A projection adjustment method and vehicle

Abstract

The application discloses a projection adjustment method and a vehicle, and is applied to the technical field of intelligent cockpits. The method comprises the following steps: collecting a sensing image and separating ambient light to obtain medium characteristic data reflecting a projection target area; dividing the projection target area into at least two subareas based on the medium characteristic data; for each subarea, independently generating a projection adjustment parameter according to the corresponding medium characteristic data, wherein the parameter comprises at least one of color compensation, brightness compensation, contrast compensation and geometric deformation compensation; and adjusting the image content of the corresponding subarea in the original image based on the projection adjustment parameter of each subarea to generate an adjusted projection image. Through real-time sensing of the medium characteristics and independent compensation according to the characteristics, the projection image can be kept clear and accurate on different material surfaces.

Description

Technical Field

[0001] This application relates to the field of smart cockpit technology, specifically to a projection adjustment method and a vehicle. Background Technology

[0002] With the increasing intelligence of automobiles, intelligent headlight projection is widely used to project graphic information onto surfaces such as the ground and walls outside the vehicle to assist driving and facilitate human-vehicle interaction. Unlike indoor projection environments, the external surface of vehicle headlight projection is not a fixed screen, but rather a constantly changing medium such as road surfaces and walls, whose color, material, and geometry are complex and uncontrollable, affecting image quality. For this unique environment, existing technologies generally employ factory calibration or preset compensation parameters based on common media. These solutions lack the ability to perceive changes in the current medium in real time. Therefore, when the vehicle travels over areas with different materials, the inability to dynamically correct projection parameters or insufficient adjustment precision can easily lead to problems such as color cast, excessive darkness, or distortion, which can severely impair the driver's ability to recognize the projected information.

[0003] Therefore, improving the projection quality of vehicle headlights has become a pressing technical problem to be solved in this field. Summary of the Invention

[0004] To address the aforementioned technical problems, this application is proposed. Embodiments of this application provide a projection adjustment method and a vehicle that, by real-time sensing and independent compensation of the optical or geometric characteristics of the projection medium, solves the problem of projection image quality distortion caused by differences in projection medium characteristics, ensuring that the projected image remains clear and accurate on various material surfaces.

[0005] According to a first aspect of this application, a projection adjustment method is provided for image projection of vehicle headlights. The method includes: acquiring a perceived image of a current projection target area and separating ambient light to obtain medium characteristic data reflecting the current projection target area; the medium characteristic data includes at least one of optical characteristics and geometric characteristics; based on the medium characteristic data, dividing the current projection target area into at least two sub-regions; for each sub-region, independently generating projection adjustment parameters according to the corresponding medium characteristic data; wherein the projection adjustment parameters include at least one of color compensation parameters, brightness compensation parameters, contrast compensation parameters, and geometric deformation compensation parameters; and adjusting the image content of the corresponding sub-region in the original image based on the projection adjustment parameters of each sub-region to generate an adjusted projection image.

[0006] As one possible implementation, dividing the current projection target region into at least two sub-regions based on the medium characteristic data includes: determining the color difference and texture difference within the current projection target region based on the medium characteristic data; and dividing the current projection target region into multiple non-overlapping homogeneous sub-regions based on the color difference and texture difference.

[0007] By comprehensively utilizing color and texture differences as segmentation criteria, the accuracy of sub-region division in multi-material mixed scenes is improved, reducing the possibility of missegmentation or omission due to relying on a single feature, thereby enhancing the accuracy of subsequent independent compensation for each sub-region.

[0008] As one possible implementation, acquiring a perceived image of the current projection target area and separating ambient light to obtain medium characteristic data reflecting the current projection target area includes: acquiring a first image without a compensation pattern being projected, the first image containing only ambient light; acquiring a second image with the compensation pattern projected, the second image containing both projection light and ambient light; obtaining a projection light reflection image based on the difference between the second image and the first image; and calculating the optical characteristics of the medium based on the projection light reflection image, a preset incident light intensity, and an incident angle obtained by a sensor; wherein the optical characteristics include at least the medium color and diffuse reflectance.

[0009] By successively acquiring a first image containing only ambient light and a second image containing both projection light and ambient light, and then subtracting them, a pure projection light reflection image is obtained. This improves the accuracy of extracting medium color and diffuse reflectance from mixed lighting, reduces the interference of ambient light changes on the estimation of medium characteristics, and provides a more reliable data foundation for color and brightness compensation.

[0010] As one possible implementation, the medium characteristic data further includes a specular mask; the method further includes: performing brightness threshold and saturation threshold detection on the perceived image, determining regions with brightness exceeding a first preset threshold and saturation lower than a second preset threshold as specular regions, and generating a specular mask corresponding to the specular regions; wherein, for each sub-region, generating projection adjustment parameters independently based on the corresponding medium characteristic data includes: when generating the projection adjustment parameters, masking pixels within the specular mask.

[0011] By generating a specular mask through dual-threshold detection and shielding the pixels within the specular mask when calculating projection adjustment parameters, the interference of specular reflection specular highlights on the average reflectivity and color statistics of the medium is reduced, improving the rationality of compensation parameter calculation and thus avoiding overcompensation caused by specular highlights.

[0012] As one possible implementation, the projection adjustment parameters also include edge enhancement parameters; the step of independently generating projection adjustment parameters for each sub-region based on the corresponding medium characteristic data includes: calculating the texture energy of the current projection target region based on the medium characteristic data; determining the edge enhancement intensity based on the preset intensity level corresponding to the texture energy; the lower the texture energy, the higher the edge enhancement intensity.

[0013] By calculating texture energy to quantify the roughness of the medium surface and adaptively determining the edge enhancement intensity accordingly, the edge sharpness and detail clarity of the projected pattern on rough surfaces are improved, while reducing the risk of noise caused by over-sharpening smooth surfaces.

[0014] As one possible implementation, the geometric deformation compensation parameters are generated by: acquiring the three-dimensional point cloud data of the current projection target area; constructing a homography transformation matrix from the projector coordinate system to the coordinate system of the current projection target area by combining the vehicle suspension pitch angle and wheel rotation angle; and generating the geometric deformation compensation parameters based on the homography transformation matrix.

[0015] By acquiring 3D point cloud data and combining it with real-time vehicle attitude parameters such as suspension pitch angle and wheel rotation angle to construct a homography transformation matrix, real-time geometric pre-correction without projection compensation pattern is achieved, which improves the geometric regularity of the projected image on dynamically changing rugged or inclined surfaces and reduces trapezoidal distortion caused by changes in surface morphology or vehicle attitude.

[0016] As one possible implementation, adjusting the image content of the corresponding sub-region in the original image based on the projection adjustment parameters of each sub-region includes: within the boundary transition zone between adjacent sub-regions, using an interpolation function with distance as the weight, mixing the projection adjustment parameters of each adjacent sub-region to obtain smooth projection adjustment parameters within the boundary transition zone.

[0017] By setting transition zones at the boundaries of adjacent sub-regions and using an interpolation function with distance as the weight to mix the adjustment parameters of adjacent regions, the brightness jumps or color blocks that may occur at the material boundaries due to independent adjustment of zones are reduced, thereby improving the visual coherence and naturalness of the projected image on multi-material surfaces.

[0018] As one possible implementation, adjusting the image content of corresponding sub-regions in the original image based on the projection adjustment parameters of each sub-region includes: performing color compensation on the pixels of the corresponding sub-regions in the original image based on the color compensation parameters corresponding to the sub-regions; adjusting the brightness or contrast of the pixels of the corresponding sub-regions in the original image based on the brightness compensation parameters or contrast compensation parameters corresponding to the sub-regions; performing geometric remapping on the sub-region images after color compensation, brightness adjustment, or contrast adjustment based on the geometric deformation compensation parameters corresponding to the sub-regions; and stitching the image content of each sub-region after geometric remapping at a preset spatial position to generate an adjusted projection image.

[0019] By independently calculating and applying its own color compensation, brightness adjustment, contrast adjustment, and geometric remapping for each sub-region, the adjustment of image content can accurately match the unique media characteristics of each sub-region. This improves the overall color accuracy, brightness uniformity, and geometric regularity of the projected image in multi-media mixed scenarios, reduces local adaptation problems caused by using a single global parameter, and enhances the adaptability to complex and ever-changing projection environments.

[0020] As one possible implementation, the method further includes: acquiring vehicle speed; when the vehicle speed is lower than a first speed threshold, performing the acquisition of the perceived image and the updating of the projection adjustment parameters at a first frequency; when the vehicle speed is between the first speed threshold and a second speed threshold, performing the acquisition of the perceived image and the updating of the projection adjustment parameters at a second frequency lower than the first frequency; when the vehicle speed is higher than the second speed threshold, stopping the acquisition of the perceived image and the updating of the projection adjustment parameters, and switching to a static projection mode; wherein, the static projection mode projects pre-stored fixed image content.

[0021] By adjusting the frequency of image acquisition and projection parameter updates according to vehicle speed, the system improves the precision and real-time tracking capability of the projection effect at low speeds, reduces unnecessary system computing power consumption and screen flicker at medium and low speeds, and reduces the impact of dynamic adjustment on system stability at high speeds, thereby improving the system's resource utilization efficiency and applicability under different driving conditions.

[0022] According to a second aspect of this application, a vehicle is provided, comprising: a vehicle headlight; and a projection adjustment system communicatively connected to the vehicle headlight, the projection adjustment system being used for the projection adjustment method as described in the first aspect or any implementation thereof.

[0023] According to a third aspect of this application, a projection adjustment device is provided, comprising: a sensing unit, configured to acquire a sensed image of a current projection target area and separate ambient light to obtain medium characteristic data reflecting the current projection target area; the medium characteristic data includes at least one of optical characteristics and geometric characteristics; a region segmentation unit, configured to segment the current projection target area into at least two sub-regions based on the medium characteristic data; a parameter generation unit, configured to independently generate projection adjustment parameters for each sub-region based on the corresponding medium characteristic data; wherein the projection adjustment parameters include at least one of color compensation parameters, brightness compensation parameters, contrast compensation parameters, and geometric deformation compensation parameters; and a projection generation unit, configured to adjust the image content of the corresponding sub-region in the original image based on the projection adjustment parameters of each sub-region, to generate an adjusted projection image.

[0024] According to a fourth aspect of this application, a computer device is provided, the computer device comprising: one or more processors; a memory; and one or more application programs, wherein the one or more application programs are stored in the memory and configured to be executed by the processor to implement the method as described in the first aspect or any implementation thereof.

[0025] According to a fifth aspect of this application, a computer-readable storage medium is provided, the storage medium storing a computer program for performing the method as described in the first aspect or any implementation thereof.

[0026] According to a sixth aspect of this application, an electronic device is provided, including a module for performing the method as described in the first aspect or any implementation thereof.

[0027] According to a seventh aspect of this application, a computer program product is provided, comprising program code for performing the method as described in the first aspect or any implementation thereof.

[0028] The projection adjustment method and vehicle provided in this application first acquire projection medium characteristic data unaffected by ambient light by collecting and sensing images and separating ambient light. This provides a reliable data foundation for subsequent adjustments, avoiding misjudgments or compensation failures caused by changes in ambient light. Second, based on this medium characteristic data, the projection area is segmented, enabling more refined operations for sub-regions of different materials, improving the flexibility and targeting of compensation. Then, compensation parameters corresponding to the medium characteristics are independently generated for each sub-region, allowing each region to receive corrections matching its surface conditions. This improves the matching degree between compensation parameters and medium characteristics, reducing image quality degradation caused by insufficient or excessive compensation. Finally, the independent adjustment parameters for each sub-region are applied to the corresponding sub-region content in the original image, resulting in targeted spatial correction of the final projected image, improving the overall accuracy and consistency of the projected image. Attached Figure Description

[0029] The above and other objects, features, and advantages of this application will become more apparent from the more detailed description of the embodiments of this application in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the embodiments of this application to explain this application and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same components or steps.

[0030] Figure 1 This is a schematic flowchart of a projection adjustment method provided in an exemplary embodiment of this application.

[0031] Figure 2 This is a schematic diagram of the projection adjustment device provided in an exemplary embodiment of this application.

[0032] Figure 3 This is a structural diagram of an electronic device provided in an exemplary embodiment of this application. Detailed Implementation

[0033] Hereinafter, exemplary embodiments according to this application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments of this application. It should be understood that this application is not limited to the exemplary embodiments described herein.

[0034] In the field of intelligent headlight image projection for vehicles, the commonly used technical solution to accurately project navigation guidance, distance warning, and light carpet patterns onto the ground or wall in front of the vehicle is to pre-store a set of projection images calibrated on a standard or ideal medium surface, or to perform uniform distortion correction offline before leaving the factory. Specifically, this solution adjusts the projected content through a fixed transformation matrix or preset brightness and color parameters. The basic working principle is to treat the target surface as a medium with consistent optical properties (such as a standard gray surface) and geometric properties (such as an ideal vertical plane), and directly project the generated content. This solution can achieve good pattern clarity and color fidelity in laboratory or calibration scenarios, and has relatively low requirements for real-time computing power on the vehicle side.

[0035] However, when this solution is applied to the varied surfaces of a vehicle during actual driving, the quality of the projected image is not high. This is because, in pursuit of minimal system response and consistent parameter calibration, the inherent static design of the solution cannot adapt to dynamic changes in the medium, easily leading to color cast, localized overexposure, geometric distortion, or loss of detail. For example, at the junction of asphalt and epoxy flooring, or in areas where wet, slippery surfaces meet dry cement surfaces, the preset global parameters cannot match the local reflectivity and color characteristics of different materials, resulting in brightness jumps or color patchwork in the projected image within the same frame, and reduced visibility of key information such as navigation arrows.

[0036] To address the aforementioned issues, a partitioning adjustment mechanism is introduced that utilizes real-time sensing of medium characteristics based on ambient light separation, region segmentation based on color and texture differences, and independent generation of multi-dimensional compensation parameters for each sub-region. This reconstructs the entire compensation chain from the original image to the final projected image, thereby improving the color accuracy, brightness uniformity, and geometric regularity of the projected image when multiple materials coexist and undergo dynamic changes. It also avoids overexposure compensation for highlights and visual splicing at partition boundaries, ensuring that the projected pattern remains clear, accurate, and stable on various material surfaces.

[0037] To implement the projection adjustment method, a projection adjustment system can be constructed. This system, applied to a vehicle, includes a projection unit, a sensing unit, and a control unit. The projection unit, employing a DLP (Digital Light Processing) optical engine module, is responsible for projecting the image content outwards. The sensing unit is used to collect information about the vehicle's surrounding environment and image data of the projection target area.

[0038] The perception unit may include a surround-view camera or a forward-view camera for capturing images of the projected target area. The perception unit may also include a combination of an ADAS (Advanced Driver Assistance Systems Camera) camera and a radar sensor, such as millimeter-wave radar or lidar, for acquiring distance and three-dimensional spatial geometry data of the projected target area. Furthermore, the perception unit may be equipped with an ambient light sensor to acquire ambient illuminance and color temperature information.

[0039] The control unit is communicatively connected to both the projection unit and the sensing unit. The control unit receives images and sensor data collected by the sensing unit, processes and analyzes them, and generates adjustment commands for the projection unit accordingly. The control unit can integrate a processing chip with AI (Artificial Intelligence) acceleration capabilities to run algorithms such as image segmentation, feature extraction, and parameter generation.

[0040] During operation, the image data collected by the sensing unit and the sensor measurements are transmitted to the control unit. Based on these inputs, the control unit calculates the medium characteristics of the current projection target area and then calculates the projection adjustment parameters for different sub-regions. The projection unit receives the adjusted image data from the control unit and projects the compensated and corrected image onto the target surface in front of the vehicle.

[0041] Figure 1 This is a schematic flowchart of a projection adjustment method provided in an exemplary embodiment of this application, applicable to image projection scenarios of vehicle intelligent headlights. The following description, in conjunction with... Figure 1 The projection adjustment method provided in the embodiments of this application will be described in detail.

[0042] In S110, a perceived image of the current projection target area is acquired and the ambient light is separated to obtain medium characteristic data reflecting the current projection target area.

[0043] The medium property data includes at least one of optical properties and geometric properties.

[0044] The current projection target area refers to the ground, wall, or other object surface currently covered by the vehicle's headlight projection beam. The perceived image can be acquired by a surround-view camera or a forward-view camera in the sensing unit.

[0045] Medium property data refers to information that characterizes the physical properties of the surface of the projected target area. Optical properties may include one or more of the following: medium color, diffuse reflectivity, specular reflectance, and texture energy; geometric properties may include the surface's three-dimensional shape, normal direction, and relative height. Medium property data can be acquired by reusing existing vehicle sensors, such as using cameras to obtain color and texture features, and using lidar or millimeter-wave radar to obtain distance and point cloud data.

[0046] In some embodiments, ambient light separation can be performed to improve the accuracy of acquiring medium property data. For example, a first image without a projection compensation pattern is acquired, the first image containing only ambient light; a second image with a projection compensation pattern is acquired, the second image containing both projection light and ambient light; a projection light reflection image is obtained based on the difference between the second image and the first image; the optical properties of the medium are calculated based on the projection light reflection image, a preset incident light intensity, and an incident angle obtained by a sensor; wherein the optical properties include at least the medium color and diffuse reflectance.

[0047] A compensation pattern is a known reference signal used to aid in calculations. It is typically a uniform, known-brightness, known-color spot that covers the entire projection area. For example, a projection unit might project a pure white, constant-brightness rectangular spot. Alternatively, it could project pure red, pure green, or pure blue spots to individually calibrate the reflectance characteristics of a specific color channel. Or it could be a checkerboard, stripe, or other pattern with known coded information, used to simultaneously acquire optical and geometric properties.

[0048] Ambient light is obtained by temporarily shutting down the projection unit. For example, the control unit instructs the projection unit to temporarily shut down, preventing it from projecting any light. Simultaneously, the control unit instructs the sensing unit (such as a camera) to capture an image containing only reflected light from external light sources like natural light or streetlights illuminating the projection target area; therefore, the first image is pure ambient light. Subsequently, the control unit instructs the projection unit to turn on and project a known compensation pattern. At the same time, the control unit instructs the sensing unit to capture another image containing two parts: projected light (the light signal reflected from the compensation pattern by the projection medium) and ambient light (the same ambient light as in the first image). Therefore, the second image is equal to the projection light superimposed on the ambient light. The control unit performs pixel-by-pixel subtraction between the second image and the first image, thus obtaining a pure image formed solely by reflections from the compensation pattern, eliminating all ambient light interference. This image directly reflects the optical response of the medium under a known light source, allowing for accurate calculation of the medium's color and diffuse reflectance.

[0049] The projected light reflection signal can be linearly separated from the mixed illumination by differential operation between the registered first and second images. The pixel values ​​on the differential image directly reflect the intensity and spectral composition of the reflected light from the medium. After obtaining the projected light reflection image, the three-channel values ​​of the medium color and the diffuse reflectance can be calculated using the Lambertian diffuse reflectance model, based on the known incident light intensity and incident angle information. The incident light intensity can be calculated based on the calibrated power of the projection light source and the light attenuation model, while the incident angle is obtained by fusing the local surface normal direction acquired by lidar or millimeter-wave radar with the vehicle attitude. This design reduces the estimation error of the medium's background color and reflectance under complex ambient light conditions. It provides a reliable benchmark for subsequent color compensation and brightness compensation, avoiding estimation bias caused by ambient light.

[0050] As one possible implementation, the control unit alternately turns the projection unit on and off, simultaneously triggering the camera to capture the first and second images. The time interval between the two frames is controlled within the reciprocal of the sensor frame rate to ensure the scene remains essentially still. After denoising and gain correction, the average value of the RGB three channels in the projection area is extracted as the medium color. The brightness value is extracted and combined with the incident angle and incident light intensity to calculate the diffuse reflectance.

[0051] By employing a time-difference method combined with physical model calculations, the color and diffuse reflectance of the medium can be accurately obtained in complex environments. This provides precise optical characteristic input for subsequent zoned adaptive adjustment, thereby synergistically enhancing the global projection quality adaptive capability. Furthermore, the method for obtaining the incident angle is not limited to lidar; three-dimensional surface information can also be acquired through stereo vision or structured light.

[0052] While estimating the diffuse reflectance of the projected light reflection image, color feature recognition can also be performed. The average value of the RGB three channels of the projected light reflection image within the target projection area is calculated to identify the basic hue of the medium, such as dark gray, blue, or red. For the projection of a white reference spot, the color of the reflected light is the medium's base color, which can be directly used as the reference for subsequent color compensation. Color features and diffuse reflectance together constitute the medium's optical properties, providing direct input for the color compensation parameters of subsequent sub-regions.

[0053] Building upon the estimation of diffuse reflectance and separation of ambient light, the specular reflection characteristics of the medium can be further estimated. For example, a simplified classification model using a BRDF (Bidirectional Reflectance Distribution Function) can be employed to determine the intensity of specular reflection based on the size, shape, and viewing angle variation of the highlight region. For the decoupling of the diffuse, specular, and ambient light components, based on the projection light reflection image obtained using the difference method between the first and second images, a spherical harmonic model can be combined to achieve higher-precision component separation.

[0054] Furthermore, echo sensors (such as LiDAR or multispectral cameras) can provide more accurate BRDF measurement data, improving the accuracy of specular reflection estimation. However, in the absence of echo sensors, basic specular reflection characteristic estimation can still be achieved by relying solely on interpolation matching between camera images and a prior material database, ensuring the applicability of the solution under different hardware configurations. Specular reflection characteristics will be used as part of the medium property data for subsequent specular area detection verification and fine-tuning of projection adjustment parameters.

[0055] In a perceived image, there may be specularly reflected highlight areas. To address the interference of these highlight areas on the estimation of medium properties, the medium property data also includes a highlight mask. The highlight mask, as part of the medium property data, is used to shield the highlight areas from interfering with subsequent parameter calculations. Brightness and saturation thresholds are detected in the perceived image. Areas with brightness exceeding a first preset threshold and saturation below a second preset threshold are identified as highlight areas, and corresponding highlight masks are generated. Based on the generated highlight mask, pixels within the highlight mask can be masked when generating projection adjustment parameters. This eliminates the contamination of medium property estimation by highlights, ensuring the reliability of the compensation parameters and preventing them from participating in the calculation of average reflectance and color statistics within that sub-region, thus avoiding bias introduced by highlights.

[0056] One possible implementation involves using dual-threshold joint detection to classify bright pixels with extremely low saturation as specular pixels. The brightness threshold can be set to a grayscale value of 220 (normalized to approximately 0.86), and the saturation threshold can be set to 0.1. After detecting the specular regions, a closed specular mask is generated using morphological dilation and erosion operations. When calculating the average color or reflectance of each sub-region, pixels within the specular mask are excluded, ensuring that the statistical results primarily reflect the characteristics of the diffuse reflective surface. This design avoids the excessively high brightness values ​​of specular regions from inflating the estimated reflectance of the entire sub-region, thereby preventing the generation of erroneous overcompensation parameters and ensuring that the projection adjustment parameters more accurately reflect the characteristics of the diffuse reflective medium, avoiding estimation bias and overexposure risks caused by specular highlights.

[0057] For example, after obtaining the perceived image, it is converted to the HSV (Hue, Saturation, Value) color space. The V and S components of each pixel are evaluated one by one, and pixels that satisfy V greater than 220 and S less than 0.1 are marked as specular candidate pixels. After connected component analysis and morphological operations, a specular mask is generated. Although the pixels within the specular mask are not included in the statistics when generating projection adjustment parameters (such as color and brightness parameters), special suppression coefficients can be applied to the specular mask area when adjusting the projected content. For example, the blue component projected into the specular area can be locally suppressed to further reduce the visual impact of the specular highlights.

[0058] By employing a high-light detection and shielding mechanism, visual interference and parameter estimation errors caused by specular reflection can be effectively eliminated, thereby synergistically enhancing the projection quality control capability under multi-media and non-ideal Lambertian surface conditions.

[0059] Understandably, the brightness threshold and saturation threshold can be calibrated according to the camera characteristics of different vehicle models.

[0060] In S120, based on the medium characteristic data, the current projection target area is divided into at least two sub-regions.

[0061] The basis for segmenting the current projection target area is the difference in medium properties within the projection target area.

[0062] In some embodiments, to improve the segmentation accuracy of sub-regions, the projection target region can be segmented into sub-regions based on various differences in the medium property data. For example, based on the medium property data, color differences and texture differences within the current projection target region can be determined; based on the color differences and texture differences, the current projection target region can be segmented into multiple non-overlapping homogeneous sub-regions.

[0063] Color differences can be quantified as color difference values ​​(e.g., based on the DeltaE color difference formula), while texture differences can be quantified as differences in grayscale co-occurrence matrix energy values. During segmentation, both types of difference information are considered, and material boundaries are only confirmed when both color and texture differences simultaneously meet preset conditions. This design ensures that shadow areas that might be misjudged based solely on color or smooth color abrupt change areas that might be missed based solely on texture can be correctly segmented, improving the robustness of segmentation to complex lighting and material overlays.

[0064] Furthermore, in the process of region segmentation based on color and texture differences, texture features can be extracted using the GLCM (Gray-Level Co-occurrence Matrix Energy) method. Texture feature parameters such as GLCM energy, contrast, and correlation are calculated for pixels corresponding to the projected target region in the projected light reflection image. These parameters are then combined with a lightweight semantic segmentation model to determine the specific type of medium, such as asphalt, epoxy flooring, floor tiles, and carpet. Based on the GLCM energy value, the texture intensity is divided into three levels: coarse (GLCM energy < 0.2), medium (0.2-0.6), and smooth (energy > 0.6). This texture level can serve as the basis for subsequently generating edge enhancement intensity and contrast compensation parameters for each sub-region.

[0065] As one possible implementation, the perceived image acquired by the camera is separated from the ambient light to obtain a projected light reflection image. Color and texture features are extracted from this image to generate a color difference map and a texture energy map. Edge detection is then performed on both feature maps, and the merged edge information is input into a region growing algorithm to obtain a segmentation mask. The seed points for the region growing algorithm can be selected through uniform sampling or based on the initial semantic segmentation results. The color difference threshold can be set to 30 deltaE, and the texture energy difference threshold can be set empirically to 0.3. These thresholds can be adjusted according to the actual application scenario.

[0066] The segmentation mechanism based on both color and texture criteria can accurately distinguish the boundaries of different materials, avoiding missegmentation that may be caused by relying on a single feature. This ensures the accuracy and effectiveness of subsequent independent compensation for each sub-region, and helps to solve the problem of inconsistent projection image quality in multi-material mixed scenes.

[0067] In other embodiments, when the medium property data is primarily composed of geometric properties, or when the optical property data is insufficient to distinguish different media (e.g., on a non-planar surface of a single color), the current projection target area can be segmented into at least two sub-regions based solely on geometric properties. Different geometric surfaces (e.g., slopes with different gradients, curved surfaces with different curvatures, or steps with abrupt height changes) cause different patterns and degrees of geometric distortion in the projected image, thus requiring segmentation into independent sub-regions to calculate geometric deformation compensation parameters for each, thereby achieving accurate correction within each sub-region.

[0068] As one possible approach, sub-region segmentation based on geometric characteristics first involves acquiring the geometric feature map of the projected target region. This can be achieved by using vehicle-mounted LiDAR or millimeter-wave radar to acquire dense 3D point cloud data, and then calculating various geometric features for each point or local region based on this 3D point cloud data. Geometric features may include: local normal directions obtained by fitting a local point cloud plane using principal component analysis, reflecting the surface tilt; surface curvature obtained by calculating the curvature of the local surface in the principal direction, used to distinguish planar, convex, or concave regions; and the rate of depth change perpendicular to the vehicle's direction of travel and the height difference within a statistical window, used to identify locations of abrupt depth changes such as ramps, potholes, or steps.

[0069] After obtaining the geometric feature map, various segmentation algorithms can be used to divide the projected target region into multiple sub-regions based on the differences in these features. For example, a region growing algorithm based on normal vectors can be used, starting from a seed point, continuously incorporating neighboring points with normal vector angles less than a preset threshold into the same sub-region, thereby achieving separation, for example, between a flat road surface and an uphill slope. Alternatively, a 3D point cloud can be projected into a 2D depth map, and edge detection operators can be used to identify locations of abrupt depth changes, thus segmenting different blocks that are continuous in depth. Furthermore, by calculating the local surface curvature and using clustering algorithms to group regions with similar curvature values ​​into one category, it is possible to distinguish between projected regions with distinctly different geometric shapes, such as flat areas and curved walls. Through these methods, projected target regions with different geometric shapes can be accurately segmented into multiple homogeneous sub-regions.

[0070] In other embodiments, when the acquired medium property data includes both optical and geometric properties, these two types of data can be used synergistically for sub-region segmentation to improve segmentation accuracy. For example, the control unit first extracts optical feature maps (such as a medium property distribution map composed of color and texture differences) and geometric feature maps (such as a surface morphology distribution map composed of local normal directions, depth gradients, or surface curvature). Then, a decision-level fusion or feature-level fusion strategy can be used to comprehensively process the two types of features. In decision-level fusion, segmentation can be performed independently based on the optical and geometric feature maps to obtain two preliminary sub-region masks, and then the final partitioning result is determined through logical operations or weighted voting based on confidence. In feature-level fusion, optical and geometric features can be concatenated into a unified multi-dimensional feature vector, which can be directly input into a clustering algorithm to achieve region division. Through this fusion strategy, stable and accurate sub-region segmentation can be achieved in scenarios where the surface color and texture are consistent but the geometric shape varies, or where the surface is flat but the material type is different.

[0071] In S130, projection adjustment parameters are generated independently for each sub-region based on the corresponding medium characteristic data.

[0072] The projection adjustment parameters include at least one of the following: color compensation parameters, brightness compensation parameters, contrast compensation parameters, and geometric deformation compensation parameters.

[0073] Using sub-regions as basic units, the system extracts medium characteristic data within each sub-region, such as medium color, diffuse reflectance, texture energy, and three-dimensional geometry, and independently calculates adjustment parameters applicable to that sub-region. As one possible implementation, color compensation parameters can be generated using a strategy combining table lookup and real-time fine-tuning. First, the system pre-sets a standard RGB compensation coefficient lookup table for common media (such as asphalt, epoxy flooring, carpet, and red brick). If the medium type of the current sub-region matches the lookup table, the corresponding baseline compensation coefficient is directly read. If not, based on the principle of minimizing color difference, a color compensation matrix is ​​calculated in real-time to ensure that the perceived color after projection reflection matches the target color. Taking a projected white target pattern as an example, the compensation coefficients for each common medium follow the following rules: the coefficient of the channel corresponding to the medium's own color decreases, the coefficient of the complementary color channel increases, the overall coefficient for dark media (such as asphalt) is greater than 1, and the overall coefficient for light media (such as light-colored carpet) is less than 1. For example, for asphalt pavement (dark gray background), the R coefficient is 1.2, the G coefficient is 1.2, and the B coefficient is 1.3; for blue epoxy flooring (blue background), the R coefficient is 1.3, the G coefficient is 1.1, and the B coefficient is 0.7; for red brick walls (red background), the R coefficient is 0.8, the G coefficient is 1.2, and the B coefficient is 1.2; and for light-colored carpet (off-white background), the R coefficient is 0.9, the G coefficient is 0.9, and the B coefficient is 0.95. By combining table lookup with real-time fine-tuning, suitable color compensation parameters can be quickly and accurately assigned to each sub-region, thereby correcting the projected color cast caused by the background color of the medium.

[0074] The brightness compensation parameter can be dynamically adjusted based on the ratio of the diffuse reflectance of the sub-region to the diffuse reflectance of the standard gray medium, and by incorporating an ambient light correction factor. For example, based on the formula: Lout = Ltarget × (Ractual / Rref) × fenv(Eamb); where Ltarget is the desired target brightness under standard medium conditions; Rref is the diffuse reflectance of the standard gray medium; Ractual is the actual diffuse reflectance of the medium; fenv(Eamb) = max(0.4, Eref / Eamb) is the ambient light correction factor, and Eref is the reference ambient illuminance (e.g., 100 lx), with a lower limit of 0.4 to prevent overcompensation. Using this formula, the system can dynamically adjust the brightness gain according to the actual diffuse reflectance of each sub-region and the ambient lighting conditions, ensuring that the final projected brightness remains consistent across media of varying brightness.

[0075] Contrast compensation parameters can be correlated with the texture roughness of the medium. For rough surfaces (such as asphalt and rubble), increasing the contrast (e.g., using a contrast enhancement factor of 1.2 to 1.5) compensates for detail loss caused by diffuse scattering. For smooth surfaces (such as ceramic tiles and glass), appropriately reducing the contrast (e.g., using a contrast reduction factor of 0.7 to 0.9) suppresses glare. Geometric deformation compensation parameters can be generated by constructing a homography transformation matrix from the projector coordinate system to the coordinate system of the current projection target area. This contrast adjustment strategy can work in conjunction with subsequent texture energy-based edge enhancement parameters to ensure both clarity and comfort of the projected image across different materials.

[0076] In addition, projection adjustment parameters can include other types of adjustment values, such as local suppression coefficients for highlight areas and edge enhancement intensity. These parameters can all be mapped to the medium property data of each sub-region.

[0077] In some embodiments, to address the problem of blurred edges of projected patterns on rough media surfaces, the projection adjustment parameters may further include edge enhancement parameters; the texture energy of the current projection target area is calculated based on media characteristic data; the edge enhancement intensity is determined according to a preset intensity level corresponding to the texture energy; the lower the texture energy, the higher the edge enhancement intensity.

[0078] Texture energy refers to the energy value calculated based on the gray-level co-occurrence matrix, used to quantify the roughness of the surface texture. Texture energy is inversely proportional to surface roughness: the lower the energy value, the rougher the surface texture, the stronger the diffuse reflection, and the more severe the blurring of the projected pattern edges. Rough surfaces (high texture intensity) diffusely scatter projected light more strongly, leading to more severe loss of edge information; therefore, stronger edge enhancement is needed to recover details. Conversely, smooth surfaces (low texture intensity) diffusely scatter weakly, and over-sharpening can actually generate noise.

[0079] By establishing a negative correlation between texture energy and edge enhancement intensity, regions with GLCM energy below 0.2 are classified as rough textures and enhanced strongly with an edge enhancement factor of 2.5-3.0; regions with energy between 0.2 and 0.6 are classified as medium textures and enhanced with an enhancement factor of 1.5-2.0; and regions with energy above 0.6 are classified as smooth surfaces and enhanced slightly or not at all. This design allows for an adaptive match between edge sharpening and the texture characteristics of the medium. It provides stronger edge recovery for rough surfaces and avoids over-sharpening and generating noise on smooth surfaces, thereby restoring the blurred edges and loss of detail in the projected pattern caused by the roughness of the medium, and maintaining the clarity of the projected image on rough surfaces.

[0080] For example, after obtaining the mask for each sub-region, the texture features of the projected light reflection image within that sub-region are extracted, its GLCM energy value is calculated, and the edge enhancement coefficient corresponding to that sub-region is obtained by looking up a table according to the hierarchical strategy. Then, when performing sub-region image adjustment on the original image, an edge enhancement filtering operation of corresponding intensity is performed on the image within that sub-region.

[0081] By employing a texture-adaptive edge enhancement scheme, the projection clarity on rough surfaces can be improved while maintaining the image purity in smooth areas, thereby further enhancing the adaptability of the projection adjustment method to different media surfaces.

[0082] It is understandable that the texture intensity grading threshold is not limited to the example value and can be optimized based on actual road material test data, such as lowering the roughness threshold to 0.15.

[0083] In some embodiments, to address the geometric distortion problem caused by changes in the geometric shape of the projected surface during vehicle dynamic driving or parking, geometric deformation compensation parameters can be generated in the following manner: acquiring three-dimensional point cloud data of the current projection target area; constructing a homography transformation matrix from the projector coordinate system to the current projection target area coordinate system by combining the vehicle suspension pitch angle and wheel rotation angle; and generating geometric deformation compensation parameters based on the homography transformation matrix.

[0084] 3D point cloud data is obtained by non-contact ranging sensors such as lidar or millimeter-wave radar, directly reflecting the actual spatial morphology of the projected target area. Vehicle suspension pitch angle and wheel rotation angle reflect the vehicle's attitude changes relative to the road surface; combining these two parameters allows for precise determination of the relative pose between the projector and the projection surface. By solving the homography transformation matrix, each pixel in the projected image can be pre-distorted and mapped onto the surface coordinate system, ensuring the projected image maintains the correct proportions and shape on rough, inclined, or curved surfaces, eliminating geometric distortions such as trapezoidal distortion. This avoids interference from calibration patterns for drivers and other road users and enables rapid response to changes in the geometry of the projection surface, such as switching from a flat road surface to a sloping wall or ramp.

[0085] As one possible implementation, the control unit receives point cloud frames from the lidar and suspension height and steering angle signals from the vehicle's CAN bus. First, the point cloud is filtered, and the local planar or surface equations of the projection target region are fitted. Then, based on the projector's intrinsic and extrinsic parameters and vehicle attitude data, a 3×3 homography transformation matrix is ​​calculated using direct linear transformation or least squares method. This matrix is ​​used as a geometric deformation compensation parameter to perform perspective transformation and remapping on the projected image, achieving pre-distortion.

[0086] By using a geometric correction scheme that fuses radar point clouds with vehicle attitude, the geometric accuracy of the projected pattern can be improved under any surface shape without projecting any calibration pattern. This helps to solve the image distortion problem in vehicle dynamic projection and improves driving safety and visual effects.

[0087] In S140, based on the projection adjustment parameters of each sub-region, the image content of the corresponding sub-region in the original image is adjusted to generate an adjusted projection image.

[0088] First, the image content of the corresponding sub-region is extracted from the original image. Then, the projection adjustment parameters corresponding to that sub-region are applied for compensation processing. Finally, the processed sub-region image content is recombined according to spatial position to generate the final projected image. For example, for the dark asphalt pavement sub-region, the brightness and contrast of pixels can be improved; for the blue epoxy flooring sub-region, color deviation can be corrected; and for areas with geometric distortion, pre-distortion remapping can be implemented.

[0089] In some embodiments, when the current projection target region is divided into at least two sub-regions, visual discontinuities may occur at the boundaries of different sub-regions. To eliminate this phenomenon, the boundary transition can be further processed when adjusting the image content of each sub-region. For example, within the boundary transition zone between adjacent sub-regions, the projection adjustment parameters of each adjacent sub-region are mixed using an interpolation function with distance as the weight to obtain smooth projection adjustment parameters within the boundary transition zone.

[0090] As one possible implementation, the width of the boundary transition band can be set to 20-30 pixels. For each pixel within the transition band, a weight is calculated based on its distance to the boundaries of the two sub-regions, for example using linear interpolation or high-order B-spline interpolation, so that the compensation coefficient changes from K of sub-region A. A Smooth transition to sub-region B K B The boundary K is determined by the formula K = w1 × K. A + w2×K B In the formula, K_boundary represents the final compensation coefficient (which can be the coefficients of each of the RGB three channels, or the brightness or contrast coefficient) of a pixel located in the transition zone between region A and region B. A K represents the compensation coefficient used within region A (calculated independently based on the medium characteristics of that region). BThe compensation coefficients used within region B are represented by w1 and w2, which are weighting coefficients that satisfy w1 + w2 = 1. Their values ​​continuously change with the pixel's position in the transition zone. For example, when a pixel is closer to region A, w1 is close to 1 and w2 is close to 0; the opposite is true when it's closer to region B. At the intermediate position, w1 = w2 = 0.5. This design ensures the continuity of the compensation parameters at geometric boundaries. At the junction of two materials, such as between floor tiles and seams, the brightness and color of the projected image will not abruptly change or appear spliced, achieving a natural visual fusion. This enhances the overall feel and aesthetics of multi-material projected images and reduces the perceived side effects of zone adjustment.

[0091] Understandably, the width of the transition band is not limited to 20-30 pixels and can be adjusted according to the projection resolution and viewing distance.

[0092] To coordinate the effects of various compensation processes, independent projection adjustment parameters can be applied at the sub-region level. For example, for each sub-region, the following processes are performed: based on the color compensation parameters corresponding to the sub-region, color compensation is applied to the pixels of the corresponding sub-region in the original image; based on the brightness compensation parameters or contrast compensation parameters corresponding to the sub-region, brightness or contrast is adjusted for the pixels of the corresponding sub-region in the original image; based on the geometric deformation compensation parameters corresponding to the sub-region, geometric remapping is performed on the sub-region image after color compensation, brightness adjustment, or contrast adjustment; the content of each sub-region image after geometric remapping is stitched together according to a preset spatial position to generate an adjusted projection image.

[0093] As an example processing sequence, color compensation, brightness adjustment, or contrast adjustment can be placed before geometric remapping to reduce the impact of pixel spatial location resampling on the accuracy of color and brightness calculations. Finally, the image content of each sub-region after geometric remapping is stitched together according to a preset spatial position to generate an adjusted projected image. It is understood that the execution order of color compensation, brightness adjustment, or contrast adjustment with geometric remapping can be adjusted as needed.

[0094] To resolve the conflict between the real-time performance of projection adjustment and the computational load of the system when vehicle speed changes, the projection adjustment method can also integrate a dynamic update mechanism that adapts to vehicle speed. For example, the vehicle speed is acquired; when the vehicle speed is below a first speed threshold, the acquisition of the perceived image and the updating of the projection adjustment parameters are performed at a first frequency; when the vehicle speed is between the first and second speed thresholds, the acquisition of the perceived image and the updating of the projection adjustment parameters are performed at a second frequency lower than the first frequency; when the vehicle speed is above the second speed threshold, the acquisition of the perceived image and the updating of the projection adjustment parameters are stopped, and the system switches to a static projection mode; wherein, the static projection mode projects pre-stored fixed image content.

[0095] As one possible implementation, the first speed threshold can be set to 5 km / h, and the second speed threshold to 30 km / h. At low speeds or when parked, the update frequency is, for example, 30-60 fps to ensure precise image tracking. At low to medium speeds, the update frequency is reduced to once every 200-500 ms for media analysis and compensation parameter calculation, but the projected image can retain the compensation results from the previous cycle to avoid screen flicker. At high speeds, due to reduced usability of the projection function and rapid road surface changes, the system automatically stops dynamic adjustment and switches to static projection mode, such as projecting simple warning symbols or disabling dynamic effects, thereby reducing the computational power consumption of the control unit and ensuring system stability. This design automatically balances projection quality and computational resources based on vehicle speed, taking into account both user experience and safety requirements under different driving conditions.

[0096] Understandably, the speed threshold and update frequency can be calibrated based on vehicle type, sensor response time, and projection system bandwidth. For example, for high-performance computing platforms, the second speed threshold can be increased to 40 km / h, and the second frequency can be maintained above 10 fps. Furthermore, the static projection mode may not be a completely fixed image, but rather updates low-dynamic content such as turn arrows based on navigation information, while maintaining the media characteristic compensation parameters at the locked state from the last low-speed setting. The first and second frequencies are not limited to fixed values ​​and can also decrease continuously according to vehicle speed.

[0097] To further enhance the practicality and personalized user experience of the projection adjustment system, scene mode management and user-customized adaptation mechanisms can be set up. For example, the system pre-stores default compensation parameter combinations for common scenarios such as epoxy flooring in underground parking garages, wet and slippery road surfaces in rainy weather, and standard asphalt road surfaces. When the medium characteristic data successfully matches any preset scene, the system automatically retrieves the corresponding scene mode. If the matching degree is lower than a preset threshold, it enters real-time calculation mode and generates projection adjustment parameters in real time according to the projection adjustment method provided in this application.

[0098] Within the same scene mode, since the brightness compensation factor already includes ambient light feedback (such as the FNV (Eamb) ambient light correction factor), the system can adapt to different ambient illuminance levels without additional mode switching. Furthermore, users can manually fine-tune parameters such as brightness, contrast, or hue via the central control screen, and the resulting offset vector is superimposed on the standard parameters of the current scene and stored in the user configuration file. Users can also name the currently adjusted parameter combination a new scene and save it. The system will proactively prompt the user whether to apply this custom mode the next time it encounters a projection target area with a media feature similarity greater than 85%, enabling continuous expansion and adaptive learning of the scene library.

[0099] This application also provides a vehicle including headlights and a projection adjustment system. The headlights are mounted at the front of the vehicle and integrate a projection optical engine (such as a DLP optical engine module) with image projection capabilities. The headlights not only provide conventional forward road lighting but also, driven by a control unit, project pre-stored or real-time generated graphic content onto external target surfaces (such as the ground or walls) to achieve information interaction. The projection adjustment system is communicatively connected to the headlights and is used in the projection adjustment method provided in this application. This allows the vehicle to project high-quality images onto the ground or walls through the headlights while driving or parked, with less influence from variations in the color, material, and geometry of the medium, thus improving the practicality and user experience of the headlight projection.

[0100] Figure 2 This is a schematic diagram of the projection adjustment device provided in an exemplary embodiment of this application, as shown below. Figure 2 As shown, the projection adjustment device 2 includes: a sensing unit 21, used to acquire a sensing image of the current projection target area and separate the ambient light to obtain medium characteristic data reflecting the current projection target area; the medium characteristic data includes at least one of optical characteristics and geometric characteristics; a region segmentation unit 22, used to segment the current projection target area into at least two sub-regions based on the medium characteristic data; a parameter generation unit 23, used to independently generate projection adjustment parameters for each sub-region according to the corresponding medium characteristic data; wherein, the projection adjustment parameters include at least one of color compensation parameters, brightness compensation parameters, contrast compensation parameters, and geometric deformation compensation parameters; and a projection generation unit 24, used to adjust the image content of the corresponding sub-region in the original image based on the projection adjustment parameters of each sub-region to generate an adjusted projection image.

[0101] As one possible implementation, the region segmentation unit 22 can be configured to: determine the color differences and texture differences within the current projection target region based on medium characteristic data; and segment the current projection target region into multiple non-overlapping homoprotic regions based on the color differences and texture differences.

[0102] As one possible implementation, the sensing unit 21 can be configured to: acquire a first image without a projected compensation pattern, the first image containing only ambient light; acquire a second image with a projected compensation pattern, the second image containing both projected light and ambient light; obtain a projected light reflection image based on the difference between the second image and the first image; and calculate the optical properties of the medium based on the projected light reflection image, a preset incident light intensity, and an incident angle obtained by the sensor; wherein the optical properties include at least the medium color and diffuse reflectance.

[0103] As one possible implementation, the medium characteristic data also includes a specular mask; the projection adjustment device 2 may further include: performing brightness threshold and saturation threshold detection on the perceived image, determining the area with brightness exceeding a first preset threshold and saturation lower than a second preset threshold as a specular area, and generating a specular mask for the corresponding specular area; wherein, the parameter generation unit 23 may be configured to: shield the pixels within the specular mask when generating projection adjustment parameters.

[0104] As one possible implementation, the projection adjustment parameters also include edge enhancement parameters; the parameter generation unit 23 can be configured to: calculate the texture energy of the current projection target area based on the medium characteristic data; determine the edge enhancement intensity according to the preset intensity level corresponding to the texture energy; the lower the texture energy, the higher the edge enhancement intensity.

[0105] As one possible implementation, the geometric deformation compensation parameters are generated as follows: acquiring the three-dimensional point cloud data of the current projection target area; constructing a homography transformation matrix from the projector coordinate system to the current projection target area coordinate system by combining the vehicle suspension pitch angle and wheel rotation angle; and generating the geometric deformation compensation parameters based on the homography transformation matrix.

[0106] As one possible implementation, the projection generation unit 24 can be configured to: within the boundary transition zone between adjacent sub-regions, use an interpolation function with distance as the weight to mix the projection adjustment parameters of each adjacent sub-region to obtain smooth projection adjustment parameters within the boundary transition zone.

[0107] As one possible implementation, the projection generation unit 24 can be configured to: perform color compensation on the pixels of the corresponding sub-region in the original image based on the color compensation parameters corresponding to the sub-region; adjust the brightness or contrast of the pixels of the corresponding sub-region in the original image based on the brightness compensation parameters or contrast compensation parameters corresponding to the sub-region; perform geometric remapping on the sub-region image after color compensation, brightness adjustment or contrast adjustment based on the geometric deformation compensation parameters corresponding to the sub-region; and stitch the contents of each sub-region image after geometric remapping according to a preset spatial position to generate an adjusted projection image.

[0108] As one possible implementation, the projection adjustment device 2 may include: acquiring the vehicle speed; when the vehicle speed is lower than a first speed threshold, performing the acquisition of a perceived image and updating the projection adjustment parameters at a first frequency; when the vehicle speed is between the first speed threshold and a second speed threshold, performing the acquisition of a perceived image and updating the projection adjustment parameters at a second frequency lower than the first frequency; when the vehicle speed is higher than the second speed threshold, stopping the acquisition of a perceived image and updating the projection adjustment parameters, and switching to a static projection mode; wherein, the static projection mode projects pre-stored fixed image content.

[0109] An electronic device includes: a processor; a memory for storing processor-executable instructions; and a processor for executing the projection adjustment method described in the embodiments provided in this application.

[0110] Below, for reference Figure 3 This application describes an electronic device according to embodiments thereof. The electronic device may be either or both of a first device and a second device, or a standalone device independent of them, which may communicate with the first device and the second device to receive acquired input signals from them.

[0111] Figure 3 A block diagram of an electronic device according to an embodiment of this application is illustrated.

[0112] like Figure 3 As shown, the electronic device 30 includes one or more processors 31 and memory 32.

[0113] The processor 31 may be a central processing unit (CPU) or other form of processing unit with data processing and / or instruction execution capabilities, and may control other components in the electronic device 30 to perform desired functions.

[0114] The memory 32 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 31 may execute the program instructions to implement the projection adjustment methods of the various embodiments of this application described above and / or other desired functions. Various contents such as input signals, signal components, and noise components may also be stored in the computer-readable storage medium.

[0115] In one example, the electronic device 30 may also include an input device 33 and an output device 34, which are interconnected via a bus system and / or other forms of connection mechanism (not shown).

[0116] When the electronic device is a standalone device, the input device 33 can be a communication network connector for receiving the collected input signals from the first device and the second device.

[0117] In addition, the input device 33 may also include, for example, a keyboard, a mouse, etc.

[0118] The output device 34 can output various information to the outside, including determined distance information, direction information, etc. The output device 34 may include, for example, a display, a speaker, a printer, and a communication network and its connected remote output devices, etc.

[0119] Of course, for the sake of simplicity, Figure 3 Only some of the components of the electronic device 30 relevant to this application are shown in this illustration; components such as buses, input / output interfaces, etc., are omitted. In addition, the electronic device 30 may include any other suitable components depending on the specific application.

[0120] The computer program product can be written in any combination of one or more programming languages ​​to perform the operations of the embodiments of this application. The programming languages ​​include object-oriented programming languages ​​such as Java and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0121] A computer-readable storage medium stores a computer program for performing the projection adjustment method described in the embodiments provided in this application.

[0122] The computer-readable storage medium may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0123] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations thereof.

Claims

1. A projection adjustment method, characterized in that, Image projection applied to vehicle headlights, the method comprising: Acquire a perceived image of the current projection target area and separate the ambient light to obtain medium characteristic data reflecting the current projection target area; the medium characteristic data includes at least one of optical characteristics and geometric characteristics. Based on the medium characteristic data, the current projection target area is divided into at least two sub-regions; For each sub-region, projection adjustment parameters are independently generated based on the corresponding medium characteristic data; wherein, the projection adjustment parameters include at least one of color compensation parameters, brightness compensation parameters, contrast compensation parameters, and geometric deformation compensation parameters; Based on the projection adjustment parameters of each sub-region, the image content of the corresponding sub-region in the original image is adjusted to generate an adjusted projection image.

2. The projection adjustment method according to claim 1, characterized in that, The step of dividing the current projection target region into at least two sub-regions based on the medium characteristic data includes: Based on the medium characteristic data, determine the color and texture differences within the current projection target area; Based on the color and texture differences, the current projection target region is divided into multiple non-overlapping homoprotic regions.

3. The projection adjustment method according to claim 1, characterized in that, The process of acquiring a perceived image of the current projection target area and separating ambient light to obtain medium characteristic data reflecting the current projection target area includes: A first image is captured without a projected compensation pattern, and the first image contains only ambient light; A second image is acquired when the compensation pattern is projected, the second image containing the projected light and ambient light; Based on the difference between the second image and the first image, a projected light reflection image is obtained; Based on the projected light reflection image, the preset incident light intensity, and the incident angle obtained by the sensor, the optical properties of the medium are calculated; wherein the optical properties include at least the medium color and diffuse reflectance.

4. The projection adjustment method according to claim 1, characterized in that, The dielectric property data also includes a high-brightness mask; the method further includes: The perceived image is subjected to brightness threshold and saturation threshold detection. Regions with brightness exceeding a first preset threshold and saturation lower than a second preset threshold are identified as highlight regions, and a highlight mask corresponding to the highlight region is generated. Specifically, generating projection adjustment parameters independently for each sub-region based on corresponding medium characteristic data includes: When generating the projection adjustment parameters, the pixels within the specular mask are masked.

5. The projection adjustment method according to claim 1, characterized in that, The projection adjustment parameters also include edge enhancement parameters; The step of independently generating projection adjustment parameters for each sub-region based on the corresponding medium characteristic data includes: Calculate the texture energy of the current projection target region based on the medium characteristic data; The edge enhancement intensity is determined based on the preset intensity level corresponding to the texture energy; the lower the texture energy, the higher the edge enhancement intensity.

6. The projection adjustment method according to claim 1, characterized in that, The geometric deformation compensation parameters are generated in the following manner: Obtain the 3D point cloud data of the current projection target area; By combining the vehicle suspension pitch angle and wheel rotation angle, a homography transformation matrix is ​​constructed from the projector coordinate system to the coordinate system of the current projection target area; The geometric deformation compensation parameters are generated based on the homography transformation matrix.

7. The projection adjustment method according to claim 1, characterized in that, The adjustment of the image content of the corresponding sub-region in the original image based on the projection adjustment parameters of each sub-region includes: Within the boundary transition zone between adjacent sub-regions, the projection adjustment parameters of each adjacent sub-region are mixed using a distance-weighted interpolation function to obtain smooth projection adjustment parameters within the boundary transition zone.

8. The projection adjustment method according to claim 1, characterized in that, The adjustment of the image content of the corresponding sub-region in the original image based on the projection adjustment parameters of each sub-region includes: Based on the color compensation parameters corresponding to the sub-region, color compensation is performed on the pixels of the corresponding sub-region in the original image; Based on the brightness compensation parameters or contrast compensation parameters corresponding to the sub-regions, the brightness or contrast of the pixels in the corresponding sub-regions of the original image is adjusted. Based on the geometric deformation compensation parameters corresponding to the sub-region, geometric remapping is performed on the sub-region image after color compensation, brightness adjustment or contrast adjustment. The image content of each sub-region after geometric remapping is stitched together according to a preset spatial position to generate an adjusted projection image.

9. The projection adjustment method according to claim 1, characterized in that, The method further includes: Get vehicle speed; When the vehicle speed is lower than a first speed threshold, the acquisition of the perceived image and the updating of the projection adjustment parameters are performed at a first frequency; When the vehicle speed is between the first speed threshold and the second speed threshold, the acquisition of the perceived image and the updating of the projection adjustment parameters are performed at a second frequency lower than the first frequency. When the vehicle speed exceeds the second speed threshold, the acquisition of the perceived image and the updating of the projection adjustment parameters are stopped, and the system switches to static projection mode; wherein, the static projection mode projects pre-stored fixed image content.

10. A vehicle, characterized in that, include: Including vehicle headlights; A projection adjustment system, which is communicatively connected to the vehicle headlights, is used to perform the projection adjustment method according to any one of claims 1-9.

Images