraction model construction module constructs a multi-medium optical field refraction model based on the acquired media refractive index data and interface morphology data, combined with Maxwell's equations and Fresnel's law. The model dimensions are adjusted according to the type of medium, and interface morphology parameters are incorporated as boundary conditions into the model construction process. At the same time, ray tracing is used to simulate the light propagation process within the model. The light rays are discretized into several light segments, and the propagation direction of each light segment in different media is calculated based on the multi-medium interface light refraction direction formula. The light intensity change is calculated using the light propagation intensity attenuation formula, and a probability model of light reflection, refraction, and scattering at the interface is established to describe the random optical phenomena in the light propagation process through probabilistic simulation.

4. The image recognition and correction system based on optical model and environmental perception according to claim 3, characterized in that, The multi-medium optical field refraction model construction module calculates the propagation direction of each light segment in different media based on the multi-medium interface light refraction direction formula. The calculation formula is as follows: ,in It is the angle of refraction of light in the second medium. It is the refractive index of the first medium, extracted from the environmental perception dataset. It is the refractive index of the second medium, extracted from the environmental perception dataset. It is the angle of incidence of light in the first medium. It is the distance from the point of incidence of the light ray to the center of the curvature of the interface. It is the radius of curvature of the medium interface. This is the interface curvature correction term, and its calculation formula is: When the interface is flat ; The change in light intensity is calculated using the formula for light intensity attenuation during propagation. The formula is as follows: ,in, It is the final intensity of light after propagation. It is the initial intensity of the light. It is the number of media layers that light passes through. It is the first The absorption coefficient of the layer medium, It is the first The scattering coefficient of the layer medium, Is the light in the first... Propagation distance in layered media It is the first The reflectivity of a medium interface is calculated using Fresnel's law and the refractive index of the medium. The formula is as follows: ,in , The first The refractive index of the medium on both sides of the interface.

5. The image recognition and correction system based on optical model and environmental perception according to claim 1, characterized in that, The light propagation path calculation module takes the environmental perception dataset as input and the multi-medium light field refraction model as its basis. It uses a numerical discretization method to solve for the light propagation path. Specifically, the multi-medium space is divided into tiny grid cells with a preset precision. The light propagation equation is discretized within each grid cell. After substituting the refractive index of the medium and the interface morphology parameters, the propagation path of the light in the cell is determined by iterative calculation through the discretization formula of the light propagation trajectory in the multi-medium space. When the light crosses the medium interface, the refraction parameters in the multi-medium light field refraction model are called to update the light propagation direction and speed. At the same time, an interference compensation sub-model is established for interference factors such as impurities and temperature gradients in the medium. By analyzing the impurity distribution data and temperature field data obtained by the environmental perception module, the influence of interference factors on light propagation is quantified, compensation terms are generated and integrated into the light propagation equation to correct the light propagation path.

6. The image recognition and correction system based on optical model and environmental perception according to claim 5, characterized in that, The light propagation path calculation module determines the propagation path of light within a unit through iterative calculation using the discretization formula of the light propagation trajectory in multi-media space. The formula is as follows: ,in, Is the light in the first... Position vectors of discrete units Is the light in the first... Position vectors of discrete units Is the light in the first... The propagation speed of each discrete unit in the medium is calculated using the following formula: ,in The speed of light in a vacuum The refractive index of the medium in which this unit is located. It is the discrete time step. Is the light in the first... The propagation direction unit vector in a discrete unit, This is the vector of light position shift caused by interference factors; simultaneously, an interference compensation sub-model is established to address interference factors such as impurities and temperature gradients in the medium, and its model formula is: ,in, It is the vector of light position shift caused by interference. It is the unit vector of the interference offset direction. It is the impurity interference coefficient. It is the concentration of impurities in the medium. It is the temperature gradient disturbance coefficient. It is the temperature gradient of the medium. It is the distance that light travels within the current medium unit.

7. The image recognition and correction system based on optical model and environmental perception according to claim 1, characterized in that, The target object light field reconstruction module, based on the calculation results of the light propagation path, adopts a reverse ray tracing method starting from the imaging plane of the image acquisition device and tracing the propagation path of each imaging ray in reverse until it reaches the surface of the target object. By integrating the coordinates, superimposing the intensity, and calibrating the phase of the reverse-traced rays, a three-dimensional light field distribution of the target object containing light field intensity, phase, and direction parameters is constructed. At the same time, in response to energy attenuation and scattering interference during the light propagation process, a signal optimization method is used to optimize the reconstructed three-dimensional light field distribution. By constructing an attenuation function and a scattering kernel, the light energy loss is compensated and the phase deviation is corrected. The light field data is then denoised and converted into a standard image data format.

8. The image recognition and correction system based on optical model and environmental perception according to claim 7, characterized in that, The target object light field reconstruction module compensates for light energy loss and corrects phase deviation by constructing an attenuation function and a scattering kernel. The phase deviation correction formula is as follows: ,in, It is the corrected phase of the optical field. It is the raw phase value extracted from the light field data. It is the number of media layers that light rays pass through from the target object to the imaging plane. It is the first The refractive index of the layered medium, It refers to the wavelength of light, preset to the center wavelength of visible light. Is the light in the first... Propagation distance in layered media It is the first Phase deviation caused by scattering from the layer medium.

9. The image recognition and correction system based on optical model and environmental perception according to claim 1, characterized in that, The image recognition and correction module employs a deep learning-based recognition model to process the standard-format image data output by the target object light field reconstruction module. It is trained on a dataset containing multi-media environmental distortion images and corrected images. By extracting target object features and performing target classification, localization, and attribute recognition, it outputs the recognition results. Simultaneously, it corrects the original distorted image using the generated standard-format image data. Image registration aligns the original distorted image with the standard-format image, calculating their geometric and grayscale deviations. Geometric transformation and grayscale correction are then used to adjust the geometric parameters and pixel values ​​of the original distorted image. After correction, the results are verified using peak signal-to-noise ratio and structural similarity index.

Images