Multi-objective optimization cooperative control method for goods shelf illumination and related equipment thereof

Abstract

The invention discloses a multi-objective optimization cooperative control method for goods shelf illumination and related equipment thereof, and belongs to the technical field of illumination control, and the method specifically comprises the steps: projecting a multi-wavelength grid detection pattern adaptive to the spacing of shelf laminates to the surface of a goods shelf, collecting a reflected light signal, and synchronously recording spatial distribution and spectrum composition data; extracting grid node coordinates by processing spatial data, calculating an offset vector relative to a reference position, analyzing spectral data to decompose wavelength components, and calculating a difference value between each wavelength light intensity and a reference value; generating a light beam direction adjustment instruction based on the direction amplitude of the offset vector, and generating a spectrum compensation instruction according to the light intensity difference, so that a basic illumination system adjusts illumination parameters accordingly; according to the invention, multi-parameter collaborative optimization of goods shelf illumination is realized, and the illumination quality in a complex environment is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of lighting control technology, specifically to a multi-objective optimization and collaborative control method for shelf lighting and related equipment. Background Technology

[0002] Shelf lighting plays a crucial role in large warehouse-style shopping malls and logistics warehousing centers, affecting not only the display of goods but also the efficiency of warehousing operations and energy consumption. As modern commercial spaces evolve towards higher floors and greater density, shelf lighting systems face multiple demands, including uniform illumination, energy control, and visual comfort. Traditional shelf lighting solutions primarily rely on fixed-installation luminaires, achieving basic lighting functions through zoned or timed dimming, which struggles to adapt to complex and ever-changing real-world lighting environments.

[0003] In existing technologies, several sensor-based lighting control schemes have emerged to address the problem of optimizing shelf lighting control. For example, a network of illuminance sensors monitors the illuminance levels in different areas of the shelf, and the brightness of the lights is adjusted based on preset thresholds. Alternatively, image sensors are used to capture images of the shelf, and image processing techniques are employed to identify shadowed areas and adjust the local lighting accordingly. These schemes, to some extent, improve the targeting of lighting control and enable dynamic adjustment of basic lighting parameters.

[0004] However, due to the three-dimensional complexity of shelf structures and the randomness of merchandise display, single-type sensors struggle to fully capture the spatial heterogeneity of light distribution. Threshold-based or simple image processing-based control methods have limited ability to handle light coupling effects, easily leading to localized over- or under-illumination. Existing solutions largely rely on direct feedback control, lacking in-depth modeling of light propagation patterns, making it difficult to achieve multi-objective collaborative optimization. Therefore, a shelf lighting control method is needed that can more accurately perceive light distribution characteristics, more effectively decouple light interference, and achieve multi-parameter collaborative optimization. Summary of the Invention

[0005] The purpose of this invention is to provide a multi-objective optimization collaborative control method and related equipment for shelf lighting, and to solve the following technical problems: Existing rack lighting control methods are prone to local lighting imbalance when dealing with high-level rack lighting scenarios because they are difficult to effectively perceive the heterogeneity of light distribution in complex spaces and lack in-depth modeling of the coupling effects of multiple parameters, thus failing to achieve multi-objective collaborative optimization.

[0006] The objective of this invention can be achieved through the following technical solutions: A multi-objective optimization collaborative control method for shelf lighting includes the following steps: S1. Project a grid detection pattern containing multi-wavelength components onto the shelf surface. The node spacing of the grid detection pattern is proportional to the shelf shelf spacing. S2. Collect the reflected light signal of the grid detection pattern on the shelf surface, and simultaneously record the spatial distribution data and spectral composition data of the reflected light signal; S3. Process spatial distribution data, extract the position coordinates of each node in the grid detection pattern, and calculate the offset vector of each node relative to the reference position. S4. Process the spectral composition data, decompose the wavelength components of the reflected light signal, and calculate the difference between the light intensity value of each wavelength component and the reference light intensity value. S5. Generate beam direction adjustment instructions for the basic lighting system based on the direction and amplitude of the offset vector, and generate spectral compensation instructions for the basic lighting system based on the light intensity difference. S6. The basic lighting system adjusts the lighting parameters according to the beam direction adjustment command and the spectral compensation command.

[0007] As a further aspect of the present invention: in step S1, the process of projecting a grid detection pattern containing multi-wavelength components onto the shelf surface is as follows: Three coherent beams of light with red, green and blue wavelengths are generated and input into a non-common-path interference structure. A multi-wavelength interference field is generated at the output of the interference structure. Dynamic phase modulation is applied to the multi-wavelength interference field. The frequency of the dynamic phase modulation is consistent with the driving frequency of the acousto-optic modulator corresponding to each wavelength. The phase-modulated interference field is converted into a grid-like light intensity distribution by a diffractive optical element. The grid-like light intensity distribution is projected onto the shelf surface to form a grid detection pattern. During the projection process, the phase delay of the acousto-optic modulator is changed synchronously. The phase delay ranges from zero to 2π. The light intensity distribution of the grid detection pattern corresponding to each phase delay is recorded.

[0008] As a further aspect of the present invention: the specific process of acquiring the reflected light signal in step S2 is as follows: Interference fringe images of the shelf surface are acquired using a monochromatic area array detector. During the sequential projection of red, green, and blue wavelengths, interference fringe image sequences of the corresponding wavelengths are acquired. Each wavelength interference fringe image sequence contains four frames of interferograms with a phase difference of π / 2. Four-step phase shift calculations are performed on the interference fringe image sequence of each wavelength to obtain the phase principal value distribution of that wavelength. Phase unwrapping operation is performed on the principal phase value distribution to obtain the continuous phase distribution of the wavelength. The continuous phase distributions of the three wavelengths are then combined into a full-color phase distribution map, which contains spatial distribution data and spectral composition data.

[0009] As a further aspect of the present invention: the specific process for processing spatially distributed data in step S3 is as follows: Complex amplitude information of each grid node is extracted from the full-color phase distribution map. The complex amplitude information includes amplitude components and phase components. The coherence matrix of the complex amplitude of adjacent grid nodes is calculated. The polarization state distribution is obtained by eigenvalue decomposition of the coherence matrix. The mapping relationship between the polarization state distribution and the surface normal vector is established. The surface normal field is reconstructed according to the mapping relationship. The surface normal field is subjected to surface integration to obtain the surface height distribution. The elevation difference field between the surface height distribution and the reference height distribution is calculated. The elevation difference field is decomposed by spherical harmonic function, and the first sixteen spherical harmonic coefficients are extracted. The three-dimensional offset vector of each grid node is calculated based on the spherical harmonic coefficients. The direction of the three-dimensional offset vector is determined by the first-order spherical harmonic coefficients, and the amplitude of the three-dimensional offset vector is determined by the second-order spherical harmonic coefficients.

[0010] As a further aspect of the present invention: the specific process for processing the spectral composition data in step S4 is as follows: The Stokes parameters for each wavelength are separated from the full-color phase distribution map. The Stokes parameters contain four polarization components. The Mueller matrix representation of the Stokes parameters for each wavelength is calculated. The depolarization coefficient distribution is obtained through the singular value decomposition of the Mueller matrix. The correspondence between the depolarization coefficient and the surface roughness is established. The bidirectional reflection distribution function of the surface is calculated based on the correspondence. The bidirectional reflectance distribution function is compared with the reference bidirectional reflectance distribution function to obtain the reflectance change at each wavelength. The light intensity difference at each wavelength is calculated based on the reflectance change. The light intensity difference is the product of the reflectance change and the reference light intensity value. The value and sign of the light intensity difference at each wavelength are recorded.

[0011] As a further aspect of the present invention: in step S5, the specific process of generating the beam direction adjustment command of the basic lighting system based on the direction and amplitude of the offset vector is as follows: A mapping relationship between the three-dimensional offset vector and the illumination beam direction is established. The mapping relationship is based on the orthogonal decomposition of the spherical harmonic coefficients. The three-dimensional offset vector is projected onto the azimuth and zenith dimensions in the spherical coordinate system. The difference between the projected component of the azimuth dimension and the reference azimuth is calculated to obtain the horizontal deflection adjustment. The difference between the projected component of the zenith dimension and the reference zenith is calculated to obtain the pitch adjustment. The beam divergence angle adjustment is calculated based on the amplitude component of the spherical harmonic function coefficients. The horizontal deflection adjustment, pitch adjustment, and beam divergence angle adjustment are encoded into beam direction adjustment commands, which include the numerical value and sign of each adjustment.

[0012] As a further aspect of the present invention: in step S5, the specific process of generating the spectral compensation command for the basic lighting system based on the light intensity difference is as follows: A correspondence between light intensity difference and spectral power distribution is established. The correspondence is based on the eigenvector analysis of the Mueller matrix. The light intensity difference of each wavelength is input into the spectral reconstruction algorithm. The spectral reconstruction algorithm outputs the target spectral power distribution. The relative difference between the target spectral power distribution and the reference spectral power distribution is calculated. The intensity adjustment coefficient of each wavelength component is calculated based on the relative difference. The intensity adjustment coefficient is proportional to the singular value of the Mueller matrix. The intensity adjustment coefficient is converted into a drive current adjustment amount, which is linearly related to the intensity adjustment coefficient. The drive current adjustment amount corresponding to each wavelength is encoded as a spectral compensation command.

[0013] As a further aspect of the present invention: In step S6, the specific process by which the basic lighting system adjusts the lighting parameters according to the beam direction adjustment command and the spectral compensation command is as follows: The analysis focuses on the horizontal deflection adjustment, pitch adjustment, and beam divergence angle adjustment in the beam direction adjustment command. Based on the horizontal deflection adjustment, the horizontal deflection mechanism of the illumination unit is adjusted; the rotation angle of the horizontal deflection mechanism is proportional to the horizontal deflection adjustment. Similarly, based on the pitch adjustment, the pitch adjustment mechanism of the illumination unit is adjusted; the rotation angle of the pitch adjustment mechanism is proportional to the pitch adjustment. Finally, based on the beam divergence angle adjustment, the optical focusing system of the illumination unit is adjusted; the change in the lens spacing of the optical focusing system is proportional to the beam divergence angle adjustment. The drive current adjustment amount for each wavelength in the spectral compensation command is analyzed. The drive current of each wavelength light-emitting element in the lighting unit is adjusted according to the drive current adjustment amount. The change in drive current is proportional to the drive current adjustment amount. Each adjustment operation is executed in the order of directional adjustment first and then spectral adjustment.

[0014] As a further aspect of the present invention: the projection of a grid detection pattern containing multi-wavelength components onto the shelf surface is based on an auxiliary detection unit, wherein the light emission direction of the light source component of the auxiliary detection unit is consistent with the light emission direction of the basic lighting system, and the optical sensor of the auxiliary detection unit establishes a data connection with the control unit of the basic lighting system. During the detection phase, the light source component of the auxiliary detection unit projects a grid detection pattern, and the optical sensor collects the reflected light signal; during the lighting adjustment phase, the control unit of the basic lighting system generates lighting parameter adjustment instructions based on the processing results of the reflected light signal.

[0015] The present invention also includes a multi-objective optimization collaborative control device for shelf lighting, for implementing the above-described multi-objective optimization collaborative control method for shelf lighting, comprising: The light source assembly is used to project a grid detection pattern containing multi-wavelength components onto the shelf surface. The node spacing of the grid detection pattern is proportional to the shelf shelf spacing. An optical sensor is used to collect the reflected light signal from the shelf surface to the grid detection pattern, and simultaneously record the spatial distribution data and spectral composition data of the reflected light signal; The signal processing unit is used to process spatial distribution data, extract the position coordinates of each node in the grid detection pattern, calculate the offset vector of each node relative to the reference position; and process spectral composition data, decompose the wavelength components of the reflected light signal, and calculate the difference between the light intensity value of each wavelength component and the reference light intensity value. The control unit is used to generate beam direction adjustment commands for the basic lighting system based on the direction and magnitude of the offset vector, and to generate spectral compensation commands for the basic lighting system based on the light intensity difference. The basic lighting system is used to adjust lighting parameters according to beam direction adjustment commands and spectral compensation commands.

[0016] The beneficial effects of this invention are: This invention projects a special multi-wavelength grid pattern onto the surface of a shelf, utilizing the principle of optical interference to precisely capture the illumination distribution on the shelf surface, effectively solving the problem of traditional methods' difficulty in accurately perceiving complex spatial illumination distribution. By analyzing the interference fringes formed by reflected light, the system can simultaneously acquire spatial location information and spectral characteristics, establishing a precise correspondence between surface deformation and illumination parameters. Based on the calculation of three-dimensional offset vectors and spectral differences, the system generates adjustment instructions for the illumination direction and spectral composition, enabling the lighting device to simultaneously optimize the beam angle and spectral composition. This method overcomes the limitations of traditional single-parameter adjustment, achieving synergistic compensation for shaded and overly bright areas, significantly improving the uniformity of high-level shelf lighting. The entire system, through a detection-then-execution workflow, ensures the accuracy and timeliness of lighting adjustments, providing a more precise and reliable lighting solution for complex warehousing environments. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 This is a flowchart illustrating the multi-objective optimization and collaborative control method for shelf lighting according to the present invention; Figure 2 This is a schematic diagram of the structure of the multi-objective optimization and collaborative control device for shelf lighting of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 As shown, the present invention is a multi-objective optimization cooperative control method for shelf lighting, comprising the following steps: S1. Project a grid detection pattern containing multi-wavelength components onto the shelf surface. The node spacing of the grid detection pattern is proportional to the shelf shelf spacing. S2. Collect the reflected light signal of the grid detection pattern on the shelf surface, and simultaneously record the spatial distribution data and spectral composition data of the reflected light signal; S3. Process spatial distribution data, extract the position coordinates of each node in the grid detection pattern, and calculate the offset vector of each node relative to the reference position. S4. Process the spectral composition data, decompose the wavelength components of the reflected light signal, and calculate the difference between the light intensity value of each wavelength component and the reference light intensity value. S5. Generate beam direction adjustment instructions for the basic lighting system based on the direction and amplitude of the offset vector, and generate spectral compensation instructions for the basic lighting system based on the light intensity difference. S6. The basic lighting system adjusts the lighting parameters according to the beam direction adjustment command and the spectral compensation command.

[0021] In S1, the process of projecting a grid detection pattern containing multi-wavelength components onto the shelf surface is as follows: The projection of a multi-wavelength grid detection pattern relies on a high-precision optical emission system. This system first generates three coherent beams of light: red, green, and blue. The coherent light generation unit typically uses narrow-linewidth semiconductor lasers. The center wavelength of the red laser source is generally selected as 635nm or 650nm, the green laser source as 532nm, and the blue laser source as 450nm or 473nm. Each laser source must be optically isolated to suppress backscattering interference and ensure the phase stability of the output beam. Simultaneously, a collimating lens is used to calibrate the output beam into parallel light, ensuring that the beam cross-sectional dimensions of the three beams match the incident ports of the subsequent optical structure.

[0022] After three coherent beams are generated, they are guided to a non-common-path interference structure via fiber couplers or a mirror array. The core feature of this structure is that the three beams propagate independently along their respective optical paths at the interference front. Each path is equipped with a precision displacement stage to fine-tune the optical path difference, preventing crosstalk caused by overlapping optical paths between beams of different wavelengths during propagation. The interference cavity contains a beam splitter and a beam combiner. The beam splitter separates each coherent beam into a reference beam and an object beam. The reference beam enters the beam combiner directly, while the object beam undergoes specific optical path adjustment before merging with the reference beam, ultimately forming a multi-wavelength interference field at the output of the interference structure. In this multi-wavelength interference field, interference fringes of different wavelengths superimpose, forming an initial intensity distribution with wavelength-distinct characteristics.

[0023] To achieve dynamic phase modulation, the multi-wavelength interference field needs to pass through a phase modulation unit composed of acousto-optic modulators. Each wavelength corresponds to an independent acousto-optic modulator, and the driving frequency of the acousto-optic modulator must match the frequency parameters of the laser source for the corresponding wavelength to ensure that the modulation signal can accurately act on the target wavelength beam. The core of dynamic phase modulation is to change the radio frequency driving signal of the acousto-optic modulator so that the beam generates a controllable phase delay when passing through the acousto-optic medium. During the modulation process, the phase delay needs to be gradually adjusted from 0 to 2π, with the adjustment step size set according to the detection accuracy requirements, usually controlled between 0.01π and 0.05π. While the phase delay changes, the data acquisition module needs to simultaneously record the light intensity distribution of the grid detection pattern under each phase delay, including the light intensity value of each pixel and the corresponding phase delay parameter, forming a complete phase-intensity mapping dataset, providing a benchmark reference for subsequent reflected light signal analysis.

[0024] The phase-modulated interference field needs to be converted into a grid-like light intensity distribution using diffractive optical elements. These elements are fabricated using binary optics, with a periodic grid structure etched onto their surface. The period of this structure (i.e., the grid node spacing) must be proportional to the shelf shelf spacing. This proportionality is typically determined based on the shelf height; for example, when the shelf shelf spacing is 30cm, the grid node spacing can be set to 3cm or 5cm to ensure the grid pattern completely covers each shelf surface and the node density is sufficient to reflect local illumination differences. The diffractive optical elements decompose the incident interference field into multiple diffraction orders propagating in specific directions. These diffraction orders converge in space to form a regular grid-like light intensity distribution. This distribution is focused by a projection lens and projected onto the shelf surface as parallel light, ultimately forming a grid detection pattern for inspection.

[0025] In step S2, the specific process of acquiring the reflected light signal is as follows: The acquisition of reflected light signals relies on a monochromatic area array detector. The detector type typically uses a scientific-grade CMOS or CCD image sensor, whose spectral response range must cover the red, green, and blue wavelength ranges, and possess high frame rate and high dynamic range characteristics to capture weak interference fringe signals and avoid light intensity saturation. The detector's pixel resolution must match the node density of the grid detection pattern, ensuring that each grid node corresponds to at least one pixel unit on the detector. Simultaneously, the imaging magnification is adjusted through the lens module to ensure that the grid pattern on the shelf surface is completely imaged within the detector's effective photosensitive area.

[0026] The acquisition process must be synchronized with the wavelength projection sequence in S1. Specifically, during the sequential projection of red, green, and blue wavelengths, the detector acquires the corresponding interference fringe image sequence for each wavelength. The projection duration for each wavelength must meet the requirement of the detector acquiring four frames of images, and the acquisition interval between adjacent frames must be coordinated with the phase adjustment period of the acousto-optic modulator to ensure that the phase difference between the four frames is strictly π / 2. Specifically, during the acquisition of the first frame, the phase delay of the acousto-optic modulator for the corresponding wavelength is set to 0; for the second frame, it is set to π / 2; for the third frame, it is set to π; and for the fourth frame, it is set to 3π / 2, forming a four-step phase-shifted image sequence. During acquisition, shutter control or gain adjustment must be used to maintain consistent exposure parameters for each frame to avoid interference from exposure differences in subsequent phase calculations.

[0027] After acquiring the interference fringe image sequence for each wavelength, a four-step phase shift calculation is required to obtain the principal phase value distribution for that wavelength. The core of the four-step phase shift calculation is to extract phase information using four frames of images with different phase shifts through intensity superposition and differential operations. During the calculation, the intensity values ​​of corresponding pixels in the four frames are first normalized to eliminate the influence of detector dark current and background light. Then, based on the phase shift interference principle, the principal phase value of each pixel is obtained through the combination of the intensity values ​​from the four frames. The numerical range of the principal phase value distribution is typically limited to between -π and π. Due to potential height abrupt changes or occlusions on the shelf surface, discontinuous regions in the principal phase value distribution may appear, requiring phase unwrapping operations to eliminate these discontinuities.

[0028] Phase unwrapping operations employ either region growing or least squares methods. The core idea is to start with a phase-continuous seed region and gradually expand it to the entire image region. By correcting phase jumps (subtracting 2π from phase differences exceeding π and adding 2π to phase differences below -π), a continuous phase distribution covering the entire imaging region is obtained. In this continuous phase distribution, the phase value of each pixel corresponds to its spatial position on the shelf surface, with phase changes reflecting height differences in spatial position.

[0029] After calculating the continuous phase distribution of the three wavelengths, they need to be synthesized into a full-color phase distribution map. The synthesis process uses an RGB color space mapping method, mapping the continuous phase distribution of the red wavelength to the red channel of the RGB space, the continuous phase distribution of the green wavelength to the green channel, and the continuous phase distribution of the blue wavelength to the blue channel. The phase value of each channel needs to be converted to a grayscale value (typically mapping the phase range from -π to π to a grayscale range of 0 to 255), and then a full-color image is generated using a color synthesis algorithm. In this full-color phase distribution map, the grayscale changes of different color channels reflect the spectral composition data, while the relationship between the spatial coordinates of pixels and the phase grayscale values ​​carries the spatial distribution data, achieving unified storage and subsequent retrieval of the two core data types.

[0030] In S3, the specific process for processing spatially distributed data is as follows: Spatial distribution data processing uses a full-color phase distribution map as input. First, it's necessary to extract the complex amplitude information of each node in the grid detection pattern. The extraction process begins by locating the grid nodes in the full-color phase distribution map. Image edge detection algorithms are used to identify node feature points (usually stripe intersections) in the grid pattern, determining the pixel coordinates of each node. For each node, the light intensity and phase values ​​are extracted from the red, green, and blue channels respectively. The light intensity value, after normalization, is used as the amplitude component of the complex amplitude, and the phase value is directly used as the phase component of the complex amplitude, thus constructing the complex amplitude parameters for each node.

[0031] After obtaining the complex amplitude information of each node, the coherence matrix of the complex amplitudes of adjacent grid nodes is calculated. The coherence matrix is ​​a 2×2 complex matrix, whose elements are composed of the combination of the complex amplitudes of two adjacent nodes. Specifically, the diagonal elements of the matrix are the squares of the complex amplitudes of the two nodes (i.e., light intensity), and the off-diagonal elements are the combination of the product of the complex amplitudes of the two nodes and their conjugates. This matrix can quantify the optical field correlation and polarization correlation characteristics between adjacent nodes.

[0032] By performing eigenvalue decomposition on the coherence matrix, two eigenvalues ​​and their corresponding eigenvectors can be obtained. The direction of the eigenvectors reflects the polarization state distribution in that region. There is an inherent mapping relationship between the polarization state distribution and the normal vector of the shelf surface. This is because the surface normal vector determines the trend of polarization state changes in reflected light—when the angle between the surface normal vector and the incident light direction changes, the amplitude ratio of the p-polarized component to the s-polarized component of the reflected light changes, leading to a polarization state shift. Based on this physical law, a model of the correspondence between polarization state parameters and the surface normal vector needs to be established through preliminary calibration experiments. During the calibration process, a standard planar sample is used, and the surface normal is changed by adjusting the sample tilt angle. Polarization state data under different normal directions are recorded, ultimately forming a mapping relationship database.

[0033] Based on the established mapping relationship, the polarization state distribution of all grid nodes in the full-color phase distribution map is analyzed to obtain the surface normal vector corresponding to each node, thereby reconstructing the surface normal field of the entire shelf surface. The surface normal field is a vector field describing the macroscopic morphology of the shelf surface, and the direction of each vector is perpendicular to the surface tangent direction at the corresponding position. To convert the surface normal field into a surface height distribution that can be directly used for position analysis, surface integration is performed on the surface normal field. The surface integration process uses a preset reference plane as the integration reference surface. By accumulating the integration along the surface tangent direction, the change in the direction of the normal vector is converted into a change in the height coordinate, and finally the height value of each grid node in three-dimensional space is output, forming the surface height distribution data.

[0034] The surface height distribution data needs to be compared with the baseline height distribution to obtain the elevation difference field. The baseline height distribution is the preset height data of the shelf in its unloaded or standard display state, which is stored in the system database through prior measurement. The elevation difference field is calculated by subtracting the baseline height value from the actual height value of each grid node. The result can be positive or negative; a positive value indicates that the actual height of the node is higher than the baseline, and a negative value indicates that it is lower than the baseline. The absolute value of the difference reflects the magnitude of the height deviation.

[0035] To extract effective positional offset information from the elevation difference field, spherical harmonic decomposition is required. Spherical harmonic decomposition unfolds an irregular three-dimensional field distribution into a linear combination of a series of orthogonal spherical harmonic functions. During the decomposition process, the order of expansion needs to be determined based on the spatial scale and detailed features of the elevation difference field. Here, the first sixteen spherical harmonic coefficients are selected as characteristic parameters because low-order spherical harmonic coefficients (orders 1-4) mainly reflect the overall trend of the elevation difference field, mid-order spherical harmonic coefficients (orders 5-16) reflect local detailed changes, and high-order coefficients (orders above 16) are mostly noise interference, which can improve the accuracy of subsequent calculations.

[0036] The three-dimensional offset vector of each grid node is calculated based on the extracted first sixteen spherical harmonic coefficients. The first-order spherical harmonic coefficient corresponds to the overall tilt direction of the elevation difference field, and its vector direction determines the direction of the three-dimensional offset vector. The second-order spherical harmonic coefficient corresponds to the curvature variation and amplitude distribution of the elevation difference field, and its value is positively correlated with the amplitude of the three-dimensional offset vector. A preset scaling factor can be used to convert the second-order spherical harmonic coefficients into specific offset amplitude values. Ultimately, the three-dimensional offset vector of each grid node contains two key parameters: direction and amplitude, providing a positional basis for the subsequent generation of beam direction adjustment commands.

[0037] In step S4, the specific process for processing the spectral composition data is as follows: Spectral composition data processing also uses a full-color phase distribution map as input. The first step is to separate the Stokes parameters for each wavelength. Stokes parameters are four-dimensional parameters describing the polarization characteristics of light, containing four polarization components: total intensity (H0), the intensity difference between horizontally and vertically linearly polarized light (H1), the intensity difference between 45-degree and 135-degree linearly polarized light (H2), and the intensity difference between right-handed and left-handed circularly polarized light (H3). During separation, the red, green, and blue channels of the full-color phase distribution map need to be processed separately: for each channel, an image polarization analysis algorithm is used, combined with the detector's polarization imaging characteristics (if the detector has no built-in polarizer, multiple images need to be acquired by rotating an external polarizer), to calculate the values ​​of the four components H0, H1, H2, and H3 for each grid node, forming an independent Stokes parameter matrix for each wavelength.

[0038] Once the Stokes parameters for each wavelength are determined, their Mueller matrix representations need to be calculated. The Mueller matrix is ​​a 4×4 real matrix describing the polarization state changes of light during propagation or reflection; its elements are derived through transformation relationships of the Stokes parameters. Specifically, standard polarization state incident light (such as horizontally linearly polarized light, vertically linearly polarized light, etc.) is selected, and the outgoing Stokes parameters after reflection from the shelf surface are recorded. Then, the elements of the Mueller matrix are solved using a system of linear equations. The Mueller matrix can fully characterize the polarization modulation characteristics of the shelf surface for specific wavelengths of light, and the singular values ​​in its matrix structure reflect the intensity and stability of the polarization state modulation.

[0039] By performing singular value decomposition on the Mueller matrix, three non-zero singular values ​​are obtained (the Mueller matrix has a rank of 3). The distribution characteristics of these singular values ​​are quantitatively correlated with the depolarization coefficient. The depolarization coefficient is a parameter that measures the degree of polarization degradation of light after reflection. The smaller the difference in the singular values ​​of the Mueller matrix, the larger the depolarization coefficient, indicating that the polarization state of the reflected light is closer to that of natural light; conversely, the larger the difference in singular values, the smaller the depolarization coefficient, indicating that the reflected light still retains strong polarization characteristics. The calculation of the depolarization coefficient needs to be based on the statistical characteristics of the singular values. A pre-defined mathematical model is used to transform the singular values ​​into specific depolarization coefficient values, thereby forming a depolarization coefficient distribution covering the entire shelf surface.

[0040] There is a clear correlation between the depolarization coefficient distribution and the surface roughness of the shelf. This is because surface roughness directly affects the scattering characteristics of reflected light: the rougher the surface, the wider the scattering angle distribution of the reflected light, the more severe the polarization state degradation, and the larger the depolarization coefficient; the smoother the surface, the more specular reflection the reflected light undergoes, the better the polarization state is preserved, and the smaller the depolarization coefficient. This correlation needs to be established through preliminary experimental calibration. During the calibration process, standard samples with different roughnesses (such as Ra=0.1μm, 0.5μm, 1μm, etc.) are used to measure the depolarization coefficient of each sample, construct a fitting curve between roughness and depolarization coefficient, and form a correlation model.

[0041] Based on the correspondence between the depolarization coefficient and surface roughness, the surface bidirectional reflectance distribution function (BRDF) at each grid node can be calculated. The BRDF describes the relationship between the reflected light intensity and the incident light intensity, incident angle, and reflection angle when light is reflected from an object's surface. Its value directly reflects the surface's ability to reflect light of a specific wavelength. During the calculation, the incident light angle parameter (determined by the projection system parameters in S1) and the roughness data corresponding to the depolarization coefficient are combined and substituted into a preset BRDF model (such as the Phong model, Cook-Torrance model, etc.) to obtain the BRDF value for each grid node at the corresponding wavelength.

[0042] To obtain the light intensity difference, the calculated bidirectional reflectance distribution function (BRDF) needs to be compared with a reference BRDF. The reference BRDF is the preset value of the shelf under standard lighting conditions (such as target lighting intensity and standard display status), which is measured and stored through previous calibration experiments. The comparison process involves subtracting the reference BRDF value from the actual BRDF value of each grid node to obtain the reflectance change at each wavelength. The product of the reflectance change and the reference light intensity value is the light intensity difference, where the reference light intensity value is the system's preset target lighting intensity, usually set according to shelf lighting standards (e.g., 500-800 lux for shopping mall shelf lighting).

[0043] Finally, the numerical values ​​and signs of the light intensity differences at each wavelength need to be recorded. The magnitude of the value reflects the degree to which the light intensity deviates from the target value, while the sign reflects the direction of deviation: a positive sign indicates that the reflected light intensity at that node is higher than the reference light intensity, and a negative sign indicates that the reflected light intensity is lower than the reference light intensity. This data will be directly used to generate subsequent spectral compensation commands, ensuring that the compensation operation can accurately correct the light intensity deviation and achieve spectral-level illumination optimization.

[0044] In step S5, the specific process of generating the beam direction adjustment command for the basic lighting system based on the direction and magnitude of the offset vector is as follows: A mapping relationship between the three-dimensional offset vector and the illumination beam direction is established. This mapping relationship is constructed based on the orthogonal decomposition of spherical harmonic coefficients. Previously, S3 extracted the first sixteen spherical harmonic coefficients, where the first-order spherical harmonic coefficients characterize the overall directional features of the three-dimensional offset vector, and the second-order spherical harmonic coefficients characterize the spatial distribution of the offset amplitude. Together, they constitute the core parameters of the mapping relationship. The orthogonal decomposition process is completed by substituting the first sixteen spherical harmonic coefficients into the orthogonal basis function set. The orthogonality of the spherical harmonic functions ensures that each three-dimensional offset vector corresponds to a unique illumination beam direction parameter, avoiding confusion in direction mapping under different offset states.

[0045] After establishing the mapping relationship, the three-dimensional offset vector needs to be projected onto the azimuth and zenith dimensions of the spherical coordinate system. The spherical coordinate system uses the installation origin of the basic lighting system as its pole. The azimuth angle is defined as the angle between the light beam on the horizontal plane and a preset reference direction (such as the direction of the main aisle of the shelf), and the zenith angle is defined as the angle between the light beam and the horizontal plane. The projection process is achieved through vector dot product operations. The three-dimensional offset vector is multiplied by the unit vectors corresponding to the azimuth and zenith angles in the spherical coordinate system to obtain the projection components in the two dimensions. The reference azimuth and reference zenith angles are preset standard lighting direction parameters of the system and need to be pre-calibrated according to the shelf layout and lighting standards. For example, when the shelves are arranged along the north-south direction, the reference azimuth angle can be set to 0 degrees (pointing due north), and the reference zenith angle can be set to 45 degrees (ensuring that the light covers the middle layer of the shelf). By calculating the difference between the projection components and the corresponding reference values, the horizontal deflection adjustment and pitch adjustment can be obtained respectively. A positive difference indicates that adjustment needs to be made in the positive direction of that dimension, and a negative difference indicates adjustment in the negative direction.

[0046] The calculation of the beam divergence angle adjustment is based on the amplitude component of the spherical harmonic function coefficients. The amplitude component of the spherical harmonic function coefficients reflects the distribution range of the three-dimensional offset vector in space. The larger the amplitude, the more concentrated the height deviation of the shelf surface in that area, requiring a smaller beam divergence angle to achieve precise local illumination; the smaller the amplitude, the more dispersed the deviation distribution, requiring a larger beam divergence angle to cover a wider area. In the calculation process, the statistical mean of the amplitude component among the first sixteen spherical harmonic coefficients is selected as the core parameter. Through the previously calibrated "mean amplitude - divergence angle adjustment" curve, the mean amplitude is converted into a specific beam divergence angle adjustment. For example, a mean amplitude of 0.5 corresponds to a 2-degree increase in the divergence angle, and a mean amplitude of 1.2 corresponds to a 3-degree decrease in the divergence angle.

[0047] After calculating the three adjustment values, they need to be encoded into beam direction adjustment commands that the basic lighting system can resolve. The encoding uses a system-defined binary frame structure. The frame header contains a command type identifier (used to distinguish beam direction commands from spectral compensation commands). The frame body is divided into three fields, corresponding to the horizontal deflection adjustment, pitch adjustment, and beam divergence angle adjustment, respectively. Each field contains a numerical bit and a sign bit. The numerical bit uses 16 bits to represent the absolute value of the adjustment (in degrees), and the sign bit uses 1 bit to represent the direction (0 for positive, 1 for negative). A check bit is set at the end of the frame to ensure the accuracy of the command transmission. The encoded command is transmitted through the communication interface between the control unit and the basic lighting system (such as a CAN bus or industrial Ethernet), awaiting a response from the actuator.

[0048] In step S5, the specific process of generating the spectral compensation command for the basic lighting system based on the light intensity difference is as follows: A correspondence between light intensity difference and spectral power distribution is established. This correspondence is constructed based on the eigenvector analysis of the Mueller matrix. Previously, S4 obtained the depolarization coefficient distribution through singular value decomposition of the Mueller matrix, while the eigenvectors of the Mueller matrix further reflect the polarization state evolution of light at different wavelengths during reflection. In the eigenvector analysis process, three non-zero eigenvectors of the Mueller matrix are extracted, each corresponding to a spectral response mode for a wavelength. By performing correlation calculations between the light intensity difference and the eigenvectors, a mapping model between the light intensity difference and the spectral power at that wavelength is established, ensuring that the light intensity deviation can be accurately converted into the adjustment requirements of the spectral power.

[0049] The intensity differences of each wavelength are input into the spectral reconstruction algorithm. This algorithm employs a linear superposition reconstruction model, using a pre-set reference spectral power distribution as the initial template. It then combines this with the power adjustment requirements corresponding to the intensity differences to output the target spectral power distribution. The reference spectral power distribution must meet the color temperature and color rendering index requirements of shelf lighting. For example, supermarket shelf lighting typically requires a neutral color temperature of 4000K and a color rendering index above 80; the reference distribution is set according to this standard. During the spectral reconstruction process, the algorithm adjusts the power proportion of the corresponding wavelength based on the sign of the intensity difference. For instance, when the intensity difference of the red wavelength is negative (lower than the reference), the algorithm increases the power proportion of red light in the target spectrum; conversely, it decreases it. The final output target spectral power distribution must simultaneously meet the requirements of intensity balance and chromaticity parameters.

[0050] The relative difference between the target spectral power distribution and the reference spectral power distribution is calculated. This relative difference is calculated as (target power - reference power) / reference power, and the result reflects the relative adjustment ratio of the power at each wavelength. Based on this relative difference, the intensity adjustment coefficient for each wavelength component is calculated. The intensity adjustment coefficient is equal to the product of the relative difference and the correction coefficient, where the correction coefficient is proportional to the singular value of the Mueller matrix. The singular value of the Mueller matrix characterizes the polarization modulation intensity of light at that wavelength. A larger singular value indicates that the polarization state of the reflected light at that wavelength is more significantly affected by surface roughness, requiring a larger correction coefficient to ensure that the intensity adjustment can offset the intensity deviation caused by polarization. For example, when the singular value of the Mueller matrix for blue light wavelength is 0.8, the correction coefficient is set to 1.2; when the singular value is 0.5, the correction coefficient is set to 0.9.

[0051] The intensity adjustment factor is converted into a drive current adjustment based on the current-luminous intensity characteristic curve of the light-emitting element. Shelf lighting systems typically use RGB tri-color LEDs as light-emitting elements. The current-luminous intensity curves of different colored LEDs exhibit a linear relationship. For example, for red LEDs in the 200mA to 500mA range, the linear correlation coefficient between luminous intensity and current is greater than 0.98. During the conversion process, the required change in luminous intensity is calculated based on the intensity adjustment factor. Then, through the inverse operation of the current-luminous intensity curve, the change in luminous intensity is converted into a drive current adjustment. For example, if the intensity adjustment factor for red light is 0.2 (requiring a 20% increase in luminous intensity), the corresponding current adjustment is 40mA (increasing from 200mA to 240mA).

[0052] Finally, the drive current adjustment amounts corresponding to each wavelength are encoded into spectral compensation commands. The commands employ a binary frame structure compatible with beam direction commands. The frame header contains a unique command type identifier, and the frame body is divided into three fields according to wavelength order (red, green, blue). Each field contains the numerical value, sign bit, and wavelength identifier bit for the drive current adjustment amount at that wavelength. A check bit is also set at the end of the frame. The encoded commands are transmitted to the drive module of the basic lighting system through the control unit's communication interface, providing a basis for subsequent current adjustment.

[0053] In S6, the specific process by which the basic lighting system adjusts the lighting parameters according to the beam direction adjustment command and the spectral compensation command is as follows: The basic lighting system first parses the beam direction adjustment command. The command parsing module built into the control unit breaks down the command according to a preset frame structure, extracting the values ​​and signs of the horizontal deflection adjustment, pitch adjustment, and beam divergence angle adjustment. During the parsing process, the module first verifies the frame end check bit to ensure that the command has not been erroneous during transmission. If the verification passes, the three adjustment values ​​are converted into control signals that the actuator can recognize. If the verification fails, the command is requested to be retransmitted.

[0054] The horizontal deflection mechanism of the lighting unit is adjusted according to the horizontal deflection adjustment amount. This mechanism consists of a stepper motor and a gear transmission assembly. The rotation angle of the stepper motor is proportional to the horizontal deflection adjustment amount. This proportion needs to be determined through prior calibration. For example, every 1 degree of horizontal deflection adjustment corresponds to 100 steps of stepper motor rotation. The control unit outputs pulse signals to the stepper motor controller. The number of pulses is calculated from the adjustment amount and the proportional coefficient. The motor drives the lighting unit to rotate around the vertical axis through gear transmission, realizing the horizontal beam deflection. During the adjustment process, the built-in angle sensor will provide real-time feedback on the actual rotation angle and compare it with the target adjustment amount. Adjustment stops when the deviation is less than 0.1 degrees.

[0055] The tilt adjustment mechanism of the lighting unit adjusts according to the tilt adjustment amount. This mechanism is also driven by a stepper motor, but its rotation axis is a horizontal axis, used to change the vertical illumination angle of the beam. The rotation angle of the tilt adjustment mechanism is also proportional to the tilt adjustment amount, and the proportional coefficient is set according to the transmission ratio of the mechanism. For example, every 1 degree of tilt adjustment corresponds to 120 motor rotations. During the adjustment process, the control unit outputs a pulse signal according to the tilt adjustment amount, and the motor drives the lighting unit to rotate around the horizontal axis. At the same time, the vertical angle sensor collects the actual angle in real time and performs closed-loop control with the target adjustment amount to ensure that the final tilt angle deviation is controlled within 0.1 degrees, avoiding the displacement of the illuminated area due to angle deviation.

[0056] The optical focusing system of the illumination unit adjusts the beam divergence angle based on the adjustment amount. This system consists of two sets of movable lenses and an electrically adjustable assembly. The change in the lens spacing is directly proportional to the adjustment amount of the beam divergence angle. When it is necessary to increase the divergence angle, the electrically adjustable assembly (usually a piezoelectric ceramic actuator) pushes the two sets of lenses away from each other, increasing the lens spacing; when it is necessary to decrease the divergence angle, the adjustable assembly pulls the lenses closer together, decreasing the lens spacing. The proportional relationship between the lens spacing and the divergence angle needs to be determined through optical simulation and experimental calibration. For example, every 0.5 mm increase in lens spacing corresponds to a 1 degree increase in divergence angle, and every 0.5 mm decrease corresponds to a 1 degree decrease in divergence angle. During the adjustment process, the lens position sensor provides real-time feedback of the spacing data, and the control unit fine-tunes the adjusting assembly based on the feedback value until the divergence angle reaches the target value.

[0057] After beam direction adjustment is completed, the basic lighting system analyzes the spectral compensation command. The command analysis module extracts the drive current adjustment amount, symbol, and wavelength identifier for each wavelength according to the frame structure. After verifying that the check bits are correct, it allocates the current adjustment requirements for each wavelength to the corresponding LED driver circuit. Each wavelength of LED is equipped with an independent constant current drive circuit. The control unit outputs an analog voltage signal (or digital control signal) to the drive circuit, and the drive circuit adjusts the output current according to the signal. The change in current is strictly proportional to the drive current adjustment amount. For example, when the red light drive current adjustment amount is 40mA (positive adjustment), the drive circuit increases the output current from 200mA to 240mA; when the blue light adjustment amount is -20mA (negative adjustment), the drive circuit reduces the current from 220mA to 200mA. During the current adjustment process, the current sampling resistor built into the drive circuit monitors the output current in real time, forming a closed-loop control to ensure that the current accuracy error is less than 1mA.

[0058] All adjustments must be performed in the strict order of "direction adjustment first, then spectral adjustment." The core reason for this sequence is that adjusting the beam direction changes the incident position and angle of light on the shelf surface, directly affecting the distribution of reflected light intensity. If spectral adjustment is performed first, subsequent direction adjustments may lead to new light intensity deviations, requiring repeated adjustments. Completing direction adjustment first ensures that the lighting position and angle meet requirements, and then spectral compensation corrects for light intensity deviations, enabling optimization of lighting parameters in one step, reducing the number of adjustments, and improving system response efficiency.

[0059] Furthermore, the projection of a multi-wavelength grid detection pattern onto the shelf surface is accomplished by an auxiliary detection unit, which mainly consists of a light source assembly and optical sensors. The core components of the light source assembly include an RGB three-color narrow-linewidth laser, an acousto-optic modulator, a non-common-path interference structure, and diffractive optical elements; its hardware configuration is completely matched to the projection system described in S1. To ensure that the detection pattern aligns with the illumination area of ​​the basic lighting system, the light output direction of the light source assembly must be consistent with that of the basic lighting system. This is achieved by fixing both to the same mounting plane using a mechanical bracket and calibrating the light output axis using a laser collimator, ensuring that the deviation between the light output axes of the two systems within a 10-meter range is less than 0.5 degrees, thus ensuring that the detection pattern accurately covers the illumination area of ​​the basic lighting.

[0060] The optical sensor of the auxiliary detection unit establishes a stable data connection with the control unit of the basic lighting system. The connection method adopts industrial Ethernet (such as Profinet or EtherNet / IP), with a communication rate of no less than 100Mbps to meet the real-time transmission requirements of reflected light signals. During the data connection process, the IP address and communication protocol of the sensor need to be configured in the control unit to ensure that the two can achieve bidirectional data interaction. The optical sensor can transmit the collected reflected light signal data (including spatial distribution data and spectral composition data) to the control unit, and the control unit can send acquisition trigger signals and parameter configuration commands (such as exposure time and acquisition frame rate) to the sensor.

[0061] During the detection phase, the workflow of the auxiliary detection unit is uniformly scheduled by the control unit. The control unit first sends a start command to the light source assembly. The light source assembly generates multi-wavelength coherent light according to the parameters in S1. After interference structure and phase modulation, the light is projected onto the shelf surface through diffractive optical elements to form a grid detection pattern. Simultaneously with the projection start, the control unit sends a acquisition trigger signal to the optical sensor. The sensor, according to the requirements in S2, acquires the corresponding interference fringe image sequence during the sequential projection of each wavelength. During acquisition, the sensor compresses the image data in real time (using a lossless compression algorithm) and transmits it to the control unit. The control unit temporarily stores the data in a buffer, awaiting retrieval by the signal processing unit.

[0062] Upon entering the lighting adjustment phase, the control unit of the basic lighting system first sends the reflected light signal data from the buffer area to the signal processing unit, triggering the data analysis processes S3 and S4. After calculating the three-dimensional offset vector and the light intensity difference, the signal processing unit feeds the results back to the control unit. Based on these results, the control unit generates the beam direction adjustment command and spectral compensation command in S5, and then sends the commands to the actuator of the basic lighting system, triggering the parameter adjustment operation in S6. After the adjustment operation is completed, the control unit sends a secondary detection command to the auxiliary detection unit. The auxiliary detection unit repeats the detection process, collecting the adjusted reflected light signal. The control unit determines whether the lighting parameters meet the standards based on the secondary detection results. If not, the adjustment process is repeated until the illumination uniformity and light intensity meet the preset standards, forming a complete closed-loop control.

[0063] Please see Figure 2 As shown, the present invention also includes a multi-objective optimization collaborative control device for shelf lighting, used to implement the above-described multi-objective optimization collaborative control method for shelf lighting, comprising: The light source assembly is used to project a grid detection pattern containing multi-wavelength components onto the shelf surface. The node spacing of the grid detection pattern is proportional to the shelf shelf spacing. An optical sensor is used to collect the reflected light signal from the shelf surface to the grid detection pattern, and simultaneously record the spatial distribution data and spectral composition data of the reflected light signal; The signal processing unit is used to process spatial distribution data, extract the position coordinates of each node in the grid detection pattern, calculate the offset vector of each node relative to the reference position; and process spectral composition data, decompose the wavelength components of the reflected light signal, and calculate the difference between the light intensity value of each wavelength component and the reference light intensity value. The control unit is used to generate beam direction adjustment commands for the basic lighting system based on the direction and magnitude of the offset vector, and to generate spectral compensation commands for the basic lighting system based on the light intensity difference. The basic lighting system is used to adjust lighting parameters according to beam direction adjustment commands and spectral compensation commands.

[0064] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A multi-objective optimization coordinated control method for shelf lighting, characterized in that, The method comprises the following steps: S1, projecting a grid detection pattern containing multiple wavelength components onto the surface of the shelf, the node spacing of the grid detection pattern being in proportional relationship with the shelf layer spacing; S2, collecting the reflected light signal of the grid detection pattern on the surface of the shelf, and synchronously recording the spatial distribution data and the spectral composition data of the reflected light signal; S3, processing the spatial distribution data, extracting the position coordinates of each node in the grid detection pattern, and calculating the offset vector of each node relative to the reference position; S4, processing the spectral composition data, decomposing the wavelength components of the reflected light signal, and calculating the difference between the light intensity value of each wavelength component and the reference light intensity value; S5, generating the light beam direction adjustment instruction of the basic lighting system according to the direction and amplitude of the offset vector, and generating the spectral compensation instruction of the basic lighting system according to the light intensity difference; S6, the basic lighting system adjusts the lighting parameters according to the light beam direction adjustment instruction and the spectral compensation instruction.

2. The multi-objective optimization coordinated control method for shelf lighting according to claim 1, characterized in that, In the S1, the process of projecting a grid detection pattern containing multiple wavelength components onto the surface of the shelf is as follows: Three coherent lights of red wavelength, green wavelength and blue wavelength are generated, and the three coherent lights are input into a non-common-path interference structure, a multi-wavelength interference field is generated at the output end of the interference structure, dynamic phase modulation is applied to the multi-wavelength interference field, and the frequency of the dynamic phase modulation is consistent with the driving frequency of the acousto-optic modulator corresponding to each wavelength; The phase-modulated interference field is converted into a grid light intensity distribution through a diffractive optical element, the grid light intensity distribution is projected onto the surface of the shelf to form a grid detection pattern, the phase delay of the acousto-optic modulator is changed synchronously during the projection process, the change range of the phase delay is from zero to 2π, and the grid detection pattern light intensity distribution corresponding to each phase delay is recorded.

3. The multi-objective optimization coordinated control method for shelf lighting according to claim 1, characterized in that, In the S2, the specific process of collecting the reflected light signal is as follows: A monochromatic area array detector is used to collect the interference fringe images of the surface of the shelf, interference fringe image sequences corresponding to the wavelengths are collected during the projection of the red wavelength, green wavelength and blue wavelength in turn, each wavelength interference fringe image sequence contains four interference images with a phase difference of π / 2, four-step phase shift calculation is performed on each wavelength interference fringe image sequence to obtain the phase principal value distribution of the wavelength; Phase unwrapping operation is performed on the phase principal value distribution to obtain the continuous phase distribution of the wavelength, and the continuous phase distributions of the three wavelengths are combined into a full-color phase distribution graph, which contains spatial distribution data and spectral composition data.

4. The multi-objective optimization coordinated control method for shelf lighting according to claim 1, characterized in that, In the S3, the specific process of processing the spatial distribution data is as follows: The complex amplitude information of each grid node is extracted from the full-color phase distribution graph, the complex amplitude information contains amplitude component and phase component, the coherence matrix of the complex amplitudes of adjacent grid nodes is calculated, the polarization state distribution is obtained through eigenvalue decomposition of the coherence matrix, the mapping relationship between the polarization state distribution and the surface normal vector is established, the surface normal field is reconstructed according to the mapping relationship, the surface height distribution is obtained by performing the surface area integration on the surface normal field, and the elevation difference field between the surface height distribution and the reference height distribution is calculated; The elevation difference value field is subjected to spherical harmonic function decomposition, and the first sixteen order spherical harmonic coefficients are extracted; a three-dimensional offset vector of each grid node is calculated according to the spherical harmonic coefficients, the direction of the three-dimensional offset vector is determined by the first order spherical harmonic coefficient, and the amplitude of the three-dimensional offset vector is determined by the second order spherical harmonic coefficient.

5. The multi-objective optimization coordinated control method for shelf lighting according to claim 1, wherein, In the S4, the specific process of processing the spectral composition data is as follows: The Stokes parameters of each wavelength are separated from the full-color phase distribution map, the Stokes parameters include four polarization components, the Mueller matrix representation of the Stokes parameters of each wavelength is calculated, the depolarization coefficient distribution is obtained by singular value decomposition of the Mueller matrix, the correspondence between the depolarization coefficient and the surface roughness is established, and the surface bidirectional reflectance distribution function is calculated according to the correspondence; The bidirectional reflectance distribution function is compared with the reference bidirectional reflectance distribution function to obtain the reflectivity change of each wavelength, the light intensity difference value of each wavelength is calculated according to the reflectivity change, the light intensity difference value is the product of the reflectivity change and the reference light intensity value, and the numerical value and the sign of the light intensity difference value of each wavelength are recorded.

6. The multi-objective optimization coordinated control method for shelf lighting according to claim 1, wherein, In the S5, the specific process of generating the light beam direction adjustment instruction of the basic lighting system according to the direction and amplitude of the offset vector is as follows: A mapping relationship between the three-dimensional offset vector and the lighting beam direction is established, the mapping relationship is based on the orthogonal decomposition of the spherical harmonic function coefficients, the three-dimensional offset vector is projected into the azimuth angle dimension and the zenith angle dimension in the spherical coordinate system, the difference between the projection component of the azimuth angle dimension and the reference azimuth angle is calculated to obtain the horizontal deflection adjustment amount; The difference between the projection component of the zenith angle dimension and the reference zenith angle is calculated to obtain the pitch adjustment amount; The light beam divergence angle adjustment amount is calculated according to the amplitude component of the spherical harmonic function coefficients, the horizontal deflection adjustment amount, the pitch adjustment amount and the light beam divergence angle adjustment amount are encoded into the light beam direction adjustment instruction, and the light beam direction adjustment instruction includes the numerical value and the sign of each adjustment amount.

7. The multi-objective optimization coordinated control method for shelf lighting according to claim 1, wherein, In the S5, the specific process of generating the spectral compensation instruction of the basic lighting system according to the light intensity difference value is as follows: A corresponding relationship between the light intensity difference value and the spectral power distribution is established, the corresponding relationship is based on the eigenvector analysis of the Mueller matrix, the light intensity difference value of each wavelength is input into a spectral reconstruction algorithm, the spectral reconstruction algorithm outputs a target spectral power distribution, the relative difference between the target spectral power distribution and the reference spectral power distribution is calculated, the intensity adjustment coefficient of each wavelength component is calculated according to the relative difference, and the intensity adjustment coefficient is proportional to the singular value of the Mueller matrix; The intensity adjustment coefficient is converted into a driving current adjustment amount, the driving current adjustment amount is linearly related to the intensity adjustment coefficient, and the driving current adjustment amount corresponding to each wavelength is encoded into the spectral compensation instruction.

8. The multi-objective optimization coordinated control method for shelf lighting of claim 1, wherein, In the S6, the specific process of adjusting the lighting parameters of the basic lighting system according to the light beam direction adjustment instruction and the spectral compensation instruction is as follows: The horizontal deflection adjustment amount, the pitch adjustment amount and the beam divergence angle adjustment amount in the light beam direction adjustment instruction are analyzed, the horizontal deflection mechanism of the lighting unit is adjusted according to the horizontal deflection adjustment amount, the rotation angle of the horizontal deflection mechanism is proportional to the horizontal deflection adjustment amount, the pitch adjustment mechanism of the lighting unit is adjusted according to the pitch adjustment amount, the rotation angle of the pitch adjustment mechanism is proportional to the pitch adjustment amount, and the optical focusing system of the lighting unit is adjusted according to the beam divergence angle adjustment amount, and the lens spacing variation of the optical focusing system is proportional to the beam divergence angle adjustment amount The wavelength drive current adjustment amount in the spectral compensation instruction is analyzed, the drive current of each wavelength light emitting element of the lighting unit is adjusted according to the drive current adjustment amount, the variation of the drive current is proportional to the drive current adjustment amount, and each adjustment operation is executed in the order of direction adjustment and spectral adjustment.

9. The multi-objective optimization coordinated control method for shelf lighting according to claim 1, wherein, The grid detection pattern containing multiple wavelength components is projected onto the surface of the shelf based on an auxiliary detection unit, the light source assembly of the auxiliary detection unit is in the same direction as the light source assembly of the basic lighting system, and the optical sensor of the auxiliary detection unit is connected to the control unit of the basic lighting system. In the detection stage, the light source assembly of the auxiliary detection unit projects the grid detection pattern, and the optical sensor collects the reflected light signal. In the lighting adjustment stage, the control unit of the basic lighting system generates the lighting parameter adjustment instruction according to the processing result of the reflected light signal.

10. A multi-objective optimization coordinated control device for shelf lighting for implementing the multi-objective optimization coordinated control method for shelf lighting according to any one of claims 1 to 9, characterized in that, Comprise: A light source assembly is used to project a grid detection pattern containing multiple wavelength components onto the surface of the shelf, and the node spacing of the grid detection pattern is proportional to the shelf layer spacing; An optical sensor is used to collect the reflected light signal of the grid detection pattern on the surface of the shelf, and to record the spatial distribution data and spectral composition data of the reflected light signal synchronously; A signal processing unit is used to process the spatial distribution data, extract the position coordinates of each node in the grid detection pattern, and calculate the offset vector of each node relative to the reference position; And process the spectral composition data, decompose the wavelength components of the reflected light signal, and calculate the difference between the light intensity value of each wavelength component and the reference light intensity value; A control unit is used to generate a light beam direction adjustment instruction of the basic lighting system according to the direction and amplitude of the offset vector, and to generate a spectral compensation instruction of the basic lighting system according to the light intensity difference; A basic lighting system is used to adjust the lighting parameters according to the light beam direction adjustment instruction and the spectral compensation instruction.

Images