A collaborative design method and system for high-uniformity rectangular spot freeform surface lenses for tunnel lighting
By constructing a freeform surface lens with a high uniformity rectangular light spot and combining it with multi-objective optimization using microrendering technology, the problems of mismatched light spots, uneven illuminance, rigid design, and difficulty in simultaneously addressing glare in traditional tunnel lighting were solved, achieving efficient and flexible tunnel lighting design.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HARBIN INST OF TECH
- Filing Date
- 2026-01-28
- Publication Date
- 2026-06-02
AI Technical Summary
Traditional tunnel lights use rotationally symmetrical lenses, which result in a mismatch between the light spot and the rectangular road surface, low light energy utilization, poor illuminance uniformity, insufficient design flexibility, high modification costs, and difficulty in achieving both anti-glare and efficient light distribution.
A high-uniformity rectangular spot freeform lens design method is adopted. The initial surface shape of the double freeform lens is constructed by Snell's law of refraction and the law of conservation of energy. Multi-objective gradient optimization is carried out by combining differentiable rendering technology, and the weights are adaptively adjusted to optimize the illuminance uniformity, spot rectangularity and anti-glare performance.
It achieves efficient rectangular light spot distribution, improves light energy utilization, enhances illuminance uniformity, reduces design cycle and modification costs, takes into account anti-glare performance, and improves the safety and comfort of tunnel lighting.
Smart Images
Figure CN122133211A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of non-imaging optics technology, specifically relating to a collaborative design method for a freeform surface lens with a high uniformity rectangular light spot for tunnel illumination. Background Technology
[0002] Traditional LED tunnel lights mostly use axisymmetric lenses or reflectors, and their light spots are mostly circular or elliptical, which is seriously mismatched with the inherent rectangular lighting area of the tunnel surface. This geometric mismatch leads to significant light energy waste, because a considerable portion of the light will inevitably be projected onto non-target areas such as the tunnel sidewalls, rather than being fully concentrated on the surface that needs illumination, resulting in low light energy utilization and poor uniformity.
[0003] Although some studies have used freeform surfaces for light distribution, existing design methods have significant drawbacks: (1) The shape of the light spot does not match the target area: the tunnel surface is rectangular, while the rotationally symmetric optical system produces a circular or elliptical light spot, resulting in a large amount of light energy being wasted in non-target areas such as the tunnel wall, resulting in low light energy utilization.
[0004] (2) Difficulty in achieving high standards of illuminance uniformity: Due to the characteristics of the light source itself and the limitations of optical design, existing luminaires are unable to achieve high uniformity of illumination within a rectangular area. The illuminance distribution formed on the road surface is often not uniform (usually below 70%), easily producing obvious bright spots and dark areas. This uneven lighting can cause visual discomfort to drivers, requiring frequent adjustment of visual focus, thus posing a potential threat to traffic safety and becoming a key pain point in tunnel lighting design.
[0005] (3) Poor design flexibility and high modification costs: Existing luminaires are seriously lacking in design flexibility and modification costs. Their optical components (such as lens arrays) are often integrated with the structural components or manufactured using integral molds. Once the geometric parameters of the tunnel project (such as width and height) change and the light distribution angle or range needs to be adjusted, the entire optical module and even the mold must be redesigned and remanufactured. This process is not only costly but also time-consuming, and cannot meet the diverse and fast-paced needs of modern tunnel engineering.
[0006] (4) It is difficult to balance anti-glare performance with efficient light distribution: In order to effectively control glare and protect the safety of oncoming drivers, it is usually necessary to add structures such as grilles and sunshades. These additional components undoubtedly increase the size and structural complexity of the entire lighting system, while also blocking some light, resulting in a decrease in the overall light output efficiency of the system, creating a contradictory situation where it is difficult to balance efficient light distribution with a comfortable visual experience. Summary of the Invention
[0007] This invention provides a method and system for the collaborative design of freeform surface lenses with high uniformity rectangular light spots for tunnel lighting. Its purpose is to solve the systemic technical problems of traditional tunnel lighting design, which uses rotationally symmetric optical systems, resulting in circular or elliptical light spots that do not match the rectangular road surface, low illuminance uniformity, rigid design, high modification costs, and difficulty in balancing anti-glare and high efficiency.
[0008] Firstly, the present invention aims to provide a collaborative design method for a high-uniformity rectangular spot freeform surface lens for tunnel lighting, comprising the following steps: S1: Determine the geometric parameters and optical indicators of the tunnel where the lighting fixtures are located. The geometric parameters include: tunnel width, fixture installation height, and fixture spacing. The optical indicators include: average illuminance of the target area, illuminance uniformity, threshold increment, and target rectangular area size. S2: Based on Snell's law of refraction and the law of conservation of energy, the initial surface shape of the double freeform lens is constructed by solving the Monge-Ampere equation; wherein, the double freeform lens includes an incident surface and an exit surface, both of which are represented by XY polynomial surfaces, and the construction of the initial surface shape is used to form a preliminary rectangular spot distribution on the target illumination surface; S3: Based on differentiable rendering technology, establish an end-to-end differentiable optical model, perform multi-objective gradient optimization on the initial surface shape, and obtain the optimal surface shape parameters. The multi-objective gradient optimization includes defining a multi-objective loss function, calculating the gradient, and simultaneously optimizing illuminance uniformity, spot rectangularity, anti-glare performance, and manufacturing constraints through an adaptive weight adjustment mechanism.
[0009] Furthermore, a preferred solution is provided: In S2, the initial surface shape is constructed based on solving the Monge-Ampere equation, expressed as: , in, It is the surface to be determined The Hessian matrix, Represents a determinant. Indicates the light intensity distribution of the light source. This represents the actual illuminance distribution.
[0010] Furthermore, a preferred solution is provided: In S3, the differentiable rendering technology includes: representing the ray tracing process as a differentiable operation, calculating the gradient of the multi-objective loss function relative to the lens surface control parameters, and simultaneously adjusting all surface parameters through a gradient descent algorithm.
[0011] Furthermore, a preferred solution is provided: In S3, the multi-objective loss function is defined as: , in, For the illuminance uniformity loss term, For rectangularity loss term, To prevent glare loss, To create a constraint loss term, This is the weighting coefficient for illuminance uniformity. Rectangularity weighting coefficient For anti-glare weighting coefficient, To generate constraint weight coefficients.
[0012] Furthermore, an optimal solution is provided: weighting coefficients. , , , The adaptive adjustment mechanism is as follows: Initialization phase: According to the tunnel lighting design specifications, priority should be given to setting the illuminance uniformity weight coefficient and the anti-glare weight coefficient; Iterative adjustment phase: During the optimization iteration process, the relative decrease rate of each loss item is monitored in real time. If the decrease rate of a certain loss item is lower than the preset threshold in continuous iterations, its weight is automatically increased. Contextualized mapping phase: Based on offline simulation data of tunnel width and installation height, a mapping database between initial weight values and tunnel parameters is established to quickly determine the initial weights for new projects.
[0013] Furthermore, a preferred embodiment is provided: S3 includes: S31: Obtain the parameters of the initial surface generated by S2, and select initial weights from the mapping database based on the tunnel parameters; S32: Under the current surface parameters, perform differentiable ray tracing to simulate light spot formation, and calculate the total loss and individual losses based on the multi-objective loss function, including illuminance uniformity loss, rectangularity loss, anti-glare loss, and manufacturing constraint loss. S33: Determine if the convergence condition is met. If it is, the optimization terminates and the current surface parameters are output; otherwise, continue to the next step. S34: Adaptive weight adjustment, updating all weights; S35: Through the differentiable rendering engine, calculate the gradient of the total loss with respect to the current surface parameters, and use a gradient-based optimization algorithm to update the lens surface parameters according to the gradient direction as the new surface parameters; S36: Apply S32-S36 to the new surface parameters until the convergence condition is met.
[0014] Furthermore, a preferred solution is provided: the convergence condition is a composite convergence condition, which must simultaneously satisfy the following: The relative change in total loss is less than ; Illuminance uniformity is greater than 0.70; The rectangularity of the light spot is greater than 0.95.
[0015] Secondly, the purpose of this invention is to propose a collaborative design system for a high-uniformity rectangular spot freeform surface lens for tunnel lighting. This system is based on a collaborative design method for a high-uniformity rectangular spot freeform surface lens for tunnel lighting, as described in any one or more of the above-mentioned schemes. The system includes: Data acquisition module: used to determine the geometric parameters and optical indicators of the tunnel where the lighting fixtures are located. The geometric parameters include: tunnel width, fixture installation height, and fixture spacing. The optical indicators include: average illuminance of the target area, illuminance uniformity, threshold increment, and target rectangular area size. Surface shape solving module: Based on the point source assumption and the law of conservation of energy, it constructs the initial surface shape of the double freeform lens by solving the Monge-Ampere equation; wherein, the double freeform lens includes an incident surface and an exit surface, both of which are represented by XY polynomial surfaces, and the construction of the initial surface shape is used to form a preliminary rectangular spot distribution on the target illumination surface; Surface optimization module: Based on differentiable rendering technology, it establishes an end-to-end differentiable optical model, performs multi-objective gradient optimization on the initial surface, and obtains the optimal surface parameters. The multi-objective gradient optimization includes defining a multi-objective loss function, calculating the gradient, and simultaneously optimizing illuminance uniformity, spot rectangularity, anti-glare performance, and manufacturing constraints through an adaptive weight adjustment mechanism.
[0016] Thirdly, the present invention aims to provide a computer device, the computer device including a memory and a processor, the memory storing a computer program, and when the processor runs the computer program stored in the memory, the processor executes a collaborative design method for a high uniformity rectangular spot freeform surface lens for tunnel lighting according to any one or more of the above-described schemes.
[0017] Fourthly, the present invention aims to provide a computer-readable storage medium for storing a computer program that executes the above-described method for co-designing a high-uniformity rectangular spot freeform surface lens for tunnel lighting.
[0018] Compared with the prior art, the advantages of the present invention are: Traditional solutions to the Monge-Ampère equation typically focus only on illuminance uniformity. This invention, however, uses it as a tool to rapidly generate high-quality initial surface patterns that already possess the characteristics of rectangular light spots. This is achieved by pre-setting the target illuminance distribution as a rectangular-like distribution with high central illuminance and smooth edge transitions, rather than a simple uniform circular distribution. This provides an excellent starting point for subsequent multi-objective collaborative optimization, reducing the number of iterations by more than 50% and solving the industry problem of simple numerical optimization easily getting trapped in local optima and experiencing slow convergence.
[0019] This invention abandons traditional sequential optimization or manual trial-and-error design methods, innovatively introducing differentiable rendering technology to establish an end-to-end differentiable optical model, thereby achieving direct, joint, and gradient optimization of freeform surface control parameters. The core of this method lies in representing the ray tracing process in optical simulation as a series of differentiable operations, allowing the gradient of the entire system's performance indicators with respect to the lens design parameters to be accurately calculated. The deep integration of the differentiable rendering gradient optimization algorithm used in this invention with the multi-objective problem of tunnel lighting represents an innovative application of cutting-edge technology in this field. In particular, the adaptive weight adjustment mechanism solves the classic problem of conflicting and unbalanced performance indicators in multi-objective optimization.
[0020] The method described in this invention achieves phased and precise targeting of design challenges through a collaborative design paradigm of "fast initial solution of the Monge-Ampère equation + differentiable rendering multi-objective fine optimization". The initial solution provides a physically reasonable starting point close to the global optimum, reducing the average number of subsequent optimization iterations by more than 50% and effectively avoiding local optima. Fine optimization simultaneously "sculpts" uniformity, rectangularity, and anti-glare performance on the "skeleton" of the initial solution, achieving comprehensive high performance that is difficult to achieve with a single method. This design paradigm shortens the traditional design cycle of several weeks to several days, achieving a double leap in efficiency and performance.
[0021] This invention is applicable to lens design scenarios for tunnel lighting. Attached Figure Description
[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.
[0023] Figure 1 This is a flowchart illustrating a collaborative design method for a high-uniformity rectangular spot freeform surface lens according to a specific embodiment of the present invention; Figure 2A flowchart of the optimization algorithm described in a specific embodiment of the present invention; Figure 3 This is a test diagram of the lighting effect according to a specific embodiment of the present invention. Detailed Implementation
[0024] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application can also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of this application with unnecessary detail.
[0025] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0026] Many specific details are set forth in the following description in order to provide a full understanding of this application. However, this application may also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.
[0027] Implementation Method 1 This embodiment proposes a collaborative design method for high-uniformity rectangular spot freeform surface lenses for tunnel lighting, such as... Figure 1 As shown, it includes the following steps: Step 1: Requirements Analysis and Parameterized Input Obtain the geometric parameters and optical properties of the tunnel where the lighting fixtures are located.
[0028] Geometric parameters include: tunnel width Lighting fixture installation height Spacing of lighting fixtures .
[0029] Optical indicators include: average illuminance of the target area Illumination uniformity Threshold increment Target rectangular region size ; Light source model: Select the .dat file (containing spatial light intensity distribution) or .ray file (ray set) of the LED chip.
[0030] Step 2: Initial surface model construction based on the law of conservation of energy and Snell's law of refraction This step aims to quickly obtain a physically feasible and near-optimal initial surface shape, providing high-quality seven points for subsequent optimization and greatly reducing the number of optimization iterations.
[0031] First, based on the law of conservation of energy, establish the mapping relationship between the light intensity distribution of the light source and the illuminance distribution of the target surface. The mathematical basis is expressed as: , in, It is the light intensity distribution of the light source. It is the illuminance distribution on the target's illuminated surface. and These represent solid angle and area elements, respectively.
[0032] The refraction path of light rays on the two curved surfaces of the lens is calculated using Snell's law of refraction, and its vector form is as follows: , in, Indicates the refractive index before refraction. Indicates the refractive index after refraction. It is the incident light direction vector. It is the direction vector of the refracted light. It is the unit normal vector at the intersection of the surfaces.
[0033] By solving the system of partial differential equations based on the above laws, the exact surface shapes of the two freeform surfaces can be determined, so that the light emitted by the LED light source is redistributed to form a uniform rectangular light spot on the target surface at a preset distance.
[0034] The Monge-Ampère Equation is used. This equation describes the conserved mapping of energy from the source angular space (u,v) to the target Cartesian space (x,y), and is expressed as: , in, It is the surface to be determined The Hessian matrix, Represents a determinant. Indicates the light intensity distribution of the light source. This represents the actual illuminance distribution.
[0035] The XY polynomial surface is represented as follows: , in, , Indicates the vertical difference in the Z-axis direction. Indicates curvature. Represents the constant of the conic surface. for The coefficient of the term, and Represent and The order of , This represents the radial distance measured from the center of the surface.
[0036] Traditional solutions to the Monge-Ampère equation typically focus only on illuminance uniformity. This invention, however, uses it as a tool to rapidly generate high-quality initial surface patterns that already possess the characteristics of rectangular light spots. This is achieved by pre-setting the target illuminance distribution as a rectangular-like distribution with high central illuminance and smooth edge transitions, rather than a simple uniform circular distribution. This provides an excellent starting point for subsequent multi-objective collaborative optimization, reducing the number of iterations by more than 50% and solving the industry problem of simple numerical optimization easily getting trapped in local optima and experiencing slow convergence.
[0037] Step 3: Multi-objective cooperative optimization loop algorithm based on differentiable rendering This method abandons traditional sequential optimization or manual trial-and-error design approaches, and innovatively introduces differentiable rendering technology to establish an end-to-end differentiable optical model, thereby achieving direct, joint, and gradient optimization of the control parameters of freeform surfaces. The core of this step lies in representing the ray tracing process in optical simulation as a series of differentiable operations, so that the gradient of the performance indicators of the entire system with respect to the lens design parameters can be accurately calculated.
[0038] like Figure 2 As shown, this step is based on differentiable rendering technology to establish an end-to-end differentiable optical model, perform multi-objective gradient optimization on the initial surface shape, and obtain the optimal surface shape parameters.
[0039] The multi-objective gradient optimization includes defining a multi-objective loss function, calculating the gradient, and simultaneously optimizing illuminance uniformity, spot rectangularity, anti-glare performance, and manufacturing constraints through an adaptive weight adjustment mechanism.
[0040] Specifically, the following steps are included: (1) Differentiable model and gradient calculation: In the optical design software, the incident and exit surfaces of the double freeform lens are parameterized into differentiable models. In this embodiment, an XY polynomial surface is used. By differentiable ray tracing, the target loss function L is calculated relative to all surface control parameters. gradient This gradient indicates the direction and magnitude by which each parameter should be adjusted to simultaneously optimize all performance objectives.
[0041] (2) Refined multi-objective loss function: The objective function F of this invention is not a simple weighted sum, but is carefully designed to directly reflect the comprehensive performance requirements of tunnel lighting: , Illuminance uniformity loss term Defined as the illuminance value at each sampling point within the target rectangular area. Standard deviation With average illuminance The square of the ratio, that is This loss term directly penalizes the brightness fluctuations within the light spot.
[0042] Rectangularity loss term By calculating the actual illuminance distribution With the ideal rectangular binary template The quantification is based on the mean square error between (1 within the target area and 0 outside the area), i.e. This loss term forces the light spot boundary to shrink towards a rectangle.
[0043] Anti-glare loss item This is specifically defined as a large-angle light intensity penalty term. The light intensity integral is calculated within the vertical angle range of 75° to 90°. Total luminous flux in the range of 0° to 90° The ratio, i.e. This loss term actively suppresses light that may cause glare.
[0044] Manufacturing constraint loss Used to ensure the manufacturability of the lens, such as when the center thickness of the lens is... Below the minimum threshold Or change the thickness Above the maximum threshold When this happens, the loss term will generate a penalty value, guiding the optimization back to the feasible parameter space.
[0045] This is the weighting coefficient for illuminance uniformity. Rectangularity weighting coefficient For anti-glare weighting coefficient, To generate constraint weight coefficients.
[0046] (3) Adaptive weight adjustment mechanism: Weighting coefficient , , , Instead of fixed empirical values, these values are determined through an adaptive adjustment mechanism, which is one of the core elements of this method and includes the following steps: A. Initialization: Based on the priority requirements for uniformity and glare threshold increment (TI) in the "Specifications for Lighting Design of Highway Tunnels", the initial setting is as follows: and Dominant and Secondly, prioritize the optimization of safety-related performance.
[0047] B. Adaptive Adjustment: During the optimization iteration process, the relative decrease rate of each loss term is monitored in real time. If a certain loss item In continuous The average rate of decline in each iteration was lower than a preset threshold. If the optimization stalls, the system will consider it to have stagnated. At this point, the system will automatically adjust its weights. improve This will inject stronger optimization momentum and break the deadlock.
[0048] Furthermore, the preset threshold Based on the learning rate of the optimization algorithm Configure the settings; the experience range is... To address the design problem of the tunnel lighting lens in this invention, Setting the concentration between 1% and 3% will yield the best results.
[0049] Furthermore, the improvement rate The calculation formula is as follows: , in, for In recent Average rate of decline over iterations This is a preset threshold. and These represent the minimum and maximum adjustment ranges, respectively. Based on simulation experiments of numerous tunnel lighting design scenarios, to balance the sensitivity and stability of the adjustment, the empirical values for each parameter are: 10% 30%; threshold 1.5%; Number of consecutive iterations :5.
[0050] C. Contextualized Mapping: By applying different tunnel geometric parameters, width... Installation height Extensive offline simulations were conducted, establishing a database mapping initial weight values to tunnel parameters. For new tunnel projects, the system can be based on its... and Quickly query and retrieve a set of near-optimal initial weights, significantly improving design efficiency and success rate.
[0051] Furthermore, step 3 includes: Step 3.1, Initialization: Load the parameters of the initial surface type generated in Step 2. The weight coefficients are retrieved from the mapping database based on tunnel parameters or initialized according to rules. .
[0052] Step 3.2, Microscopic ray tracing and loss calculation: Under the current parameters Next, perform differentiable ray tracing to simulate light spot formation. Calculate the total loss based on the aforementioned loss function formula. and the losses of each item .
[0053] Step 3.3, Convergence Check: Determine if the convergence conditions are met. The convergence conditions are complex and must be satisfied simultaneously: (i) Total loss The relative change is less than ; (ii) Illuminance uniformity ; (iii) Spot rectangularity .
[0054] If satisfied, the optimization terminates, and the current surface parameters are output. Otherwise, proceed to the next step.
[0055] Step 3.4, Adaptive Weight Adjustment (Inner Loop Decision): Calculate the most recent loss for each component. The average rate of decline in each iteration. The aforementioned adaptive weight adjustment mechanism is invoked to determine whether and how to update the weight coefficients. .
[0056] Step 3.5, Gradient Calculation and Parameter Update: Calculate the total loss using a differentiable rendering engine. For the current surface parameters gradient A gradient-based optimization algorithm is used to update the lens surface parameters according to the gradient direction: , in, This is the learning rate.
[0057] Step 3.6, Iteration: Let Then proceed to step 3.2 and continue execution until convergence.
[0058] Implementation Method 2 This embodiment is a further illustrative example of the collaborative design method for a high-uniformity rectangular spot freeform surface lens for tunnel lighting described in Embodiment 1.
[0059] like Figure 3 As shown, the double freeform lens designed using the method described in Embodiment 1 is applied to an LED lamp. The LED lamp cover is placed at a height of 1.07m. The light emitted by the LED light source module is distributed twice through its precise optical surface, and finally forms a uniform rectangular lighting area on the road surface.
[0060] Implementation Method 3 This embodiment proposes a collaborative design system for a high-uniformity rectangular spot freeform surface lens for tunnel lighting, based on the collaborative design method for a high-uniformity rectangular spot freeform surface lens for tunnel lighting as described in Embodiment 1. The system includes: Data acquisition module: used to determine the geometric parameters and optical indicators of the tunnel where the lighting fixtures are located. The geometric parameters include: tunnel width, fixture installation height, and fixture spacing. The optical indicators include: average illuminance of the target area, illuminance uniformity, threshold increment, and target rectangular area size. Surface shape solving module: Based on the point source assumption and the law of conservation of energy, it constructs the initial surface shape of the double freeform lens by solving the Monge-Ampere equation; wherein, the double freeform lens includes an incident surface and an exit surface, both of which are represented by XY polynomial surfaces, and the construction of the initial surface shape is used to form a preliminary rectangular spot distribution on the target illumination surface; Surface optimization module: Based on differentiable rendering technology, it establishes an end-to-end differentiable optical model, performs multi-objective gradient optimization on the initial surface, and obtains the optimal surface parameters. The multi-objective gradient optimization includes defining a multi-objective loss function, calculating the gradient, and simultaneously optimizing illuminance uniformity, spot rectangularity, anti-glare performance, and manufacturing constraints through an adaptive weight adjustment mechanism.
[0061] It is understood that the present invention has been described through some embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.
Claims
1. A collaborative design method for a high-uniformity rectangular spot freeform surface lens for tunnel lighting, characterized in that, Includes the following steps: S1: Determine the geometric parameters and optical indicators of the tunnel where the lighting fixtures are located. The geometric parameters include: tunnel width, fixture installation height, and fixture spacing. The optical indicators include: average illuminance of the target area, illuminance uniformity, threshold increment, and target rectangular area size. S2: Based on Snell's law of refraction and the law of conservation of energy, the initial surface shape of the double freeform lens is constructed by solving the Monge-Ampere equation; wherein, the double freeform lens includes an incident surface and an exit surface, both of which are represented by XY polynomial surfaces, and the construction of the initial surface shape is used to form a preliminary rectangular spot distribution on the target illumination surface; S3: Based on differentiable rendering technology, establish an end-to-end differentiable optical model, perform multi-objective gradient optimization on the initial surface shape, and obtain the optimal surface shape parameters. The multi-objective gradient optimization includes defining a multi-objective loss function, calculating the gradient, and simultaneously optimizing illuminance uniformity, spot rectangularity, anti-glare performance, and manufacturing constraints through an adaptive weight adjustment mechanism.
2. The method for collaborative design of a high-uniformity rectangular spot freeform surface lens for tunnel lighting according to claim 1, characterized in that, In S2, the initial surface shape is constructed based on solving the Monge-Ampere equation, as follows: ， in, It is the surface to be determined The Hessian matrix, Represents a determinant. Indicates the light intensity distribution of the light source. This represents the actual illuminance distribution.
3. The method for collaborative design of a high-uniformity rectangular spot freeform surface lens for tunnel lighting according to claim 1, characterized in that, In S3, the differentiable rendering technology includes: representing the ray tracing process as a differentiable operation, calculating the gradient of the multi-objective loss function relative to the lens surface control parameters, and simultaneously adjusting all surface parameters through a gradient descent algorithm.
4. The method for collaborative design of a high-uniformity rectangular spot freeform surface lens for tunnel lighting according to claim 1, characterized in that, In S3, the multi-objective loss function is defined as: ， in, For the illuminance uniformity loss term, For rectangularity loss term, To prevent glare loss, To create a constraint loss term, This is the weighting coefficient for illuminance uniformity. Rectangularity weighting coefficient For anti-glare weighting coefficient, To generate constraint weight coefficients.
5. The method for collaborative design of a high-uniformity rectangular spot freeform surface lens for tunnel lighting according to claim 4, characterized in that, Weighting coefficient , , , The adaptive adjustment mechanism is as follows: Initialization phase: According to the tunnel lighting design specifications, priority should be given to setting the illuminance uniformity weight coefficient and the anti-glare weight coefficient; Iterative adjustment phase: During the optimization iteration process, the relative decrease rate of each loss item is monitored in real time. If the decrease rate of a certain loss item is lower than the preset threshold in continuous iterations, its weight is automatically increased. Contextualized mapping phase: Based on offline simulation data of tunnel width and installation height, a mapping database between initial weight values and tunnel parameters is established to quickly determine the initial weights for new projects.
6. The method for collaborative design of a high-uniformity rectangular spot freeform surface lens for tunnel lighting according to claim 5, characterized in that, S3 includes: S31: Obtain the parameters of the initial surface generated by S2, and select initial weights from the mapping database based on the tunnel parameters; S32: Under the current surface parameters, perform differentiable ray tracing to simulate light spot formation, and calculate the total loss and individual losses based on the multi-objective loss function, including illuminance uniformity loss, rectangularity loss, anti-glare loss, and manufacturing constraint loss. S33: Determine if the convergence condition is met. If it is, the optimization terminates and the current surface parameters are output; otherwise, continue to the next step. S34: Adaptive weight adjustment, updating all weights; S35: Through the differentiable rendering engine, calculate the gradient of the total loss with respect to the current surface parameters, and use a gradient-based optimization algorithm to update the lens surface parameters according to the gradient direction as the new surface parameters; S36: Apply S32-S36 to the new surface parameters until the convergence condition is met.
7. The method for co-designing a high-uniformity rectangular spot freeform surface lens for tunnel lighting according to claim 6, characterized in that, The convergence condition is a composite convergence condition, which must be satisfied simultaneously: The relative change in total loss is less than ; Illuminance uniformity is greater than 0.70; The rectangularity of the light spot is greater than 0.
95.
8. A collaborative design system for a high-uniformity rectangular spot freeform surface lens for tunnel lighting, characterized in that, The system is based on a collaborative design method for a high-uniformity rectangular spot freeform surface lens for tunnel lighting as described in any one of claims 1-7. The system includes: Data acquisition module: used to determine the geometric parameters and optical indicators of the tunnel where the lighting fixtures are located. The geometric parameters include: tunnel width, fixture installation height, and fixture spacing. The optical indicators include: average illuminance of the target area, illuminance uniformity, threshold increment, and target rectangular area size. Surface shape solving module: Based on the point source assumption and the law of conservation of energy, it constructs the initial surface shape of the double freeform lens by solving the Monge-Ampere equation; wherein, the double freeform lens includes an incident surface and an exit surface, both of which are represented by XY polynomial surfaces, and the construction of the initial surface shape is used to form a preliminary rectangular spot distribution on the target illumination surface; Surface optimization module: Based on differentiable rendering technology, it establishes an end-to-end differentiable optical model, performs multi-objective gradient optimization on the initial surface, and obtains the optimal surface parameters. The multi-objective gradient optimization includes defining a multi-objective loss function, calculating the gradient, and simultaneously optimizing illuminance uniformity, spot rectangularity, anti-glare performance, and manufacturing constraints through an adaptive weight adjustment mechanism.
9. A computer device, characterized in that, The computer device includes a memory and a processor. The memory stores a computer program. When the processor runs the computer program stored in the memory, the processor executes a collaborative design method for a high-uniformity rectangular spot freeform surface lens for tunnel lighting according to any one of claims 1-7.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program that executes a collaborative design method for a high-uniformity rectangular spot freeform surface lens for tunnel lighting, as described in any one of claims 1-7.