System for quantitatively determining pavement marking quality

Abstract

A system for quantitatively determining a quality of a pavement marking disposed along a roadway includes one or more controllers in wireless communication with a plurality of vehicles. The one or more controllers receive image data collected by the plurality of vehicles representative of a pavement marking disposed along a roadway. The one or more controllers execute instructions to determine at least one of a color distance measurement between an average color space value of the pavement marking and an ideal marking color space value and a marking intensity contrast between the pavement marking and a roadway.

Description

Technical Field

[0001] This disclosure relates to a system and method for quantitatively determining the quality of road signs by measuring the color distance between the current sign color and the original or ideal sign color, and by measuring the sign intensity contrast between the road surface and the road sign. Background Technology

[0002] Road markings are used to convey information to road users. Specifically, road markings can indicate specific sections of the road to be used, provide information about the conditions ahead, and indicate where passage is permitted. For example, yellow lines separate traffic traveling in opposite directions, while white lines separate lanes traveling in the same direction.

[0003] Road markings wear down over time, necessitating periodic repainting. In some cases, subjective visual inspection can be used to determine the aesthetic quality acceptance of road markings. However, current quality inspections have a qualitative component and do not provide a uniform assessment, thus leading to variations in quality based on the perception of the specific individual performing the visual inspection. Alternatively, some government agencies and municipalities may repaint road markings based on manual intervention or a pre-set schedule rather than the current quality of the markings. Currently, there is no method to measure the degradation of the paint color or the contrast in visibility of road markings compared to surrounding road materials.

[0004] Therefore, although current road marking repainting techniques have achieved their intended purpose, there is still a need in the field for a method to objectively determine the current condition of road markings. Summary of the Invention

[0005] According to several aspects, a system for quantitatively determining the quality of road markings set along a road surface is disclosed. The system includes one or more controllers that wirelessly communicate with multiple vehicles, wherein the one or more controllers receive image data representing road markings set along a road surface collected by the multiple vehicles. The one or more controllers execute instructions to convert image frames based on the image data into grayscale image frames, wherein the grayscale image frames retain data indicating the original color space values ​​of the image frames. The one or more controllers execute instructions to create a grayscale filter by providing one or more color masks that separate only the original color space values ​​representing the road markings, and then combining the outputs of the one or more color masks. The one or more controllers execute instructions to separate the original color space values ​​representing the road markings from the grayscale image frames using the grayscale filter to determine a filtered grayscale image frame. The one or more controllers execute instructions to determine an average color space value corresponding to the road markings and an average color space value corresponding to the road surface based on the filtered grayscale image frames. The one or more controllers execute instructions to determine at least one of a color distance measurement between the average color space value of the road markings and an ideal marking color space value, and a marking intensity contrast between the road markings and the road surface.

[0006] In one aspect, one or more controllers execute instructions to filter a masked grayscale image frame by assigning binary values ​​to pixels of the filtered grayscale image frame using a Boolean mask. Pixels representing road markings are assigned a first binary value, and pixels representing the road surface are assigned a second binary value.

[0007] In another aspect, one or more controllers execute instructions to determine the average color space value of the pixels representing a first binary value of the road surface marking. One or more controllers execute instructions to determine the average color space value of the pixels representing a second binary value of the road surface marking, and identify the color represented by the average color space value representing the road surface marking.

[0008] In another aspect, one or more controllers identify the boundary between the road markings and the road surface in a filtered grayscale image frame.

[0009] In one aspect, one or more controllers execute instructions to correct the brightness of grayscale image frames in order to remove discoloration from image data representing the road surface.

[0010] In another aspect, one or more color masks include a first color mask that separates only the color space values ​​representing a first color and a second color mask that separates only the color space values ​​representing a second color.

[0011] In another aspect, the first color is yellow, and the second color is white.

[0012] In one aspect, the intensity contrast of the marker is determined by the following formula:

[0013]

[0014] Where, μ M It is the mark intensity, and μ p It refers to road surface strength.

[0015] In another scenario, one or more controllers execute instructions to create a map that plots color distance measurements of road markings at a specific geographic location, where the map provides visual indicators where road markings need to be repainted.

[0016] In another aspect, one or more controllers execute instructions to determine the dominant color space values ​​of road markings and the road surface based on the original color space values ​​of the filtered grayscale image frames. The one or more controllers determine multiple clusters, each representing the dominant color space values ​​of the road markings and the road surface. The one or more controllers determine the Euclidean distance between the average color space value of each dominant color space value of the road markings and the ideal marking color space value.

[0017] In one respect, road markings are lane markings.

[0018] In another aspect, a method for quantitatively determining the quality of road markings set along a road surface is disclosed. The method includes receiving image data representing road markings set along a road surface, collected by one or more controllers from multiple vehicles. The method further includes converting image frames based on the image data into grayscale image frames by the one or more controllers, wherein the grayscale image frames retain data indicating the original color space values ​​of the image frames. The method also includes creating a grayscale filter by providing one or more color masks that separate only the original color space values ​​representing the road markings, and then combining the outputs of the one or more color masks. The method further includes separating the original color space values ​​representing the road markings from the grayscale image frames using the grayscale filter to determine a filtered grayscale image frame. The method also includes determining an average color space value corresponding to the road markings and an average color space value corresponding to the road surface based on the filtered grayscale image frame. Finally, the method includes determining at least one of a color distance measurement between the average color space value of the road markings and an ideal marking color space value, and a marking intensity contrast between the road markings and the road surface.

[0019] In another aspect, the method includes filtering a masked grayscale image frame by assigning binary values ​​to pixels of the filtered grayscale image frame using a Boolean mask, wherein pixels representing road markings are assigned a first binary value and pixels representing road surfaces are assigned a second binary value.

[0020] In another aspect, the method includes determining an average color space value for a pixel representing a first binary value of a road marking. The method also includes determining an average color space value for a pixel representing a second binary value of the road surface, and identifying the color represented by the average color space value representing the road marking.

[0021] In one aspect, the method includes identifying the boundary between road markings and the road surface of a color-filtered grayscale image frame.

[0022] In another aspect, the method includes correcting the brightness of grayscale image frames to remove discoloration in image data representing the road surface.

[0023] In another aspect, the method includes determining the sign intensity contrast, which is determined by the following formula:

[0024]

[0025] Where, μ M It is the mark intensity, and μ p It refers to road surface strength.

[0026] In one aspect, the method includes creating a map of color distance measurements of road markings at specific geographic locations, wherein the map provides visual indicators where road markings need to be repainted.

[0027] In another aspect, the method includes determining the dominant color space values ​​of road markings and the road surface based on the original color space values ​​of a grayscale image frame filtered by color. The method also includes: determining multiple clusters, each cluster representing the dominant color space values ​​of the road markings and the road surface; and determining the Euclidean distance between the average color space value of each dominant color space value of the road markings and the ideal marking color space value.

[0028] In one aspect, a system for quantitatively determining the quality of lane markings set along a road surface is disclosed. The system includes one or more controllers that wirelessly communicate with multiple vehicles, wherein the one or more controllers receive image data representing lane markings set along the road surface collected by the multiple vehicles. The one or more controllers execute instructions to convert image frames based on the image data into grayscale image frames, wherein the grayscale image frames retain data indicating the original color space values ​​of the image frames. The one or more controllers create a grayscale filter by providing one or more color masks that separate only the original color space values ​​representing the lane markings, and then combining the outputs of the one or more color masks. The one or more controllers separate the original color space values ​​representing the lane markings from the grayscale image frames using the grayscale filter to determine a filtered grayscale image frame. The one or more controllers determine an average color space value corresponding to the lane markings and an average color space value corresponding to the road surface based on the filtered grayscale image frame. The one or more controllers determine at least one of the Euclidean distance between the average color space value of the lane markings and an ideal marking color space value, and the marking intensity contrast between the lane markings and the road surface.

[0029] Further areas of application will become apparent from the description provided herein. It should be understood that the descriptions and specific examples are for illustrative purposes only and are not intended to limit the scope of this disclosure. Attached Figure Description

[0030] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of this disclosure in any way.

[0031] Figure 1A This is a schematic diagram of a system for quantitatively determining the quality of road markings, disclosed according to an exemplary embodiment, the system including a computing system that wirelessly communicates with multiple vehicles;

[0032] Figure 1B This is a schematic diagram of road markings installed along the road surface according to an exemplary embodiment;

[0033] Figure 2 This is a schematic diagram illustrating the computing system shown in FIG1 according to an exemplary embodiment; and

[0034] Figure 3 The illustration according to the exemplary embodiment is used for determining Figure 1B The flowchart shows a method for improving the quality of road markings. Detailed Implementation

[0035] The following description is exemplary in nature and is not intended to limit this disclosure, application or use.

[0036] refer to Figure 1A and Figure 1B An exemplary system 10 is disclosed for quantitatively determining the quality of road markings 12 set along a road surface 14 of a road 16. The road surface 14 represents a surface used for vehicle or pedestrian traffic. In such a way... Figure 1B In the examples shown and described below, road marking 12 is a white or yellow dashed line indicating a center lane. However, it should be understood that these figures are merely exemplary in nature, and road marking 12 is not limited to lane markings and may include other shapes and colors. For example, in another embodiment, road marking 12 is a symbol, such as a diamond indicating a lane reserved for high-capacity vehicles or a bicycle indicating a lane reserved for cyclists. Furthermore, although road marking 12 is described, in one embodiment, system 10 may also be used to determine the quality of road signs, such as stop signs, exit ramp signs, and traffic signs.

[0037] In such Figure 1A In the example shown, system 10 includes a computing system 20, which includes one or more controllers 26 that wirelessly communicate with a fleet or multiple vehicles 22. The one or more controllers 26 receive and aggregate image data 18 collected from the multiple vehicles 22. The multiple vehicles 22 can include any type of vehicle with wireless capability to connect to the computing system 20, such as, but not limited to, cars, trucks, SUVs, vans, or motorhomes. Each vehicle 22 includes one or more cameras 24 for capturing image data 18, which represents road markings 12 positioned along road surface 14. As described below, the computing system 20 performs image processing algorithms and quantitative techniques to determine color distance measurements between the current marking color and the original or ideal marking color, and between the current road surface color and the ideal road surface color. In one embodiment, the ideal marking color and the ideal road surface color are based on standards from various government agencies, such as the Department of Transportation (DoT) of the federal and state governments. Furthermore, it should be understood that although this disclosure describes the Euclidean distance model as an exemplary use case for determining color distance measurements, other color difference formulas, such as urban block models or CIELAB color space models, may also be used. System 20 also determines the sign intensity contrast between road surface 14 and road sign 12.

[0038] The color distance measurement between the current and ideal sign colors provides an objective measure of how the paint color of the road sign deteriorates over time, and the sign intensity contrast provides an objective measure of the visibility of road sign 12 relative to road surface 14. Specifically, the color distance measurement between the current and ideal sign colors indicates how much the current sign color deviates from the ideal sign color over time. For example, a larger Euclidean distance indicates a greater degree of deterioration in the paint color of road sign 12 over time. The sign intensity contrast is measured between road sign 12 and road surface 14 and provides an objective measure of the visibility comparison between road sign 12 and road surface 14, where a higher contrast indicates higher visibility, and a lower contrast indicates lower or worse visibility.

[0039] Figure 2 This is a block diagram of the calculation system 20, which includes one or more controllers 26 for determining the color distance measurement between the current sign color and the ideal sign color, and the sign intensity contrast between the road surface 14 and the road sign 12. The calculation system 20 includes a collection module 30, a preprocessing module 32, a feature extraction module 34, a color filtering module 36, an Euclidean distance module 38, a contrast module 40, and an evaluation module 42. (Reference) Figure 1A , Figure 1B and Figure 2 The collection module 30 of the computing system 20 receives image data 18 from multiple vehicles 22, wherein the image data 18 represents road markings 12 on a specific segment of road 16. It should be understood that since the image data 18 is collected from more than one vehicle 22, there may be multiple sets of image data 18 for a specific segment of road 16. The image frame 44 is then sent to the preprocessing module 32 of the computing system 20.

[0040] Preprocessing module 32 performs one or more preprocessing techniques on image frame 44 to generate a representation road 16 based on image frame 44. Figure 1B A grayscale image frame 50 of a specific road section's road markings 12. Figure 2 In the example shown, the preprocessing module 32 includes a grayscale block 52 and a luminance block 54. The grayscale block 52 converts image frame 44 in color space values ​​into a grayscale image frame 50. It should be understood that although image frame 44 is converted to grayscale, grayscale image frame 50 still indicates the original color space values. That is, grayscale image frame 50 retains data indicating the original color space values ​​of the original image frame 44 within each pixel. Luminance correction is then performed on grayscale image frame 50 to remove discoloration in the image data representing road surface 14 by luminance block 54. In a non-limiting embodiment, luminance block 54 performs a gamma correction algorithm to remove discoloration; however, it should be understood that other techniques may also be used.

[0041] Then, the brightness-corrected grayscale image frame 50 is sent to the feature extraction module 34. In such a way... Figure 2 In the example shown, the feature extraction module 34 includes a color space block 60, a color mask block 62, a common mask 64, a grayscale filter 68, and an edge detection block 70. The color space block 60 of the feature extraction module 34 converts the grayscale image frame 50 into an image frame represented by color space values ​​72. For example, in one embodiment, the darkened grayscale image 50 is converted into an image frame, wherein the road marking 12 is separated from the road surface 14 based on the common mask 64. The common mask 64 is created to represent the image frame 44 using Hue, Saturation, and Luminance (HSL) color space values; however, it should be understood that other color spaces can also be used to determine color difference measurements and contrast, such as, but not limited to, Hue, Saturation, Value (HSV), Red-Blue-Green (RGB) color space, or Y′UV color space.

[0042] Then, the image frame represented in color space value 72 is sent to color mask block 62, which includes one or more color masks to separate the original color space values ​​representing the road markings. In one embodiment, color mask block 62 includes a first color mask 62A and a second color mask 62B; however, more than two color masks may also be used. First color mask 62A is created by separating only the color space values ​​representing the first color from the image frame represented in color space value 72, and second color mask 62B is created by separating only the color space values ​​representing the second color from the image frame represented in color space value 72. In one embodiment, since many road markings are either white or yellow, the first color is yellow, the second color is white, and the white mask corresponds to white markings, and the yellow mask corresponds to yellow markings. However, it should be understood that other colors may also be used for the first and second colors.

[0043] The first color mask 62A and the second color mask 62B are combined to create a bitwise OR mask, which is the common mask 64. A bitwise OR operation is performed to create the common mask 64 from the two-color mask (i.e., white and yellow). The common mask 64 separates only the color space values ​​representing the first or second color from the image frame represented by color space values ​​72. The outputs of the common mask 64 are combined to create a grayscale filter 68. The grayscale filter 68 separates the original color space values ​​representing the first and second colors from the grayscale image frame 50, thus serving as a bitwise AND mask. Therefore, the grayscale filter 68 is created by first providing one or more color masks that separate only the original color space values ​​representing the road marking 12 (yellow and white in this example), and then combining the outputs of one or more color masks (in this example, the output is all yellow or white road color).

[0044] Grayscale filter 68 receives a grayscale image frame 50 representing a road marking 12 for a specific segment of road 16 from preprocessing module 32, and separates the original color space values ​​representing a first color and a second color from the grayscale image frame 50 to determine a filtered grayscale image frame 80, wherein the road marking 12 is individually visible compared to the road surface 14. In this example, grayscale filter 68 separates the original color space values ​​representing yellow and white color values ​​from the grayscale image frame 50, which are common colors chosen for road markings.

[0045] The filtered grayscale image frame 80 is then sent to the edge detection block 70. In one embodiment, the filtered grayscale image frame 80 may first be filtered using any type of image denoising technique (such as Gaussian blur) before being sent to the edge detection block 70. Gaussian blur denoising has been used to reduce image noise, and in this example, it is achieved by reducing the size that may be visible in the road surface portion of the image frame. The edge detection block 70 identifies the boundary between the road surface marker 12 and the road surface 14 of the filtered grayscale image frame 80. For example, in one embodiment, the edge detection algorithm is a Canny edge detector; however, it should be understood that other algorithms may also be used. The filtered grayscale image frame 80 is then sent to the filtering module 36.

[0046] The color filtering module 36 includes a Boolean mask 82, a first color filter 84, and a second color filter 86. Based on the filtered grayscale image frame 80 received from the feature extraction module 34, the color filtering module 36 determines the average color space values ​​88 and 90 corresponding to road marking 12 and road surface 14, respectively. Specifically, the Boolean mask 82 determines the masked filtered grayscale image frame 92 by assigning binary values ​​to the pixels of the filtered grayscale image frame 80, wherein pixels representing road marking 12 are assigned a first binary value, and pixels representing road surface 14 are assigned a second binary value. For example, in one embodiment, pixels representing road marking 12 are assigned a 1, while pixels representing road surface 14 are assigned a 0. Of course, this embodiment can be reversed, such that road marking 12 is assigned a 0, and road surface 14 is assigned a 1.

[0047] Then, the first color filter 84 determines the average color space value 88 of the road sign 12 and the average color space value 90 of the road surface 14 based on the masked, filtered grayscale image frame 92. Specifically, the first color filter 84 determines the average color space value 88 corresponding to all pixels representing the first binary values ​​of the road sign 12. Then, the first color filter 84 determines the average color space value 90 corresponding to all pixels representing the second binary values ​​of the road surface 14. The average color space values ​​88 and 90 are then sent to the second color filter 86. The second color filter 86 identifies the color represented by the average color space value 88 representing the road sign 12. For example, if the road sign 12 is yellow, then the second color filter 86 determines that the color represented by the average color space value of the road sign 12 is yellow.

[0048] Then, the average color space values ​​88 and 90, and the color representing the average color space value 88 of the road marking 12, are sent to both the Euclidean distance module 38 and the contrast module 40. The Euclidean distance module 38 determines the Euclidean distance between the average color space value 88 and the ideal sign color space value. It should be understood that HSL, RGB, and Y′UV color spaces can be used to calculate the Euclidean distance and color space values. In one embodiment, the Euclidean distance is determined based on Equations 1-3, as follows:

[0049] Formula 1

[0050]

[0051] Formula 2

[0052]

[0053] Formula 3

[0054]

[0055] Wherein, if the average color space value 88 is expressed in RGB color space, then Equation 1 is used; if the average color space value 88 is expressed in HSL or HSV color space, then Equation 2 is used; if the average color space value 88 is expressed in Y′UV color space, then Equation 3 is used. LM represents the average color space value 88 of road marking 12, I represents the ideal marking color, R, G, B represent the values ​​of red, green and blue, H represents the hue value, V represents the value, L represents the brightness, Y is the brightness component, and U and V represent the chromaticity components.

[0056] The contrast module 40 determines the sign intensity contrast between road surface 14 and road surface sign 12 based on the average color space value 88 of road surface sign 12 and the average color space value 90 of road surface 14, where the average color space values ​​88 and 90 represent the average intensity of the target sign. Sign intensity contrast measures the difference between the intensity of road surface sign 12 and the intensity of the surrounding background area (such as road surface 14 in this example). Therefore, a higher sign intensity contrast value indicates that it is easier for individuals and autonomous vehicles to perceive road surface sign 12. In one embodiment, the sign intensity contrast is determined based on the road surface sign intensity value and the road surface intensity value, and specifically, it is determined by Equation 4, which is:

[0057]

[0058] Where, μ M It is the mark intensity, and μ p It refers to road surface strength.

[0059] In one embodiment, one or more controllers 26 of system 20 receive data indicating the geographic location of road marking 12. For example, in one embodiment, Global Positioning System (GPS) coordinates are included in image data 18 transmitted from multiple vehicles 22 to one or more controllers 26. Figure 1A One or more evaluation modules 42 of controller 26 create a map 98 that plots the Euclidean distance 38 and sign intensity contrast of road markings 12 at a specific geographic location, wherein map 98 provides visual indicators of road marking quality and visibility. For example, map 98 can indicate where lane markings on a specific segment or length of road need to be repainted or re-marked to improve road marking quality and / or visibility.

[0060] Figure 3 This shows the determination along road 16 ( Figure 1B A flowchart of an exemplary method 200 for improving the quality of road markings 12 on road surface 14 is provided. Specifically, method 200 determines a color distance measurement between the current marking color and an ideal marking color, as well as the marking intensity contrast between road surface 14 and road markings 12. General Reference Figure 1A , Figure 1B , Figure 2 and Figure 3 Method 200 begins at block 202. In block 202, preprocessing modules 30 of one or more controllers 26 convert image frame 44 into grayscale image frame 50. As described above, grayscale image frame 50 retains data indicating the original color space values ​​of the original image frame 44. Also as described above, grayscale block 52 of preprocessing module 30 converts image frame 44 in color space values ​​into grayscale image frame 50, and then luminance block 54 of preprocessing module 30 corrects the luminance of grayscale image frame 50 to remove discoloration in the image data representing road surface 14. Method 200 can then proceed to block 204.

[0061] In block 204, the color space block 60 of the feature extraction module 34 converts the grayscale image frame 50 into an image frame represented by color space value 72. Then, method 200 can proceed to block 206.

[0062] In block 206, one or more feature extraction modules 34 of controller 26 create a grayscale filter 68 by first providing one or more color masks that separate only the raw color space values ​​representing the road marking 12, and then combining the outputs of the one or more color masks together. In this example, the first color mask 62A separates yellow, and the second color mask 62B separates white; yellow and white are common colors for road markings. Method 200 can then proceed to block 208.

[0063] In box 208, grayscale filter 68 separates the original color space values ​​representing road marking 12 from the darkened grayscale image frame 50 to determine the filtered grayscale image frame 80. In this example, grayscale filter 68 separates the original color space values ​​representing yellow and white color values, which are common colors chosen for road markings. Method 200 can then proceed to box 210.

[0064] In box 210, the edge detection block 70 of the feature extraction module 34 identifies the boundary between the road marking 12 and the road surface 14 in the color-filtered grayscale image frame 80. Then, method 200 can proceed to box 212.

[0065] In block 212, the color filter module 36 determines the average color space value 88 corresponding to road sign 12 and the average color space value 90 corresponding to road surface 14 based on the filtered grayscale image frame 80. Specifically, as described above, the Boolean mask of the color filter module 36 determines the masked filtered grayscale image frame 92 by assigning binary values ​​to the pixels of the filtered grayscale image frame 80. The first color filter 84 determines the average color space value 88 corresponding to all the binary values ​​representing road sign 12 and the average color space value 90 corresponding to all the binary values ​​representing road surface 14. The second color filter 86 identifies the color represented by the average color space value representing road sign 12. Then, method 200 can proceed to blocks 214A and 214B.

[0066] In box 214A, Euclidean distance module 38 determines the color distance measurement between the average color space value 88 and the ideal sign color space value. Equations 1-3 as described above can be used to determine the Euclidean distance. In box 214B, contrast module 40 determines the sign intensity contrast, as described in Equation 4 above. Method 200 can proceed to box 216.

[0067] In box 216, one or more controllers 26 create a map 98 that plots the Euclidean distances and sign intensity contrasts of road signs 12 at a specific geographic location. Map 98 provides visual indicators of the road sign quality and visibility of the road signs 12. For example, map 98 can indicate where road signs 12 on a specific segment or length of road need to be repainted or re-illuminated to improve road sign quality and / or visibility. Method 200 can then conclude.

[0068] Return to reference Figure 1B and Figure 2 In one alternative embodiment, the quality of the road marking 12 is determined based on a dominant color method. Specifically, in one embodiment, the system 20 further includes a dominant color module 100, which determines the dominant color space values ​​for both the road marking 12 and the road surface 14. The dominant color module 100 includes a clustering block 102, a histogram block 104, and a drawing block 106.

[0069] Clustering block 102 determines the dominant color space values ​​of road signs 12 and road surface 14 based on the original color space values ​​of the filtered grayscale image frame 80. Dominant color module 100 determines N clusters 180, each cluster representing the dominant color space value of road signs 12 and road surface 14, where the number N can be any number at least 2. Specifically, clustering block 102 determines the number N of clusters based on the original color space values ​​of the filtered grayscale image frame 80 of road signs 12 and road surface 14, where a higher value of N indicates greater color variation and degradation in road signs 12 and road surface 14. Histogram block 104 can then create a histogram showing the distribution of the N clusters 180, each cluster representing the dominant color of road signs 12 and road surface 14. Drawing block 106 can then create a color map showing all dominant colors in image frame 44. Alternatively, all original color space values ​​(i.e., RGB, HSV, or Y′UV color spaces) of the dominant colors of road signs 12 and road surface 14 can be presented.

[0070] Then, each of the N clusters 180 representing the dominant color space value is sent to the Euclidean distance module 138, which determines the Euclidean distance between each dominant color space value of the road sign 12 and the ideal sign color space value. As mentioned above, the Euclidean distance is determined based on Equations 1-3. In one embodiment, the Euclidean distance between the dominant color space value and the ideal sign color space of the road sign 12 is sent to the evaluation module 42. The evaluation module 42 creates a map 98 that plots the Euclidean distances of the road sign 12 at a specific geographic location.

[0071] Referring generally to the accompanying drawings, the disclosed system offers various technical effects and benefits. Specifically, the disclosed system provides a method for quantitatively determining the quality of road markings that does not consider human perception, which in turn can lead to uniform results, rather than being based on variations in color perception between different individuals. The disclosed method allows individual municipalities and government agencies to repaint based on specific quality standards, which in turn can reduce repainting costs. Finally, the disclosed method is able to explore perceived color variations in road markings (e.g., using dominant color space methods), which was previously impossible.

[0072] In a system-on-a-chip (SoC), a controller can refer to electronic circuitry, combinational logic circuitry, a field-programmable gate array (FPGA), a processor (shared, dedicated, or grouped) that executes code, or a combination of some or all of the above, or a portion thereof. Additionally, the controller can be microprocessor-based, such as a computer having at least one processor, memory (RAM and / or ROM), and associated input and output buses. The processor can operate under the control of an operating system residing in memory. The operating system can manage computer resources so that computer program code, embodied as one or more computer software applications, such as applications residing in memory, can have instructions executed by the processor. In alternative embodiments, the processor can directly execute the application, in which case the operating system can be omitted.

[0073] The description in this disclosure is merely exemplary in nature, and changes that do not depart from the spirit and scope of this disclosure are intended to remain within its scope. Such changes should not be considered as departing from the spirit and scope of this disclosure.

Claims

1. A system for quantitatively determining the quality of road markings installed along a road surface, the system comprising: One or more controllers that wirelessly communicate with multiple vehicles, wherein the one or more controllers receive image data collected by the multiple vehicles representing road markings set along the road surface, wherein the one or more controllers execute instructions to: The image frame based on the image data is converted into a grayscale image frame, wherein the grayscale image frame retains data indicating the original color space values ​​of the image frame; A grayscale filter is created by providing one or more color masks that separate only the original color space values ​​representing the road markings, and then combining the outputs of the one or more color masks together. The original color space values ​​representing the road markings are separated from the grayscale image frame using the grayscale filter to determine the filtered grayscale image frame. The average color space value corresponding to the road marking and the average color space value corresponding to the road surface are determined based on the filtered grayscale image frame. Determine at least one of the following: a color distance measurement between the average color space value of the road sign and the ideal color space value of the sign, and a sign intensity contrast between the road sign and the road surface; The road markings and the dominant color space values ​​of the road surface are determined based on the original color space values ​​of the filtered grayscale image frame. Multiple clusters are identified, each cluster representing the road markings and the dominant color space values ​​of the road surface; and Determine the Euclidean distance between the average color space value of each dominant color space value of the road marking and the ideal marking color space value.

2. The system according to claim 1, wherein, The one or more controllers execute instructions to: A Boolean mask is used to filter a masked grayscale image frame by assigning binary values ​​to the pixels of the filtered grayscale image frame, wherein pixels representing road markings are assigned a first binary value and pixels representing the road surface are assigned a second binary value.

3. The system according to claim 2, wherein, The one or more controllers execute instructions to: Determine the average color space value of the pixel representing the first binary value of the road marking; Determine the average color space value of the pixel representing the second binary value of the road surface; as well as The markings are represented by colors derived from the average color space values ​​of the road markings.

4. The system according to claim 1, wherein, The one or more controllers identify the boundary between the road markings and the road surface in the filtered grayscale image frame.

5. The system according to claim 1, wherein, The one or more controllers execute instructions to: The brightness of the grayscale image frame is corrected to remove color distortion in the image data representing the road surface.

6. The system according to claim 1, wherein, The one or more color masks include a first color mask that separates only the color space values ​​representing a first color and a second color mask that separates only the color space values ​​representing a second color.

7. The system according to claim 6, wherein, The first color is yellow, and the second color is white.

8. The system according to claim 1, wherein, The intensity contrast of the marker is determined by the following formula: Where, µ M It is the mark strength, and µ p It refers to road surface strength.

9. The system according to claim 1, wherein, The one or more controllers execute instructions to: Create a map that plots color distance measurements of the road signs at specific geographic locations, wherein the map provides visual indicators where the road signs need to be repainted.

Images