Method for reducing atmospheric turbulence effects in laser measurement of railway rail straightness

Abstract

The application discloses a method for reducing the atmospheric turbulence effect in railway rail flatness laser measurement, comprising the following steps: S1, collecting a light spot image and extracting a part with high light spot energy; S2, processing the original light spot image by using a log algorithm, then performing normalization processing on the image, and extracting the part with high light spot energy by using a threshold method to obtain a light spot image; S3, taking the contours of the light spot image in four directions, fitting in the four directions respectively, obtaining four fitted circles, finding two circles with the smallest radius difference, and re-fitting the contours of the two circles into a circle, and the center of the circle is the position of the light spot; S4, repeatedly performing S1-S3 for multiple times to obtain the average value of multiple light spots. Through the fixed threshold method, the log algorithm, the fitting in four directions and other operations, the positioning error caused by the irregular light spot formed by the laser light spot affected by the atmospheric turbulence can be reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to a method for reducing the effect of atmospheric turbulence in laser measurement of railway rail flatness. BACKGROUND

[0002] Spot center positioning is a precise positioning technology of the energy center position of a laser spot, and is a key technology commonly used in optical measurement, which has a wide range of applications in the fields of laser communication, target tracking, weapon systems, laser radar and laser ranging. In atmospheric laser communication, after the laser passes through the turbulent atmosphere, phenomena such as beam drift, expansion and intensity fluctuation occur, resulting in different degrees of distorted spots detected by the receiving antenna. Fast and accurate positioning of the center of the imaged spot is one of the key technologies to realize laser communication.

[0003] At present, the spot positioning methods used in the system include centroid method, geometric center method, Hough transform method, ellipse fitting method and Gaussian fitting method. The centroid method and the geometric center method can accurately position the uniform spot, have fast calculation speed, and have high positioning accuracy for symmetrical and uniformly distributed spots. When the spot distortion degree becomes larger, the accuracy will also be reduced to a certain extent. Zhou Hongfeng et al. used an iterative threshold method for rough positioning, and then performed interpolation processing to achieve sub-pixel positioning of the CCD spot center, which improved the positioning accuracy and stability to a certain extent. Zhang Feizhou et al. statistically analyzed the characteristics of the broken spot. With the increase of turbulence intensity and the deterioration of beam quality, the total spot expands and the number of broken spots increases, while the radius of the broken spot changes little. Shen Baoliang et al. gave a statistical deviation curve of the influence of turbulence intensity on spot centroid positioning accuracy through simulation. The results show that improving the reliability of the central part of the CCD camera can improve the overall positioning accuracy. Liu Yunqing, Jiang Huilin et al. conducted in-depth analysis on the centroid algorithm and the geometric center tracking method in combination with atmospheric turbulence and detectors, and obtained and experimentally verified the conclusion that the tracking error of the geometric center algorithm is less than that of the centroid error under atmospheric conditions. However, the stability and accuracy of the geometric center method under turbulence conditions can still be improved. Wu Zekai et al. improved the algorithm based on circle fitting, added acquisition processing, denoising and filtering, and multiple iterations, which improved the accuracy of detecting the spot center position and enhanced the anti-interference ability. In 2019, WANG X et al. proposed a power-weighted sub-pixel positioning algorithm, which was applied to a tiltmeter CCD measurement system and experimentally verified. The experiment showed that the algorithm had high positioning accuracy for regular spots in an indoor non-turbulent environment. The positioning stability in a turbulent environment needs to be further improved. The current adaptive optics technology can detect and correct the beam wavefront phase distortion caused by atmospheric turbulence in real time, but it is still limited by the computing power, calculation speed and measurement error, which has a great influence. SUMMARY

[0004] The present application aims at overcoming the deficiencies of the prior art, and provides a method for reducing the atmospheric turbulence effect in railway rail flatness laser measurement, which can reduce the positioning error caused by the irregular laser spot formed by the atmospheric turbulence affecting the laser spot through fixed threshold method, log algorithm, fitting in four directions and other operations.

[0005] The purpose of the present application is achieved by the following technical scheme: a method for reducing the atmospheric turbulence effect in railway rail flatness laser measurement, comprising the following steps:

[0006] S1, collect the spot image and extract the part with high spot energy, set the area of the extracted part with high spot energy as A, and set the area of the part with high spot energy extracted without the influence of atmospheric turbulence as B; when the area A is greater than or equal to B, the traditional centroid method or barycenter method is used to position the spot, and when A is less than B, step S2 is performed on the original spot image;

[0007] S2, the original spot image is processed by using the log algorithm, then the image is normalized, and then the part with high spot energy is extracted by using the threshold method to obtain the spot image;

[0008] S3, the contours of the spot image in the upper, lower, left and right directions are taken, fitting is performed in the four directions respectively, four fitted circles are obtained, the two circles with the smallest radius difference are found, and the contours of the two circles are refitted as a circle, and the center of the circle is the position of the spot;

[0009] S4, repeat S1-S3 multiple times to obtain the average value of multiple spots.

[0010] The specific implementation method of step S1 is to extract the spot image by using a fixed threshold algorithm: set the fixed threshold as T, g(x, y) as the gray value of the image in x row y column, and f(x, y) as the original pixel value of the image in x row y column:

[0011]

[0012] After binarizing the image, the number of threshold values of 0 is added to obtain the area A; a spot image formed at a distance of 5m between the laser emitting end and the receiving target surface is collected, the fixed threshold T is set, and the number of threshold values of 0 after binarization is added to obtain the area B.

[0013] The specific implementation method of step S2 is:

[0014] S21, calculate the natural logarithm of the absolute value of each pixel point src(I) on the image by using the log() function:

[0015]

[0016] C is a predetermined constant.

[0017] S22, the natural logarithm is normalized and converted to the interval of 0-255;

[0018] S23, the part with high spot energy is extracted by fixed threshold algorithm, and the spot image is obtained.

[0019] The beneficial effects of the present application are: the present application can reduce the positioning error caused by irregular spot formed by the influence of atmospheric turbulence on laser spot through fixed threshold method, log algorithm, fitting in four directions and other operations. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 the flow chart of the method for reducing the effect of atmospheric turbulence in laser measurement of railway rail flatness of the present application;

[0021] Figure 2 the spot image collected by the present application.

[0022] Figure 3 the picture after the collected image is binarized by fixed threshold value;

[0023] Figure 4 the picture after the original image is normalized by log image processing of the present application;

[0024] Figure 5 the picture after threshold segmentation;

[0025] Figure 6 the center of the circle found by the present application. DETAILED DESCRIPTION

[0026] The technical solutions of the present application will be further described below in combination with the drawings.

[0027] As shown in the drawings, the method for reducing the effect of atmospheric turbulence in laser measurement of railway rail flatness of the present application comprises the following steps: Figure 1 S1, collect the spot image and extract the part with high spot energy, set the area of the extracted part with high spot energy as A, and set the area of the part with high spot energy extracted without the influence of atmospheric turbulence as B; when the area A is greater than or equal to B, it indicates that the atmospheric turbulence is not disturbed, and the traditional centroid method or barycenter method is used to position the spot; when A is less than B, it indicates that the spot is affected by atmospheric turbulence, causing excessive energy loss, small high-energy area and instability, which needs to be further processed, and step S2 is performed on the original spot image;

[0028]

[0029] ​The specific implementation method of this step is: a fixed threshold algorithm is used to extract the light spot image: a fixed threshold T is set, g(x, y) is the gray value of the image at x row y column, and f(x, y) represents the original pixel value of the image at x row y column:

[0030]

[0031] The light spot image collected in this embodiment is shown in Fig. 1, and the high-energy area of the light spot is extracted by the fixed threshold method, as shown in Fig. 2. After binarization of the image, the number of threshold values of 0 is added, and the area A is obtained. A light spot image is collected at a distance of 5 m between the laser emitting end and the receiving target surface, and the fixed threshold T is set. After binarization, the number of threshold values of 0 is added, and the area B is obtained. Figure 2 Figure 3

[0032] S2, the original light spot image is processed by using the log algorithm, then the image is normalized, and then the high-energy part of the light spot is extracted by the threshold method to obtain the light spot image; the specific implementation method is:

[0033] S21, the absolute value of each pixel point src(I) on the image is calculated by using the log() function, so that the data is more stable:

[0034]

[0035] C is a predetermined constant;

[0036] S22, the natural logarithm is normalized and converted to the interval of 0-255, which can enhance the image and increase the contrast of the image, so that the area affected by atmospheric turbulence is further strengthened, which is closer to the current situation of the original light spot, and the edge of the light plate is clearer. A new image is generated as shown in Fig. 3. Figure 4

[0037] S23, the high-energy part of the light spot is extracted by the fixed threshold algorithm to obtain the light spot image, as shown in Fig. 4. Figure 5

[0038] S3, because of the position of the light spot, the influence of atmospheric turbulence or other factors, the image processed according to S3 cannot completely obtain an accurate light spot image, and fitting a circle according to the boundary of the current image is also inaccurate. Therefore, the contour of the light spot image is taken in four directions, fitting is performed in four directions respectively, four fitted circles are obtained by using the least square method, two circles with the smallest radius difference are found, the circle with large error change is excluded, and the contours of the two circles are refitted as a circle. The center of the circle is the position of the light spot, as shown in Fig. 5. Figure 6

[0039] The least square method is used to fit a circle: the equation of the circle is: ​​​​​

[0040] (x-x c ) 2 +(y-y c ) 2 = R 2 (3)

[0041] (x c ,y c ) is the center of the fitted circle, (x, y) is the scatter data of the circle, and R is the radius of the fitted circle.

[0042] The square of the sum of the distances from the existing contour points to the fitted circle is:

[0043]

[0044] When f is the minimum, the circle is found.

[0045] S4, repeatedly S1-S3 multiple times to obtain the average value of multiple light spots. Due to the influence of atmospheric turbulence under outdoor environment, the laser will appear large fluctuations. The position of the light spot is not fixed, but always fluctuates in a range. By recording the position of the light spot multiple times, the average value is finally taken, so that the position of the light spot is closer to the value under the condition that the light spot is not disturbed.

[0046] Those skilled in the art will appreciate that the embodiments described herein are presented for the purpose of helping the reader understand the principles of the present application, and should be understood as not limiting the scope of protection of the present application to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations according to the technical inspiration disclosed in the present application without departing from the essence of the present application, and these modifications and combinations are still within the scope of protection of the present application.

Claims

1. A method for reducing the effect of atmospheric turbulence in laser measurement of the flatness of a railway rail, characterized in that, Comprise the following steps: S1, collect the light spot image and extract the light spot energy high part, set the extracted light spot energy high part area A, set the area of the extracted light spot energy high part not affected by atmospheric turbulence B; When area A is greater than or equal to B, the traditional centroid method or centroid method is used to locate the light spot, when A is less than B, step S2 is executed on the original light spot image; S2, the original light spot image is processed by log algorithm, then the image is normalized, and the light spot energy high part is extracted by threshold method to obtain the light spot image; The specific implementation method is: S21. Use the log() function to calculate the value of each pixel in the image. The natural logarithm of the absolute value: （2）； is a predetermined constant; S22, the natural logarithm is normalized and converted to the interval of 0~255; S23, the light spot energy high part is extracted by fixed threshold algorithm to obtain the light spot image; S3, the contours of the light spot image in the upper, lower, left and right directions are taken, fitting is performed in the four directions respectively, four fitted circles are obtained, the two circles with the smallest radius difference are found, and the contours of the two circles are refitted as a circle, and the center of the circle is the position of the light spot; S4, repeat S1~S3 multiple times to obtain the average value of multiple light spots.

2. The method for reducing the effect of atmospheric turbulence in laser measurement of the flatness of railway rails according to claim 1, characterized in that, The step S1 is specifically implemented by using a fixed threshold algorithm to extract the light spot image, wherein a fixed threshold is T, represents the gray value of the image in the xth row and yth column, represents the original pixel value of the image in the xth row and yth column. （1）； After the image is binarized, the number of threshold value 0 is added, which is area A; Collect a light spot image formed at a distance of 5m between the laser emitting end and the receiving target surface, fix the threshold value T, add the number of threshold value 0 after binarization, which is area B.

Images