A method for measuring the retroreflective index of a scale based on a polarized double camera

Abstract

The application discloses a kind of mark line reflectivity measurement methods based on polarized dual camera, including following measurement steps: S101, install test assembly on the trunk lid of vehicle, S102, determine the homography matrix of two cameras, S103, control camera and artificial light source by hardware trigger, respectively collect target image with target road marking, S104, for the image shot by camera, including homographic transformation in processing, make image alignment, the present application is installed on the trunk lid of car by setting installation assembly, main control, an artificial light source, two cameras, three polarizing plates and computer, start car can be carried out shooting and image acquisition, it is more convenient to operate, the road marking reflectivity is obtained by shooting image, image segmentation and calculation, the prompt effect of road marking can be detected in time, it is convenient for people to maintain the section of poor reflective effect, so that road marking reaches ideal use effect.

Description

Technical Field

[0001] This invention relates to the field of road marking testing technology, specifically a method for measuring the reflectivity of road markings based on a polarization dual-camera system. Background Technology

[0002] Road markings are markings on the road surface that use lines, arrows, words, vertical markings, raised pavement markers, and delineators to convey traffic information such as guidance, restrictions, and warnings to road users. Their function is to regulate and guide traffic. They can be used in conjunction with signs or used alone. Road markings usually use reflective materials, which can reflect light through a light source to remind drivers to drive according to the rules. Inspection of road markings can detect damage to the markings in a timely manner, facilitating timely maintenance. However, the current road marking detection instruments are relatively expensive and bulky, making them inconvenient for widespread use, which leads to the inability to detect and maintain road markings in a timely manner. Summary of the Invention

[0003] This invention provides a method for measuring the reflectivity of road markings based on a polarization dual-camera system. This method can effectively solve the problem mentioned in the background art that current road marking detection instruments are expensive, bulky, and inconvenient to use widely, thus leading to the inability to detect and maintain road markings in a timely manner.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for measuring the reflectance of a criterion based on a polarization dual-camera system, comprising the following measurement steps: S101. Install the test component on the vehicle's trunk lid; S102. Determine the homography matrix of the two cameras; S103. Control the camera and artificial light source through hardware triggers to acquire target images with target road markings respectively; S104. Perform processing on the images captured by the camera, including homography transformation, to align the images. S105 performs image segmentation on the target image, distinguishing between road markings and the road background to obtain a better and more complete road marking map; S106. Calculate and process the obtained road marking map to obtain the corresponding reflectivity R.

[0005] According to the above technical solution, in step S102, the formula for calculating the homography matrix of the camera is as follows: ; in, This is the intrinsic parameter matrix of the camera; This is the actual extrinsic parameter matrix of the camera; The ideal extrinsic parameter matrix for the camera; and These are the matrix inversion and pseudo-inversion operations, respectively.

[0006] According to the above technical solution, in S106, the 0-255 values ​​represented by the RGB values ​​corresponding to the pixel values ​​are processed to measure the light intensity I of the artificial light source and the light intensity measured by the camera equipped with a polarizer aligned with the artificial light source. The light intensity measured by a camera equipped with a polarizer perpendicular to the direction of the artificial light source. To obtain the corresponding reflected light intensity without external interference. Then calculate the reflectivity R; Reflected light intensity The calculation formula is as follows: ; in, The light intensity measured by a camera equipped with a polarizer aligned with the direction of the artificial light source; The light intensity measured by a camera equipped with a polarizer perpendicular to the direction of the artificial light source; The formula for calculating reflectance R is as follows: ; Where I is the measured light intensity of the artificial light source; The intensity of the reflected light.

[0007] According to the above technical solution, in S101, the test components include an installation component for mounting on the trunk lid of a vehicle, a hardware synchronization trigger, a main control, an artificial light source, two cameras, three polarizers, and a computer; The mounting assembly includes a shaft that is perpendicular to the horizontal axis of the vehicle's trunk lid and parallel to the horizontal axis in the horizontal direction. Two cameras are spaced apart on the shaft, and the optical axes of the two cameras and the optical axis of the artificial light source are aligned with the axial direction of the shaft. The shaft is clamped between the two cameras.

[0008] According to the above technical solution, in S101-S106, the measurement of the reflectivity of the markings requires the cooperation of multiple modules. The modules involved in the reflectivity measurement include the test component module and the camera module installed on the vehicle trunk lid, as well as the hardware trigger module, the target image module, the image processing module, the image segmentation module, and the calculation module. The hardware trigger module connects the triggers correctly to the two cameras according to the hardware triggers in the two cameras and the main control, so as to achieve simultaneous shooting by different cameras and control the switching of the light source according to the trigger signal. The target image module uses hardware triggers to control two cameras and an artificial light source to acquire target images with target road markings. The image processing module performs image correction and other preprocessing operations using a homography matrix. The image segmentation module uses the PointRend image segmentation model to segment the road markings in the image from the background road; The calculation module calculates the reflectivity of road markings according to the calculation formula provided above, and obtains the reflectivity of the corresponding road markings.

[0009] According to the above technical solution, the PointRend image segmentation model mainly consists of the following four parts: I. It includes two stages of feature processing: fine-grained features and coarseprediction. II. Select N "difficult points" from coarse prediction. For each difficult point, obtain its "feature vector". The point feature vector mainly consists of two parts: the feature vector of the corresponding point of fine-grained features and the feature vector of the corresponding point of coarse prediction. Concatenate these two feature vectors into a single vector. III. Predict this "feature vector" using an MLP network, update the coarse prediction, which means making a new prediction and classification for this difficult point; IV. Continue repeating steps I-III to render the corresponding segmented image (mask) from coarse to fine and from point to surface.

[0010] According to the above technical solution, a computer device includes at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit performs the steps of the method for measuring the reflectivity of road markings.

[0011] According to the above technical solution, a computer-readable storage medium stores a computer program executed by a computer device, which, when the computer program is run on the computer device, causes the computer device to perform the steps of the method for measuring the reflectivity of road markings.

[0012] Compared with existing technologies, the beneficial effects of this invention are as follows: By installing components, a main control unit, an artificial light source, two cameras, three polarizers, and a computer on the trunk lid of a car, the optical axes of the two cameras and the artificial light source are aligned and do not interfere with each other. Based on the homography transformation processing of the two cameras, images of the target road markings are acquired separately. Image segmentation is then performed to separate the road markings, conveniently obtaining the road edge image. The reflectivity is calculated based on the pixel values ​​of the road marking image. By installing the testing device on a car, images can be captured while the car is in motion, making operation more convenient. Through image capture, image segmentation, and calculation of the road marking reflectivity, the effectiveness of road markings can be promptly detected, facilitating maintenance of road sections with poor reflectivity and ensuring the road markings achieve their ideal performance. Attached Figure Description

[0013] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0014] In the attached diagram: Figure 1 This is a flowchart of the steps for measuring the reflectivity of road edges according to the present invention; Figure 2 This is a schematic diagram of the installation of the test components provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the target image provided in an embodiment of the present invention; Figure 4 This is the image processed by the PointRend neural network; Figure 5 This describes how to use the PointRend neural network; Figure 6 This is a schematic diagram of a system module for measuring the reflectivity of road markings provided in an embodiment of the present invention. Detailed Implementation

[0015] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0016] Example: Figure 1-5 As shown, the present invention provides a technical solution, a method for measuring the reflectance of a marking based on a polarization dual-camera, comprising the following measurement steps: S101. Install the test component on the vehicle's trunk lid; S102. Determine the homography matrix of the two cameras; S103. Control the camera and artificial light source through hardware triggers to acquire target images with target road markings respectively; S104. Perform processing on the images captured by the camera, including homography transformation, to align the images. S105 performs image segmentation on the target image, distinguishing between road markings and the road background to obtain a better and more complete road marking map; S106. Calculate and process the obtained road marking map to obtain the corresponding reflectivity R.

[0017] In S102, the formula for calculating the homography matrix of the camera is as follows: ; in, This is the intrinsic parameter matrix of the camera; This is the actual extrinsic parameter matrix of the camera; The ideal extrinsic parameter matrix for the camera; and These are the matrix inversion and pseudo-inversion operations, respectively.

[0018] According to the above technical solution, in S106, the 0-255 values ​​represented by the RGB values ​​corresponding to the pixel values ​​are processed to measure the light intensity I of the artificial light source and the light intensity measured by the camera equipped with a polarizer aligned with the artificial light source. The light intensity measured by a camera equipped with a polarizer perpendicular to the direction of the artificial light source. To obtain the corresponding reflected light intensity without external interference. Then calculate the reflectivity R; Reflected light intensity The calculation formula is as follows: ; in, The light intensity measured by a camera equipped with a polarizer aligned with the direction of the artificial light source; The light intensity measured by a camera equipped with a polarizer perpendicular to the direction of the artificial light source; The formula for calculating reflectance R is as follows: ; Where I is the measured light intensity of the artificial light source; The intensity of the reflected light.

[0019] In S101, the test components include a mounting assembly for mounting on the vehicle trunk lid, a hardware synchronization trigger, a master control, an artificial light source, two cameras, three polarizers, and a computer. The mounting assembly includes a shaft that is perpendicular to the horizontal axis of the vehicle's trunk lid and parallel to the horizontal axis in the horizontal direction. Two cameras are spaced apart on the shaft to capture road markings. The optical axes of the two cameras and the optical axis of the artificial light source are aligned with the axial direction of the shaft, and the shaft is clamped between the two cameras.

[0020] like Figure 6 As shown, the modules involved in reflectivity measurement include a test component module and a camera module installed on the vehicle trunk lid, as well as a hardware trigger module, a target image module, an image processing module, an image segmentation module, and a calculation module. The hardware trigger module connects the triggers correctly to the two cameras and the main control based on the trigger signals, enabling simultaneous shooting by different cameras and controlling the switching of the light source. The target image module uses hardware triggers to control two cameras and an artificial light source to acquire target images containing the target road markings. The image processing module performs image correction and other preprocessing operations using the homography matrix. The image segmentation module uses the PointRend image segmentation model to segment the road markings in the image from the background road, making it easier for subsequent processing. The calculation module calculates the reflectivity of road markings according to the calculation formula provided above, and obtains the corresponding reflectivity of road markings.

[0021] The PointRend image segmentation model mainly consists of the following four parts: I. As seen in the application of PointRend, it involves two stages of feature processing: fine-grained features and coarse prediction. The backbone network is ResNet, so the fine-grained features are the stage 2 output of ResNet, which is the fine segmentation result after 4x downsampling. The coarse prediction is the prediction result of the detector head, i.e., the result before upsampling and restoration to the original image. Specific code: points = sampling_points(out, x.shape[-1] / / 16, self.k, self.beta); II. Select N "difficult points" from the coarse prediction, which are points whose results differ significantly from surrounding points, such as points on the object's edge. For each difficult point, obtain its "feature vector." The point feature vector consists of two parts: the feature vector of the corresponding point in the fine-grained features and the feature vector of the corresponding point in the coarse prediction. Concatenate these two feature vectors into a single vector. Specific code follows: coarse = point_sample(out, points, align_corners=False) fine = point_sample(reS102, points, align_corners=False) feature_representation = torch.cat([coarse, fine], dim=1) III. Next, an MLP network is used to predict this "feature vector" and update coarseprediction, which is equivalent to making a new prediction for this difficult point and classifying it. Specific code: rend = self.mlp(feature_representation) IV. Continue repeating steps I-III, rendering the corresponding mask from coarse to fine and from point to surface, which is the segmented image.

[0022] A computer device includes a processing unit and a storage unit, wherein the storage unit stores a computer program that, when executed by the processing unit, causes the processing unit to perform steps of a method for measuring the reflectivity of road markings.

[0023] A computer-readable storage medium storing a computer program executable by a computer device, which, when run on the computer device, causes the computer device to perform steps of a method for measuring the reflectivity of road markings.

[0024] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for measuring the reflectance of a gradation line based on a polarization dual-camera system, characterized in that: The measurement steps include the following: S101. Install the test component on the vehicle's trunk lid; S102. Determine the homography matrix of the two cameras; S103. The camera and artificial light source are controlled by a hardware trigger to acquire target images with target road markings respectively; S104. Perform processing on the images captured by the camera, including homography transformation, to align the images. S105 performs image segmentation on the target image, distinguishing between road markings and the road background to obtain a better and more complete road marking map; S106. Calculate and process the obtained road marking map to obtain the corresponding reflectivity R; In step S106, the 0-255 values ​​represented by the RGB values ​​corresponding to the pixel values ​​are processed to measure the light intensity I of the artificial light source and the light intensity measured by the camera equipped with a polarizer aligned with the artificial light source. The light intensity measured by a camera equipped with a polarizer perpendicular to the direction of the artificial light source. To obtain the corresponding reflected light intensity without external interference. Then calculate the reflectivity R; Reflected light intensity The calculation formula is as follows: ； in, The light intensity measured by a camera equipped with a polarizer aligned with the direction of the artificial light source; The light intensity measured by a camera equipped with a polarizer perpendicular to the direction of the artificial light source; The formula for calculating reflectance R is as follows: ； Where I is the measured light intensity of the artificial light source; The intensity of the reflected light.

2. The method for measuring the reflectance of a gradation line based on a polarization dual-camera according to claim 1, characterized in that, In step S102, the formula for calculating the homography matrix of the camera is as follows: ； in, This is the intrinsic parameter matrix of the camera; This is the actual extrinsic parameter matrix of the camera; The ideal extrinsic parameter matrix for the camera; and These are the matrix inversion and pseudo-inversion operations, respectively.

3. The method for measuring the reflectance of a gradation line based on a polarization dual-camera according to claim 1, characterized in that, In S101, the test components include a mounting assembly for mounting on the trunk lid of a vehicle, a hardware synchronization trigger, a main control, an artificial light source, two cameras, three polarizers, and a computer. The mounting assembly includes a shaft that is perpendicular to the horizontal axis of the vehicle's trunk lid and parallel to the horizontal axis in the horizontal direction. Two cameras are spaced apart on the shaft, and the optical axes of the two cameras and the optical axis of the artificial light source are aligned with the axial direction of the shaft. The shaft is clamped between the two cameras.

4. The method for measuring the reflectance of a gradation line based on a polarization dual-camera according to claim 1, characterized in that, In S101-S106, the measurement of the reflectivity of the markings requires the cooperation of multiple modules. The modules involved in the reflectivity measurement include the test component module and camera module installed on the vehicle trunk lid, as well as the hardware trigger module, target image module, image processing module, image segmentation module and calculation module. The hardware trigger module connects the triggers correctly to the two cameras according to the hardware triggers in the two cameras and the main control, so as to achieve simultaneous shooting by different cameras and control the switching of the light source according to the trigger signal. The target image module uses hardware triggers to control two cameras and an artificial light source to acquire target images with target road markings. The image processing module performs image correction and other preprocessing operations using a homography matrix. The image segmentation module uses the PointRend image segmentation model to segment the road markings in the image from the background road; The calculation module calculates the reflectivity of road markings according to the calculation formula provided above, and obtains the reflectivity of the corresponding road markings.

5. The method for measuring the reflectance of a gradation line based on a polarization dual-camera according to claim 4, characterized in that, The PointRend image segmentation model mainly consists of the following four parts: I. It includes two stages of feature processing: fine-grained features and coarse prediction. II. Select N "difficult points" from coarse prediction. For each difficult point, obtain its "feature vector". The point feature vector mainly consists of two parts: the feature vector of the corresponding point of fine-grained features and the feature vector of the corresponding point of coarse prediction. Concatenate these two feature vectors into a single vector. III. Predict this "feature vector" using an MLP network, update the coarse prediction, which means making a new prediction and classification for this difficult point; IV. Continue repeating steps I-III to render the corresponding segmented image (mask) from coarse to fine and from point to surface.

Images