A roll parameter monitoring method in PFD screen based on FPGA

Abstract

The present invention belongs to the field of graphic monitoring technology and relates to a method for monitoring roll parameters in a PFD screen based on an FPGA. The method comprises the following steps: step 1: acquiring a PFD display screen and performing binarization processing on the display screen; step 2: calculating the coordinates of each pixel in the binarized PFD display screen image; step 3: searching for a horizon line feature in the binarized PFD display screen image; step 4: calculating a horizon line slope based on the obtained horizon line feature; and step 5: calculating a roll angle based on the horizon line slope.

Description

Technical Field

[0001] The present invention relates to the field of graphic monitoring technology, in particular to monitoring roll in a PFD screen of an aircraft digital display, and specifically to a method for monitoring roll parameters in a PFD screen based on FPGA. Background Art

[0002] The digital display of an aircraft is an important component for realizing human-computer interaction. The digital display receives parameter information collected by various sensors through the onboard bus and draws it on the display screen through the GPU. The pilot obtains various parameter information of the aircraft through different screens of the digital display. Among them, the PFD screen is responsible for displaying the most important flight parameters, such as airspeed, pressure altitude, pitch angle, roll angle, metric Mach number, etc. Figure 1 shown.

[0003] To improve the reliability and mean time between failures of digital displays, it's necessary to monitor the display output by the GPU. Furthermore, real-time monitoring of the video data stream directly output to the GPU screen can enhance monitoring confidence. Given the limited processing resources of CPUs and GPUs, the need for real-time monitoring, and the flexibility of programming, FPGAs offer a promising approach.

[0004] Images drawn by the GPU undergo rendering and anti-aliasing processing. While these processes can make the images appear higher quality to the human eye, they actually introduce interference when it comes to FPGA-based graphics monitoring. Furthermore, the graphics boundaries exhibit different values at different roll angles, presenting major challenges in achieving high-reliability graphics monitoring and recognition using FPGAs. Summary of the Invention

[0005] Purpose of the invention: The present invention provides a roll parameter monitoring method in a PFD screen based on FPGA. The parameter monitoring method is implemented based on color boundary recognition, which can effectively resist the graphic distortion caused by GPU rendering and anti-aliasing in the video path, improve the mean time between failures, and ensure flight safety.

[0006] The technical solution of the present invention is:

[0007] A method for monitoring roll parameters in a PFD screen based on FPGA, the method comprising the following steps:

[0008] Step 1: Collect the PFD display image and perform binarization processing on the display image;

[0009] Step 2: Calculate the coordinates of each pixel in the binary PFD display image;

[0010] Step 3: Find the sky-ground line feature in the binary PFD display image;

[0011] Step 4: Calculate the slope of the sky-ground line based on the obtained sky-ground line characteristics;

[0012] Step 5: Calculate the roll angle based on the slope of the horizon.

[0013] Furthermore, in step 1, the collected PFD display image is a 24-bit color image, where the upper 8 bits represent the red component, the middle 8 bits represent the green component, and the lower 8 bits represent the blue component.

[0014] Furthermore, a special color extraction method is used to perform binarization processing;

[0015] If the red component R of a pixel is 0x59<=R<=0x65, the green component G is 0x25<=G<=0x30, and the blue component B is 0x00<=B<=0x0B, then the pixel is determined to be brown and binarized to binary 00;

[0016] If the red component R of a pixel is 0x00<=R<=0x0B, the green component G is 0x69<=G<=0x73, and the blue component B is 0xA5<=B<=0xB5, then the pixel is determined to be green and binarized to binary 01;

[0017] If the red component R of a pixel has a value of R>=0x8C, the green component G has a value of G>=0x8C, and the blue component B has a value of B>=0x8C, then the pixel is determined to be white and binarized to binary 10;

[0018] Other colors are uniformly binarized to binary 11.

[0019] Furthermore, a threshold-based method is used for binarization processing; the grayscale data of a pixel is calculated; and the grayscale data of the pixel is compared with the preset brown, blue, and white grayscale thresholds to determine the color of the pixel, and the pixel is binarized as 00, 01, 10, or 10.

[0020] Furthermore, the process of finding the ground-sky line feature is as follows: in the binary PFD display image, points that meet the ground-sky line feature are found from top to bottom and from left to right.

[0021] Next, from top to bottom and from left to right, scan each row and column of N vertical pixels as a group to determine whether there is a point in the group that meets the ground-ground line characteristics. N is 5 to 7.

[0022] Furthermore, the process of determining whether a point is on the sky-ground line is as follows: The process of determining whether a point meets the sky-ground line characteristics is as follows: From top to bottom, select five rows of pixels as a unit and scan row by row, then within a unit, from left to right, group five consecutive vertical pixels together and determine whether the pixel colors in the group meet the following characteristics: any one of the 1st to 3rd pixels has a binary value corresponding to blue, the 4th pixel has a binary value corresponding to a color other than brown, and the 5th pixel has a binary value corresponding to brown; then the coordinates of the 5th pixel in the group are taken as the point on the sky-ground line.

[0023] Furthermore, the process of determining whether a point is on the sky-ground line is as follows: The process of determining whether a point meets the sky-ground line characteristics is as follows: from top to bottom, select seven rows of pixels as a unit and scan them row by row. Then, within a unit, from left to right, each time, seven consecutive vertical rows of pixels are grouped together and determine whether the pixel colors in the group meet the following characteristics:

[0024] The first pixel's binary value corresponds to blue, any of the 2nd to 4th pixels' binary values correspond to blue, the 5th pixel's binary value corresponds to a color other than brown, and the 6th and 7th pixels' binary values all correspond to brown; the coordinates of the 6th pixel in this group are taken as the point on the sky-ground line.

[0025] Furthermore, in step 4, a linear fit is performed on all points on the obtained sky-ground line, and the slope of the fitted line is used as the slope K of the sky-ground line.

[0026] Furthermore, in step 5, arctan(K) is calculated based on the slope to obtain the roll angle; when the horizontal line coincides with the horizon line after being rotated clockwise by an acute angle, the roll angle is negative, otherwise the roll angle is positive.

[0027] The present invention has the following beneficial effects:

[0028] 1. Based on the actual display screen, the colors representing the sky and the earth are binarized separately;

[0029] 2. Then perform detection based on the characteristics of the sky and earth line to find the points that meet the characteristics of the sky and earth line.

[0030] 3. Calculate the slope based on the characteristic points of left 8 and right 8, and then get the roll value based on the lookup table. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 This is a schematic diagram of the PFD screen and roll angle display;

[0032] Figure 2 Schematic diagram of detection of feature points on the sky-ground line;

[0033] Figure 3This is a principle block diagram of a roll parameter monitoring system in a PFD screen implemented based on FPGA;

[0034] Figure 4 Schematic diagram of the correspondence between pixels and coordinate points in an image and the video line and field synchronization signals. DETAILED DESCRIPTION

[0035] The present invention is described in detail below in conjunction with specific implementation processes. The following examples will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several variations and improvements can be made without departing from the scope of the present invention. These all fall within the scope of protection of the present invention.

[0036] In actual operation, according to the display characteristics of the roll angle, the angle between the horizon and the horizon is the roll angle. With the upper left corner as the origin and the horizontal line to the right as the positive x-axis, the angle of the x-axis rotated clockwise to the horizon is an acute angle, and the roll angle is negative. Otherwise, it is positive. The roll angle in the PFD is displayed as follows: Figure 1 shown.

[0037] A roll parameter monitoring system in PFD screen based on FPGA, architecture Figure 3 As shown, FPGA performs corresponding monitoring functions according to the externally provided video data stream (DATA, VS, HS, DE) and the data information to be monitored (coordinate range, reference value), and outputs the monitoring results.

[0038] The external input video data is a 24-bit color image. First, the input video data is binarized, converting the RGB color data into the required binary data stream. Simultaneously, a counter is used to generate the image coordinate position corresponding to each video data point based on the field sync (VS), line sync (HS), and data valid signal (DE). This one-to-one correspondence between the binary pixel data and the (x, y) coordinates is fed into the module for searching the horizon feature and calculating the roll angle. Finally, the roll angle value and valid signal are obtained.

[0039] A) Binarization module

[0040] Binarization converts the input 24-bit RGB data stream into a single-bit data stream, thereby saving resources and simplifying subsequent data extraction, template matching and other processing processes.

[0041] Image binarization is essentially a filtering algorithm that removes useless background data and extracts the desired feature data. There are many different binarization algorithms available, and in practice, different binarization algorithms should be selected based on different application scenarios.

[0042] This article takes the common threshold-based binarization and special color extraction-based binarization as examples to briefly introduce their logical implementation principles.

[0043] The principle of threshold-based binarization is as follows. In the input 24-bit RGB data, the upper 8 bits represent the red component, the middle 8 bits represent the green component, and the lower 8 bits represent the blue component. First, convert the RGB data to grayscale data. There are two common conversion formulas:

[0044] Gray=R*0.299+G*0.587+B*0.114 (1)

[0045] Gray=R*0.212+G*0.715+B*0.072 (2)

[0046] The grayscale value is then compared with the set grayscale thresholds for brown, blue, and white to determine the pixel's color. The pixel is then binarized to 00, 01, 10, or 10, representing brown, blue, white, or other colors, respectively. Floating-point operations should be avoided as much as possible in FPGAs. Therefore, the multiplication coefficients in the above formula are multiplied by 1000 to convert them to integers. Accordingly, the thresholds for grayscale binarization are also multiplied by 1000 before comparison.

[0047] The principle of binarization based on color extraction is shown in Table 1. The red component, green component, and blue component of the input 24-bit RGB data are compared with the pre-set color component values. If the red component R of a pixel is 0x59<=R<=0x65, the green component G is 0x25<=G<=0x30, and the blue component B is 0x00<=B<=0x0B, then the pixel is determined to be brown and binarized to binary 00.

[0048] If the red component R of a pixel is 0x00<=R<=0x0B, the green component G is 0x69<=G<=0x73, and the blue component B is 0xA5<=B<=0xB5, then the pixel is determined to be green and binarized to binary 01;

[0049] If the red component R of a pixel has a value of R>=0x8C, the green component G has a value of G>=0x8C, and the blue component B has a value of B>=0x8C, then the pixel is determined to be white and binarized to binary 10;

[0050] Other colors are uniformly binarized to binary 11.

[0051] B) Coordinate generation module

[0052] The function of the coordinate generation module is to calculate the position coordinates of each incoming pixel data in the entire image based on the video line, field synchronization signal and data valid signal. Generally, the upper left corner of the image is selected as the coordinate origin. Taking an image of 800×600 as an example, the corresponding relationship between its coordinates and pixel points is as follows: Figure 4 shown.

[0053] Among them, the X-axis coordinate coor_x is generated according to the input video data valid signal DE and pixel clock CLK, corresponding to Figure 4 Horizontal coordinate. The rising edge of DE starts coor_x counting. While DE is high, coor_x increments by 1 on each rising edge of CLK and is cleared on a falling edge of DE, repeating this cycle. For 800×600 video data, with 800 pixels per line, the range of coor_x is [0,799].

[0054] The generation sequence of the Y-axis coordinate coor_y is as follows Figure 4 As shown vertically. At the beginning of each video frame, i.e., on the rising edge of the VS signal, coor_y is started and cleared. Then, on each falling edge of DE, coor_y is incremented by 1 until it is cleared to 0 on the rising edge of VS in the next video frame. This cycle repeats. For 800×600 video data, with a total of 600 lines of valid data, the range of coor_y is [0,599].

[0055] C) Feature point search and roll angle calculation

[0056] According to Figure 1 The feature points shown are searched for the sky, earth and sky line features on the entire PFD page, from left to right and from top to bottom; find points in 7 consecutive rows in the same column that meet the following features at the same time, such as Figure 2 As shown in the figure: the first row is the color of the sky after binarization, any of the 2nd to 4th rows is the color of the sky after binarization, the 5th row is not the color of the ground after binarization, and the 6th and 7th rows are the colors of the ground after binarization. The coordinates of the 6th row are the points on the sky-sky line.

[0057] Then perform linear fitting on all points on the sky-ground line and find the slope;

[0058] Then, in order to simplify and avoid wasting resources, only the 8 points on the leftmost Horizontal Line and the 8 points on the rightmost Horizontal Line are selected for linear fitting to obtain the slope;

[0059] Then, the inverse function is obtained based on the slope to obtain the value of the roll angle;

[0060] You can also use the arctan lookup table to substitute the slope to find the roll angle.

[0061] The definition of the ground-ground line feature is that seven consecutive rows in the same column meet the relevant features. It can also be defined using five rows, that is, excluding the original first and last row features.

[0062] Lookup table creation:

[0063] Initialize a 16-bit x 2048 ROM to store the roll value (the detection range is [-60, 60]). When the ROM address is 0, the initial value of tan(roll) is 0. When the ROM address is 1, tan(roll) = 2 / 2047. When the read address is 2, tan(roll) = 2x2 / 2047. Similarly, when the read address is 2047, tan(roll) = 2047*2 / 2047. This way, a roll value between 0 and arctan2 is obtained.

Claims

1. A method for monitoring roll parameters in a PFD image based on FPGA, characterized by: The method comprises the following steps: Step 1: Capture the PFD display screen and perform binarization on the display screen; the captured PFD display screen is a 24-bit color image, with the upper 8 bits representing the red component, the middle 8 bits representing the green component, and the lower 8 bits representing the blue component; binarization is performed using a special color extraction method; if the red component R value of a pixel is 0x59 <= R <= 0x65, the green component G value is 0x25 <= G <= 0x30, and the blue component B value is 0x00 <= B <= 0x0B, then the pixel is judged to be brown and binarized to binary 00; if the red component R value of a pixel is 0x00 <= R <= 0x0B, the green component G value is 0x69 <=G <= 0x73, and the blue component B value is 0xA5 <= B <= 0xB5, then the pixel is judged to be green and binarized to binary 01; if the red component R value of a pixel is R >= 0x8C, the green component G value is G >= 0x8C, if the blue component B value is B>=0x8C, the pixel is determined to be white and binarized to binary 10; other colors are uniformly binarized to binary 11; Step 2: Calculate the coordinates of each pixel in the binary PFD display image; Step 3: Find the sky-ground line feature in the binary PFD display image; Step 4: Calculate the slope of the sky-ground line based on the obtained sky-ground line characteristics; Step 5: Calculate the roll angle based on the slope of the horizon.

2. The method according to claim 1, wherein: Use a threshold-based approach to perform binarization processing; calculate the grayscale data of a pixel; and compare the grayscale data of the pixel with the preset brown, blue, and white grayscale thresholds to determine the color of the pixel and binarize the pixel to 00, 01, 10, or 11.

3. The method according to claim 1, wherein: The process of finding the ground-ground line feature is as follows: in the binary PFD display image, points that meet the ground-ground line feature are found from top to bottom and from left to right.

4. The method according to claim 3, wherein: The process of finding points that meet the characteristics of the sky and the earth line is as follows: from top to bottom, select five rows of pixels as a unit and scan them line by line. Then, within a unit, from left to right, take five consecutive vertical pixels as a group and determine whether the pixel colors in the group meet the following characteristics: If any one of the 1-3 pixels has a binary value corresponding to blue, the 4th pixel has a binary value corresponding to a color other than brown, and the 5th pixel has a binary value corresponding to brown, the coordinates of the 5th pixel in the group are taken as the point that meets the sky-ground line feature.

5. The method according to claim 3, wherein: The process for finding points that meet the ground-ground line characteristics is as follows: From top to bottom, select seven rows of pixels as a unit and scan them row by row. Then, within each unit, from left to right, group seven consecutive vertical pixels together and determine whether the pixel colors in that group meet the following characteristics: The first pixel's binary value corresponds to blue, any of the 2nd to 4th pixels' binary values correspond to blue, the 5th pixel's binary value corresponds to a color other than brown, and the 6th and 7th pixels' binary values all correspond to brown. The coordinates of the 6th pixel in this group are taken as the point that meets the sky-ground line feature.

6. The method according to claim 3, wherein: In the step 4, linear fitting is performed on all points that meet the characteristics of the sky-ground line, and the slope of the fitted line is used as the slope K of the sky-ground line.

7. The method according to claim 6, characterized in that: In step 5, arctan(K) is calculated based on the slope to obtain the roll angle. When the horizontal line coincides with the horizon line after being rotated clockwise by an acute angle, the roll angle is negative; otherwise, the roll angle is positive.

Images