HUD software defined dimming system and method based on GPU calculation

Abstract

The invention relates to the technical field of HUD light control, in particular to an HUD software-defined dimming system and method based on GPU computing, the system adopts a Compute Shader to construct a parallel computing framework, and real-time dynamic regulation and control of N dimming partitions are realized through a GPU parallel stream processor. Specifically, rendered image data is matched with a lamp panel, region division is carried out on LED lamp beads on the lamp panel, and a coordinate mapping table is output; pixel-level parallel calculation: traversing each pixel of the rendered image, if the pixel is a non-black pixel, triggering dimming, lightening a corresponding partition on a lamp panel, and performing edge optimization; and compressing the data according to bits and transmitting the compressed data to a light controller. Therefore, the backlight control instruction period is shortened to be within 3ms, the real-time requirement is met, and delta Elt is kept; a static contrast gt is achieved on the premise of color precision 2; 3000: 1, and the overall power consumption of the system is reduced.

Description

Technical Field

[0001] This application relates to the technical field of HUD lighting control, and particularly to a HUD software-defined dimming system and method based on GPU computing. Background Art

[0002] In the current mainstream automotive HUD display technologies are TFT (Thin Film Transistor) and DLP (Digital Light Processing), with a static contrast ratio ranging from 1000:1 to 1500:1. To improve the display effect, the prior art adopts two solutions:

[0003] Solution A: Increase the power of the display device to directly increase the brightness. However, this results in a power consumption increase of more than 30% (measured data), affecting the battery life of electric vehicles.

[0004] Solution B: Hardware-level Local Dimming, which realizes zoned dimming through an independent control chip (such as Texas Instruments DLP5530-Q1). However, the following problems exist: 1. The hardware cost increases by 15%-20% (BOM analysis data); 2. Lack of flexibility, the hardware solution cannot adapt to irregular partitions, and the shape of the dimming area is limited (only rectangular partitions are supported); 3. The dynamic response delay > 5ms (measured data), affecting real-time performance. Further, both of the above two solutions cannot balance the contrast and energy consumption issues, and the existing dimming algorithms generate dark areas at the partition junctions, resulting in the loss of graphic edge information. Summary of the Invention

[0005] One of the problems in the background art that the present invention aims to solve.

[0006] To this end, the present invention provides a HUD software-defined dimming system and method based on GPU computing.

[0007] The technical solution adopted by the present invention to solve its technical problems is:

[0008] A HUD software-defined dimming system based on GPU computing, which is used for the lighting control of HUD. The dimming system includes,

[0009] A GPU acceleration module, and

[0010] A lighting controller, which receives the dimming instruction from the GPU acceleration module and drives the LED light board to be lit in zones;

[0011] Specifically, the GPU acceleration module includes:

[0012] A frame buffer parsing unit, which extracts the current frame image data from the rendering pipeline;

[0013] A Computer Shader computing unit, which is connected to the frame buffer parsing unit and executes a real-time dimming algorithm to achieve pixel-level parallel computing;

[0014] A data compression module, which is connected to the Computer Shader computing unit, encodes the dimming instructions, and the data compression module is connected to the lighting controller.

[0015] Further, in the Computer Shader computing unit, the lamp board is divided into regions, and the partition types of the region division include one of rectangular partition, honeycomb partition, and annular partition.

[0016] Further, in the Computer Shader computing unit, edge optimization is implemented: if the rendering image pixel is close to the boundary of the partition, adjacent partitions are lit simultaneously.

[0017] Further, if the distance between the edge of the rendering image and the boundary of the partition is less than 30 pixels, edge optimization is implemented to light the adjacent partitions.

[0018] Further, the data compression module compresses the lightMap array by bit; and transmits it to the lighting controller through DMA.

[0019] An HUD software-defined dimming method based on GPU computing, including,

[0020] Step 1, dynamic partition mapping, matching the rendered image data with the lamp board, dividing the LED lamp beads on the lamp board into regions, and outputting a coordinate mapping table;

[0021] Step 2, pixel-level parallel computing, traversing each pixel of the rendered image, triggering dimming if it is a non-black pixel, lighting the corresponding partition on the lamp board, and performing edge optimization;

[0022] Step 3, compressing and transmitting the data by bit to the lighting controller.

[0023] Further, in the step 1, the lamp board is partitioned in a rectangular partition manner, and the width of each rectangular partition is The height is where W total is the overall width of the lamp board, H total is the overall height of the lamp board, C column is the overall number of columns of the lamp board partitions, and R row is the overall width of the lamp board partitions.

[0024] Further, in the step 1, the lamp board is partitioned in a honeycomb partition manner, and the side length of the honeycomb partition is Among them, S is the side length of the hexagon, and N column is the number of columns in the horizontal direction.

[0025] The beneficial effects of the present invention are as follows:

[0026] In terms of the system architecture, a parallel computing framework is constructed using a computer graphics compute shader to achieve hardware-level acceleration of the backlight control algorithm on a heterogeneous computing platform of a graphics processing unit (GPU). This architecture breaks through the computing bottleneck of traditional fixed-function hardware dimming solutions, shortening the backlight control instruction cycle to within 3 ms, meeting the real-time requirements (60 Hz refresh rate) of in-vehicle HUD systems.

[0027] An adaptive backlight control algorithm based on pixel-level brightness analysis is developed to achieve real-time dynamic regulation of N dimming zones through the parallel stream processors of the GPU. Tested by the CIE 1931 standard, under the premise of maintaining a color accuracy of ΔE < 2, this solution achieves a static contrast ratio > 3000:1 (measured value 3186:1), which is 12 times higher than the traditional solution.

[0028] Innovatively, a software-hardware decoupling design is adopted, replacing a dedicated ASIC chip with a programmable rendering pipeline. Measured data shows that when maintaining the same light output power, the overall power consumption of the system is reduced by 42.7% (typical value), and the BOM cost is reduced by more than 37% (based on the quotation analysis of Tier1 suppliers).

[0029] A parametric dimming model is constructed to support the definition of dimming regions with irregular polygons (such as honeycomb, hexagon, etc.), continuous gradients, and dynamic variable shapes. The Voronoi diagram segmentation algorithm is used to achieve smooth edge processing of dimming zones with arbitrary shapes.

[0030] An edge light compensation algorithm based on a convolutional neural network (CNN) is developed to predict the brightness of a 3x3 neighborhood in the shader unit. This mechanism reduces the incidence of dark areas at the edges of the graphics from the industry average of 6.3% to below 0.8% by dynamically expanding the effective illumination area.

[0031] This solution successfully resolves the classic contradiction of "high contrast requirements and limited power consumption budget" in the field of in-vehicle displays, and its technical indicators (contrast ratio ≥ 3000:1, power consumption ≤ 1.8 W / dm 2 , and the morphological adaptability coefficient > 0.92).

[0032] The topology-independent characteristic of this technical framework opens up a new paradigm for styling integration in automotive HUD systems. Vehicle manufacturers can freely define the morphological parameters of dimming zones based on the vehicle model design language (such as the hexagonal matrix for sports models and the star array for luxury models), achieving an organic unity of technical performance and aesthetic design. Description of the Drawings

[0033] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0034] Figure 1 It is a schematic structural diagram of the HUD software-defined dimming system based on GPU computing in the present invention.

[0035] Figure 2 It is a schematic flow diagram of the HUD software-defined dimming method based on GPU computing in the present invention.

[0036] Figure 3 It is a schematic diagram of the rendering matching of the light board under the rectangular partition in the present invention.

[0037] Figure 4 It is a schematic diagram of the light board in the honeycomb partition in the present invention.

[0038] Figure 5 It is a schematic diagram of the honeycomb color partition in the present invention.

[0039] Figure 6 It is a schematic diagram of the honeycomb color partition with a common area added in the present invention.

[0040] Figure 7 It is a schematic diagram of the rendering matching of the light board under the honeycomb partition in the present invention.

[0041] Figure 8 It is a schematic diagram of the backlight effect of 192 rectangular partitions in the present invention.

[0042] Figure 9 It is a schematic diagram of the backlight effect of 1080 rectangular partitions in the present invention. Detailed Embodiments

[0043] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, and therefore only showing the components related to the present invention.

[0044] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. are based on the orientation or positional relationships shown in the drawings. These are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present invention. In addition, features defined as "first", "second" may explicitly or implicitly include one or more of such features. In the description of the present invention, unless otherwise specified, the meaning of "a plurality" is two or more.

[0045] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, the terms "mounted", "connected", "coupled" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0046] An HUD software-defined dimming system based on GPU computing includes an HMI software layer and an embedded control layer.

[0047] The HMI software layer integrates a GPU acceleration module, which specifically includes a frame buffer parsing unit, a Computer Shader calculation unit, and a data compression module. The frame buffer parsing unit is used to extract the current frame image data from the rendering pipeline, and the image data is a binary black-and-white image; the Computer Shader calculation unit is used to execute a real-time dimming algorithm according to the image data and transmit the calculated dimming instruction to the data compression module; the data compression module encodes the dimming instruction into a bitstream (such as 48 partitions → 6-byte array).

[0048] The embedded control layer is set as a lighting controller, and the lighting controller receives the bitstream instruction of the data compression module through the I2C bus and drives the LED light board partitions (PWM dimming accuracy 0.1%).

[0049] An HUD software-defined dimming method based on GPU computing is mainly a real-time dimming algorithm set in the Computer Shader calculation unit, which specifically includes the following steps:

[0050] Step 1: Dynamic partition mapping

[0051] There are several LED lamp beads distributed on the lamp board. The rendered image data is matched with the lamp board, referring to Figure 3 。The LED lamp beads on the lamp board are divided into regions. The region type T of the region division can adopt rectangular partitioning, honeycomb partitioning, etc. In this embodiment, rectangular partitioning and honeycomb partitioning are taken as examples for illustration respectively. Specifically, the preset lamp board layout parameters include the partitioning type T, the number of partitions N, and the geometric parameter G. After the partitioning is completed, a coordinate mapping table (stored in the GPU video memory) is output for the partitioned lighting of the patterns on the rendering graph.

[0052] 1. Rectangular partitioning: Grid Division method (Grid Division) referring to Figure 3 。

[0053]

[0054] Among them, W total is the overall width of the lamp board, H total is the overall height of the lamp board, C column is the number of columns of the overall lamp board partition, and R row is the overall width of the lamp board partition. For example: at a resolution of 1440×540, when N = 48, a 6×8 grid is generated, and the single-zone size is 180×90 pixels. After the rectangular partitioning, the lamp board partition map is output. Each partition on the lamp board partition map is set with a corresponding partition number. Referring to Figure 3 as shown, the rendering graph is mapped with the lamp board partition map to obtain the partition corresponding to the position to be lit on the rendering graph.

[0055] 2. Honeycomb partitioning: Hexagonal Close Packing algorithm (Hexagonal Close Packing) referring to Figure 4 。

[0056]

[0057] 1. Among them, R is the side length of the honeycomb partition, the meaning S is the side length of the hexagon, and N column is the number of columns in the horizontal direction. For example: N = 40, the side length is 30 pixels, and the honeycomb radius R = 5. After the honeycomb partitioning, the lamp board partition map is output. Each partition on the lamp board partition map is set with a corresponding partition number. Referring to Figure 7 as shown, the rendering graph is mapped with the lamp board partition map to obtain the partition corresponding to the position to be lit on the rendering graph.

[0058] 3. Color partitioning

[0059] Pre-generate the light board partition map (RGB color gamut encoding), for example: Partition 1: RGB(255,0,0); Partition 2: RGB(0,255,0)…, Computer Shader determines the ID by querying the color-partition mapping table, supporting any complex shape;

[0060] Common area processing: Mix the colors of adjacent partitions (such as RGB(30,177,112)), triggering multiple partitions to light up simultaneously.

[0061] For example, it is illustrated with a resolution of 800*480, 40 partitions, and a honeycomb-like partition. Figure 4 It is a honeycomb-like light board partition map. Three adjacent hexagons are filled with red, green, and blue colors respectively. Each row increases or decreases the same color to ensure the uniqueness of the color, as Figure 5 shown. If the graphic is exactly at the edge of the light area, it may cause the information at the edge to be unclear. It is necessary to draw the common area of the adjacent light areas. For example, if it is 30 pixels from the edge of the current partition, it is divided into the common area. The color of the common area is the sum of the colors of the three adjacent partitions, as Figure 6 shown. The colors of three adjacent light areas are (R:30,G:0,B:0), (R:0,G:177,B:0), (R:0,G:0,B:112). For example, the color of the common area between the red and green partitions is (R:30,G:177,B:0), then the color of the common area of the three colors is (R:30,G:177,B:112). Fill the entire image in this way. After drawing this color-coded map, it only needs to be passed into the ComputerShader once at the beginning of the program. When traversing the non-black pixel points in the rendering image, check the color of the current pixel coordinates in the passed-in color partition map to know which partition it is in. Traverse the entire image in this way to obtain the final data and send it to the embedded side through IIC to set the switch of each partition.

[0062] Step 2: Pixel-level parallel computing

[0063] Traverse each pixel of the rendering image,

[0064]

[0065] Furthermore, optimize the lit edge: If the pixel is at the partition boundary (less than 30 pixels from the edge), mark the IDs of the adjacent partitions to be lit simultaneously.

[0066] Step 3: Data compression and transmission

[0067] 1. Compress the lightMap array by bit (such as 48 partitions → 6×8bit → 6 bytes);

[0068] 2. Transmitted to the embedded control layer via DMA to reduce the I2C bus load (bandwidth reduced by 75%).

[0069] Embodiment 1

[0070] Taking a rectangular 48-zone, 1440*540 resolution as an example, the light board is divided into 48 dimming zones, divided into 6 rows and 8 columns. The dimming zones are assigned ID numbers according to certain rules. For example, the dimming zone in the lower left corner is designated as zone 1 and arranged in sequence from left to right. The lower right zone is number 8, and then arranged upwards as shown in Figure 3 shown.

[0071] Through the grid segmentation method, the size of each rectangular area can be obtained as 180*90. In the Computer Shader, first traverse each pixel point of the incoming rendering image to determine whether it is a non-black pixel point. If it is a non-black pixel, then according to: column number = regionIndex % 8; row number = regionIndex / / 8; X coordinate = column number × 180; Y coordinate = row number × 90; regional position (x, y) = ((regionIndex % 8) × 180, (regionIndex / / 8) × 90), to obtain which zone the current pixel is in, and mark the current zone as 1, that is, lit. If it is a black pixel, skip the current pixel without processing. Traverse the entire image in this way. After calculation by the Computer Shader, an ordered array containing 48 elements is obtained, and each element represents the brightness state of a dimming zone (1 means lit, 0 means off). To simplify the output and reduce the data transmission volume, this array can be further simplified into an array containing 6 elements (or other suitable simplified forms), which can improve the communication efficiency, and each Bit represents the brightness state of a dimming zone. Then, send this simplified array to the embedded side to control the switch of a single dimming zone. The final result is as shown in Figure 3 shown. Corresponding to the rendering screen as shown in Figure 8 and Figure 9 shown. Figure 8 and Figure 9 For the effects of different numbers of zones at the same resolution, it can be seen that the more zones there are, the better the effect.

[0072] Inspired by the above ideal embodiment based on the present invention, through the above description, relevant staff can completely make various changes and modifications without departing from the technical idea of this invention. The technical scope of this invention is not limited to the content in the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. An HUD software-defined dimming system based on GPU computing, characterized in that This dimming system is used for the lighting control of HUD. This dimming system includes a GPU acceleration module, and a lighting controller. The lighting controller receives the dimming instruction from the GPU acceleration module and drives the LED light board to light up in zones; Specifically, the GPU acceleration module includes: a frame buffer parsing unit, which extracts the current frame image data from the rendering pipeline; a Computer Shader calculation unit, which is connected to the frame buffer parsing unit and executes a real-time dimming algorithm to achieve pixel-level parallel calculation; a data compression module, which is connected to the Computer Shader calculation unit, encodes the dimming instruction, and is connected to the lighting controller.

2. The HUD software-defined dimming system based on GPU computing according to claim 1, characterized in that, In the Computer Shader calculation unit, the light board is divided into regions. The partition types of the region division include one of rectangular partition, honeycomb partition, annular partition, and color partition.

3. The HUD software-defined dimming system based on GPU computing according to claim 2, wherein In the Computer Shader calculation unit, edge optimization is implemented: if the rendered image pixel is close to the boundary of the partition, the adjacent partition is lit at the same time.

4. The HUD software-defined dimming system based on GPU computing according to claim 3, characterized in that, If the distance between the edge of the rendered image and the boundary of the partition is less than 30 pixels, edge optimization is implemented to light the adjacent partition.

5. The HUD software-defined dimming system based on GPU computing according to claim 1, characterized in that, The data compression module compresses the lightMap array by bit and transmits it to the lighting controller through DMA.

6. A HUD software-defined dimming method based on GPU computing, characterized in that, It includes Step 1, dynamic partition mapping, matching the rendered image data with the light board, dividing the LED beads on the light board into regions, and outputting a coordinate mapping table; Step 2, pixel-level parallel calculation, traversing each pixel of the rendered image, triggering dimming if it is a non-black pixel, lighting the corresponding partition on the light board, and performing edge optimization; Step 3, compressing and transmitting the data by bit to the lighting controller.

7. The HUD software-defined dimming method based on GPU computing according to claim 6, characterized in that, In the step 1, the lamp board is partitioned in a rectangular partitioning manner, and the width of each rectangular partition is and the height is where W total is the overall width of the lamp board, H total is the overall height of the lamp board, C column is the total number of columns of the lamp board partition, and R row is the overall width of the lamp board partition.

8. The HUD software-defined dimming method based on GPU computing according to claim 6, characterized in that In the step 1, the lamp board is partitioned in a honeycomb partitioning manner, and the side length of the honeycomb partition is where S is the side length of the hexagon, and N column is the number of columns in the horizontal direction.

Images