A PFD picture empty speed parameter monitoring method based on FPGA implementation

Abstract

The present application belongs to the technical field of graphic monitoring, and relates to a PFD picture airspeed parameter monitoring method based on FPGA, comprising the following steps: step one: making an airspeed character matching template; step two: collecting a PFD picture airspeed display image, and storing the image after binarization processing; step three: matching the stored airspeed image with the airspeed character matching template based on sliding error matching, to obtain the coordinates of the airspeed image corresponding to the minimum matching error on the airspeed character matching template; step four: judging the numerical characters displayed in the airspeed image according to the coordinates; and step five: obtaining and outputting the airspeed value according to the numerical characters and the summation of the weights of the respective numerical characters.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of graphic monitoring, and particularly aims at the monitoring of airspeed in PFD picture of digital display of an aircraft, and more particularly relates to a method for monitoring airspeed parameter in PFD picture based on FPGA. BACKGROUND

[0002] The digital display of an aircraft is an important component for realizing man-machine interaction. The digital display receives parameter information collected by various sensors through an onboard bus, and draws the information on a display screen through a GPU. The pilot obtains various parameter information of the aircraft through different pictures of the digital display, and the PFD picture undertakes the display of the most important flight parameters, such as airspeed, barometric altitude, pitch angle, roll angle, and metric Mach number, as shown in FIG. 1. Figure 1

[0003] In order to improve the reliability of the digital display and the mean time between failures, it is necessary to monitor the display picture output by the GPU. Furthermore, real-time monitoring of the video data stream directly output by the GPU picture can improve the confidence of the monitoring. Considering the limited CPU and GPU processing resources, the real-time monitoring requirements, and the flexibility of programming, the use of FPGA is a very promising method.

[0004] The image drawn by the GPU is subjected to rendering and anti-aliasing processing. Although these processes can make the picture look better to the human eye, they introduce interference for the graphic monitoring by the FPGA. Moreover, the airspeed parameter is characterized by rolling display, and at a certain moment, there will be two numbers in the airspeed display window, which will interfere with the correct identification of the airspeed parameter. These are the main difficulties in the identification of the graphic monitoring by the FPGA with high reliability. SUMMARY

[0005] The present application provides a method for monitoring airspeed parameter in PFD picture based on FPGA, which is realized based on minimum template matching error and can effectively resist the graphic distortion caused by GPU rendering and anti-aliasing in the video path. The correct airspeed number or airspeed fault flag is extracted in the rolling airspeed display window, the display reliability is improved, the mean time between failures is improved, and the flight safety is ensured.

[0006] The technical solution of the present application is as follows:

[0007] A method for monitoring airspeed parameter in PFD picture based on FPGA, comprising the following steps:

[0008] Step 1: making an airspeed character matching template;

[0009] ​Step two: collect the PFD picture airspeed display image, and store the binarized airspeed display image;

[0010] Step three: match the stored airspeed image with the airspeed character matching template based on sliding error matching, and obtain the coordinates of the airspeed image corresponding to the minimum matching error on the airspeed character matching template;

[0011] Step four: determine the numerical characters displayed in the airspeed image according to the coordinates;

[0012] Step five: obtain and output the airspeed value by summing the numerical characters and the weight of each numerical character.

[0013] Further, in step one, the airspeed character matching template includes thirteen vertically arranged characters, from top to bottom, 0, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, 1, and "block"; "block" represents a white rectangle with diagonal stripes.

[0014] Further, in step one, the airspeed character matching template is obtained by sequentially making character templates of each character and then sequentially splicing the character templates of each character according to the character display interval in the PFD picture,

[0015] Any character template is made by the following method:

[0016] Collect different images corresponding to the same character under different display conditions, set the initial value of the binarization threshold, and perform binarization processing on different images to obtain multiple binarized images, select one binarized image as the character template sample, and match it with other binarized images of the character, when the total number of points that cannot be matched between any pair of binarized images does not exceed 5% of the total points of the character template sample, select the character template as the final template, otherwise adjust the initial value of the binarization threshold or select a new binarized image as the character template sample for iterative matching until the requirement is met.

[0017] Further, in step two, the PFD picture is collected and binarized, and the binarized image including only the airspeed numerical value is cropped out, and the binarized images of the hundreds, tens, and units are respectively stored.

[0018] Further, in step three, the binarized images of the hundreds, tens, and units are respectively matched with the airspeed character matching template by sliding error matching;

[0019] Sliding on the air speed character matching template in the direction from top to bottom, calculating the matching error between the binary image and the air speed character matching template every time a pixel point is slid, and recording the longitudinal position of the binary image and the air speed character matching template at this time, when the binary image is slid to the bottom, the matching is completed, and the longitudinal position coordinate Y corresponding to the minimum matching error is output 百 , Y 十 , Y 个 .

[0020] Further, in step four, first, the unit place longitudinal position coordinate Y 个 of the unit place binary image and the air speed character matching template is matched to determine the character corresponding to the unit place; then, the judgment result of the unit place character and the longitudinal coordinate position Y 十 , the longitudinal coordinate position Y 百 of the ten place and the hundred place are used to determine the characters corresponding to the ten place and the hundred place respectively.

[0021] Further, in step four, the judgment process of the character corresponding to the unit place is as follows:

[0022] The unit place longitudinal position coordinate Y 个 is compared with the unit place longitudinal coordinate threshold division result in the air speed character matching template to obtain the unit place character recognition result, as shown in the following table:

[0023]

[0024] Wherein, O_10 represents the starting longitudinal coordinate of the first 0 character in the air speed character matching template; 0_i represents the starting longitudinal coordinate of the character i, i takes the value of 9-1, 0_0 represents the starting longitudinal coordinate of the second 0 character; 0_END represents the coordinate between the second 0 character and the second 1 character.

[0025] Further, in step four, according to the judgment result of the unit place character, the longitudinal coordinate position Y 十 , the longitudinal coordinate position Y 百 of the ten place and the hundred place are compared with the non-unit place longitudinal coordinate threshold division result in the air speed character matching template, and the judgment of the characters corresponding to the ten place and the hundred place is as shown in the following table:

[0026]

[0027] H / D_10_1 is the starting longitudinal coordinate of the first 0 character in the air speed character matching template, and H / D_10_2 is the longitudinal coordinate position of the lower part divided by 7 / 3 along the longitudinal direction of the first 0 character;

[0028] H / D_i_1 and H / D_i_2 are the longitudinal coordinate positions of dividing the character i in the airspeed character matching template along the longitudinal direction 3 / 4 / 3; H / D_0_1 and H / D_0_2 are the longitudinal coordinate positions of dividing the second 0 character in the airspeed character matching template along the longitudinal direction 3 / 4 / 3; H / D_H1_1 and H / D_H1_2 are the longitudinal coordinate positions of dividing the second 1 character in the airspeed character matching template along the longitudinal direction 3 / 4 / 3; H / D_B_1 is the longitudinal coordinate position of dividing the block character in the airspeed character matching template 3 / 7, and H / D_B_2 is the end longitudinal coordinate of the block character in the airspeed character matching template.

[0029] Further, the step one further comprises: making an airspeed fault mark template, the airspeed fault mark being SPD; the airspeed fault mark template only comprising the character SP; and the airspeed fault mark template being made in the same way as the airspeed character matching template.

[0030] Further, the method further comprises: step six, collecting an airspeed display image of a PFD picture and performing a binaryzation processing, matching the binaryzated airspeed display image with the airspeed fault mark template based on a sliding error matching, when the minimum matching error is higher than 10% of the total character points of the airspeed fault mark template, determining that the airspeed fault mark is not displayed, and the airspeed is normal; otherwise, determining that the airspeed fault mark is displayed, and the airspeed is abnormal.

[0031] The present application has the following beneficial effects:

[0032] 1. The actual display picture is taken as a basis to obtain a specially-made airspeed identification pattern, and the pattern can effectively resist the problems of blurred pattern boundary and slightly deviated pattern display position after binaryzation caused by rendering and anti-aliasing processing during GPU drawing.

[0033] 2. The binaryzated airspeed image is matched in the airspeed template by sliding, and the values of the hundreds, tens and units of the airspeed are obtained according to the minimum matching error, or the airspeed fault mark is recognized. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 Fig. 1 is a schematic diagram of a PFD picture and airspeed display.

[0035] Figure 2 Fig. 2 is a principle block diagram of an airspeed parameter monitoring system realized based on FPGA in a PFD picture.

[0036] Figure 3 Fig. 3 is a matching schematic diagram of an airspeed character matching template.

[0037] Figure 4 Fig. 4 is a schematic diagram of airspeed character scrolling display.

[0038] Figure 5 schematic diagram for airspeed fault flag display;

[0039] Figure 6 schematic diagram for airspeed character matching template;

[0040] Figure 7 schematic diagram for airspeed fault flag template;

[0041] Figure 8 schematic diagram for airspeed value recognition module. DETAILED DESCRIPTION

[0042] The application will be described in detail below in conjunction with the specific implementation process. The following examples will help those skilled in the art to further understand the application, but do not limit the application in any form. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the application. These are all within the scope of protection of the application.

[0043] The airspeed parameter monitoring system, as shown in Figure 2 , is composed of three parts, an airspeed value recognition module, an airspeed calculation output module, and an airspeed fault flag recognition module.

[0044] The airspeed value recognition module includes: a hundred, ten, and unit airspeed value binary image cache RAM, an airspeed value recognition template ROM, a graphic sliding matching and matching error calculation function, and a minimum template matching error position extraction function. The principle block diagram is shown in Figure 8 , and the template matching process is shown in Figure 3 .

[0045] The airspeed calculation output module includes: a minimum template matching error position-value solving function and an airspeed calculation function.

[0046] The airspeed fault flag recognition module includes: an airspeed display window binary image cache RAM, an airspeed fault flag template ROM, and a graphic sliding matching and matching error function.

[0047] The airspeed value recognition module:

[0048] 1) The airspeed display window image input to the airspeed value recognition module has been binarized, and the binarization method is consistent with that used when the airspeed template is made.

[0049] 2) The hundred, ten, and unit airspeed value binary image cache RAMs respectively store the images of the corresponding regions of the hundred, ten, and unit airspeed values in the airspeed display window.

[0050] 3) The airspeed value recognition template ROM is initialized when the FPGA is powered on, and stores the airspeed value recognition template.

[0051] 4) When sliding matching, read the binary pattern from the RAM of the hundred, ten and unit airspeed value binary pattern cache respectively, and slide on the airspeed value recognition template from left to right and from top to bottom. Calculate the matching error between the airspeed image and the airspeed template once every 1 pixel point. The matching error is the number of unmatched characters in the binary image and the template. For example, Figure 3

[0052] 5) When the template matching error is obtained by sliding matching each time, record the vertical position of the binary airspeed image on the airspeed value recognition template at this time. When the airspeed image slides to the lower right corner of the airspeed value recognition template, the sliding is completed. According to the cumulative results in the sliding matching process, output the vertical coordinate of the airspeed image on the airspeed value recognition template corresponding to the minimum template matching error.

[0053] 6) In the above process, the hundreds, tens and units are independently recognized.

[0054] Airspeed calculation output module

[0055] 1) According to the vertical coordinate of the minimum template matching error of the hundreds, tens and units output by the airspeed value recognition module on the airspeed value recognition template, obtain the number represented by the airspeed character at this position.

[0056] 2) Further, since the airspeed value display effect is vertical scrolling, there is a case where there are two groups of values in a display window, as shown in Figure 4 At this time, the corresponding airspeed value judgment needs to be performed, such as Table 1 and Table 2.

[0057] 3) After obtaining the characters at each position, calculate the final airspeed value by summing the weights of the hundreds, tens and units.

[0058] Airspeed fault flag recognition module:

[0059] The working principle is the same as that of the airspeed value recognition module, except that only the value of the minimum template matching error needs to be recorded, and the position of the image on the template does not need to be known. Whether the minimum matching error is less than the threshold value is used to judge whether there is an airspeed fault flag.

[0060] Production of airspeed recognition architecture image matching template:

[0061] ​The template is made by collecting the same character images in different display conditions, determining the binary threshold, and binarizing the character images. The binarization result of one image is selected as the character template of the number, and is matched with the binarization images of other same character images. When the total number of points that cannot be matched between any pair of images does not exceed 5% of the total points of the character template, the current character template is selected for subsequent airspeed value recognition template making. If it does not meet the requirement, the binarization threshold is changed and the binarization result of a new image is selected as the character template for iteration until the requirement of 5% is met.

[0062] When the character templates of all characters are obtained, the character templates are spliced according to the spacing between the characters in the rolling display of the airspeed value, to obtain an airspeed character matching template.

[0063] The airspeed character matching template contains 13 characters in total, 11 different characters, and is composed of 0->9->8->…->2->1->0->1->block characters from top to bottom. The airspeed character matching template is shown in Figure 6 The block is a white rectangle with diagonal stripes.

[0064] The airspeed fault flag and the airspeed fault flag template are shown in Figure 5 and Figure 7 In order to save the airspeed display window binary pattern cache RAM and the airspeed fault flag template ROM storage space, the character template of the airspeed fault flag only contains “S” and “P”.

[0065] 1) In the iterative template making method, when the airspeed value recognition template is spliced from single character templates, the distance between the horizontal and vertical boundaries of the airspeed value recognition template and the center of the character should be greater than the distance between the boundaries and the character in the airspeed character binarization image, so that the character image can slide left and right on the template when matching. This making method is also applicable to the making of the airspeed fault flag template.

[0066] 2) The airspeed character matching template is composed of 0->9->8->…->2->1->0->1->block characters from top to bottom, and the center spacing between each character is equal to the center spacing between the characters when the airspeed displayed on the PFD screen is rolling.

[0067] 3) Recognition of the hundreds, tens and units of the airspeed value, especially when the airspeed display is in a rolling state.

[0068] In making the airspeed value recognition template, the mapping relationship between the vertical coordinates of the template and the number characters is obtained. Further, the mapping relationship between the vertical coordinates and the hundreds and tens of the number characters also needs to add the value of the units to distinguish the carry and the borrow.

[0069] Based on the minimum template matching error obtained by the sliding matching, the longitudinal position of the current airspeed character image on the airspeed value recognition template is obtained, and according to the mapping relationship between the template longitudinal coordinate and the digital character, the current airspeed character is obtained. Further, when the airspeed value is in a rolling state, the minimum error matching position is between the two characters in the template, and the ten-bit and hundred-bit character values need to be combined with the one-bit character value for judgment, and the one-bit character does not need to be processed in this way.

[0070] 1) When the template is made, a mapping relationship table about the template longitudinal coordinate and the character value is formed, and the displayed character at this time is obtained according to the longitudinal coordinate of the minimum template matching error in the template;

[0071] 2) When there is a rolling situation, the longitudinal coordinate of the minimum matching error of the airspeed image in the template is between two characters, and according to the one-bit value, the hundred-bit and ten-bit characters are determined, which can refer to Table 1, Table 2 and Figure 6 , wherein the scales in Table 1, Table 2 and Figure 6 are obtained according to the sensor airspeed value obtained during GPU drawing and the actual picture display.

[0072] Table 1 Hundred-bit / ten-bit airspeed value recognition template minimum template matching error longitudinal coordinate-airspeed value mapping table

[0073]

[0074] Table 2 One-bit airspeed value recognition template minimum template matching error longitudinal coordinate-airspeed value mapping table

[0075]

[0076]

[0077] Airspeed fault flag recognition

[0078] Based on the sliding matching, whether the airspeed fault flag is detected is judged according to whether the minimum matching error is less than a threshold value.

[0079] When the minimum matching error is higher than the 10% threshold value of the total character points of the template, it is considered that there is no airspeed fault flag at this time.

[0080] Embodiment

[0081] The airspeed parameter monitoring method described in the application has the following specific steps,

[0082] I. Template making:

[0083] Collect the same character image in different display conditions, determine the binary threshold, and select the binary result of an image as the character template of the number. Match the binary image of the same character with other binary images. When the total number of points that cannot be matched between any pair of images does not exceed 5% of the total points of the character template, select the current character template for subsequent airspeed value recognition template making. If it does not meet the requirement, change the binary threshold and select a new image binary result as the character template for iteration until the 5% requirement is met.

[0084] When all the character templates of the characters are obtained, the character templates are spliced according to the spacing between the characters in the rolling display of the airspeed value to obtain an airspeed character matching template.

[0085] The airspeed character matching template contains 13 characters in total, 11 different characters, and is composed of 0->9->8->…->2->1->0->1->block characters from top to bottom. The airspeed character matching template is shown in Figure 6 The block is a white rectangle with diagonal stripes.

[0086] The airspeed fault flag and the airspeed fault flag template are shown in Figure 5 and Figure 7 To save the airspeed display window binary pattern buffer RAM and the airspeed fault flag template ROM storage space, the character template of the airspeed fault flag only contains "S" and "P".

[0087] In the making of the airspeed value recognition template, the mapping relationship between the template longitudinal coordinates and the number characters is shown in Table 1 and Table 2. Further, the mapping relationship between the longitudinal coordinates and the hundreds and tens of number characters also needs to add the value of the units to distinguish the carry and the retreat.

[0088] II. Airspeed character and fault recognition

[0089] In the FPGA initialization, the ROMs in the airspeed value recognition module and the airspeed fault flag recognition module that store the airspeed value character templates and the airspeed fault flag templates are initialized. The templates should be larger than the cutout area.

[0090] After the PFD picture is processed by binaryzation (the binaryzation method is consistent with that used in template making), the binary image containing only the airspeed area and the coordinates of the image are transmitted to:

[0091] 1. Airspeed value recognition module: The input video data is cut into hundreds, tens, and units, and exists in the respective RAMs. When the airspeed area is stored, the airspeed value comparison start flag is generated.

[0092] 2. The airspeed fault flag recognition module, the input video stream is stored in a separate RAM; when the airspeed region is stored, the airspeed fault flag comparison start flag is generated;

[0093] When the airspeed value recognition module generates the airspeed value comparison start flag, taking the recognition of the unit value as an example, the airspeed region image is read from the binary image cache RAM, and is slid on the airspeed value template in the direction from left to right and from top to bottom. The matching error between the airspeed image and the airspeed template is calculated once every 1 pixel point. When the matching error is obtained every time, the longitudinal position of the airspeed image on the template at this time is recorded. When the airspeed image is slid to the right bottom corner of the template, the sliding is completed, and the longitudinal coordinate of the airspeed image on the template corresponding to the minimum matching error is output.

[0094] After the airspeed calculation output module receives the longitudinal coordinate of the airspeed image on the template corresponding to the minimum matching error of the airspeed character, the value characters of the hundreds place, the tens place and the unit place are obtained according to the scale shown in FIG. 3 and Table 1 and Table 2. Figure 6 Similarly, the value characters of the hundreds place and the tens place are obtained according to FIG. 3 and Table 1, and the current airspeed value is obtained by summing the weights of the hundreds place, the tens place and the unit place, and is output. Figure 6

[0095] The airspeed fault flag recognition module reads the airspeed region image from the binary image cache RAM, and slides on the airspeed fault flag template in the direction from left to right and from top to bottom. The matching error between the airspeed image and the airspeed template is calculated once every 1 pixel point. When the minimum matching error is higher than the 10% threshold of the total character point number of the template, it is considered that there is no airspeed fault flag at this time, and the corresponding recognition result is output.

[0096] The preferred embodiments disclosed herein are only preferred embodiments of the present application. The embodiments are selected and specifically described in the specification in order to better explain the principles and practical applications of the present application, and are not limitations of the present application. Any modifications and changes made by those skilled in the art within the scope of the specification shall fall within the scope of the present application.​

Claims

1. A method for monitoring airspeed parameters in a PFD image based on FPGA implementation, characterized in that: The method includes the following steps: Step 1: Create an airspeed character matching template; Step 2: Acquire the airspeed display image on the PFD screen, binarize the airspeed display image, and then store it; Step 3: Based on the sliding error matching, match the stored airspeed image with the airspeed character matching template to obtain the coordinates of the airspeed image corresponding to the minimum matching error on the airspeed character matching template; specifically as follows: Binary images of the hundreds, tens, and units digits are matched with an airspeed character matching template using a sliding error method. The template is slid downwards, and the matching error between the binarized image and the template is calculated for each pixel slid. The vertical position of the binarized image relative to the template is recorded at this point. When the binarized image reaches the bottom, the matching is complete, and the vertical coordinate Y corresponding to the minimum matching error is output. 百 Y 十 Y 个 ; Step 4: Determine the numerical characters displayed in the airspeed image based on the coordinates; first, determine the vertical position coordinate Y of the units digit based on the matching template between the binarized image of the units digit and the airspeed character. 个 Determine the character corresponding to the units digit; then, based on the result of determining the units digit and the vertical coordinate Y of the tens digit... 十 Hundreds digit vertical coordinate position Y 百 Determine the characters corresponding to the tens and hundreds places separately; based on the vertical coordinate Y of the units place. 个 The results of the unit digit ordinate thresholding in the airspeed character matching template are compared to obtain the unit digit character recognition result; based on the unit digit character judgment result, the tens digit ordinate position Y is determined. 十 Hundreds digit vertical coordinate position Y 百 Compare the results of the non-unit ordinate thresholding in the airspeed character matching template to determine the character corresponding to the tens or hundreds place. Step 5: Calculate the airspeed value by summing the numerical characters and their respective weights, and then output the result.

2. The method according to claim 1, characterized in that: In step one, the airspeed character matching template includes thirteen vertically arranged characters, from top to bottom: 0, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0, 1, "block"; "block" represents a white rectangle with diagonal stripes.

3. The method according to claim 2, characterized in that: In step one, character templates for each character are created sequentially, and then the character templates are sequentially spliced ​​together according to the character display spacing in the PFD screen to obtain the airspeed character matching template. A character template for any given character is created in the following way: Collect different images of the same character under different display conditions, set an initial value for the binarization threshold, and perform binarization processing on different images to obtain multiple binarized images. Select one binarized image as the character template sample for the character, and match it with other binarized images of the character. When the total number of unmatched points between any pair of binarized images does not exceed 5% of the total number of points in the character template sample, select the character template as the final template. Otherwise, adjust the initial value of the binarization threshold or select a new binarized image as the character template sample for iterative matching until the requirements are met.

4. The method according to claim 3, characterized in that: In step two, the PFD image is captured and binarized, and a binarized image containing only the airspeed value is cropped out. The binarized images of the hundreds, tens, and units digits are extracted and stored separately.

5. The method according to claim 4, characterized in that: In step four, the character corresponding to the units digit is determined as shown in the table below: Where O_10 represents the starting ordinate of the first 0 character in the airspeed character matching template; 0_i represents the starting ordinate of character i, where i takes values ​​from 9 to 1; 0_0 represents the starting ordinate of the second 0 character; and 0_END represents the coordinates between the second 0 character and the second 1 character.

6. The method according to claim 5, characterized in that: In step four, the character corresponding to the tens or hundreds place is determined as shown in the table below: H / D_10_1 is the starting ordinate of the first 0 character in the airspeed character matching template, and H / D_10_2 is the ordinate position of the first 0 character divided along the vertical 7 / 3 and located at the bottom. H / D_i_1 and H / D_i_2 are the vertical coordinates of the character 'i' in the airspeed character matching template, divided along the vertical axis of 3 / 4 / 3; H / D_0_1 and H / D_0_2 are the vertical coordinates of the second '0' character in the airspeed character matching template, divided along the vertical axis of 3 / 4 / 3; H / D_H1_1 and H / D_H1_2 are the vertical coordinates of the second '1' character in the airspeed character matching template, divided along the vertical axis of 3 / 4 / 3; H / D_B_1 is the vertical coordinate of the block character in the airspeed character matching template, divided along the vertical axis of 3 / 7; and H / D_B_2 is the vertical coordinate of the end of the block character in the airspeed character matching template.

7. The method according to claim 6, characterized in that: Step one further includes: creating an airspeed fault sign template, wherein the airspeed fault sign is SPD; the airspeed fault sign template contains only the character SP; the method of creating the airspeed fault sign template is the same as that of creating the airspeed character matching template.

8. The method according to claim 7, characterized in that: The method further includes: Step 6: Acquire the airspeed display image on the PFD screen and perform binarization processing. Based on sliding error matching, match the binarized airspeed display image with the airspeed fault sign template. When the minimum matching error is higher than 10% of the total number of character points in the airspeed fault sign template, it is determined that no airspeed fault sign is displayed and the airspeed is normal; otherwise, it is determined that an airspeed fault sign is displayed and the airspeed is abnormal.

Images