Mileage control road surface image rapid splicing method

Abstract

The application discloses a kind of mileage control road surface image fast splicing methods.High-definition camera is installed at the top of car tail, camera positive direction keeps certain angle with ground, encoder is installed on wheel, and mileage information is obtained by the rotation of wheel;According to mileage information, control camera to collect road surface image;The collected road surface image is perspective transformation, and field of view is adjusted to road surface directly above, and 'ladder' road surface is transformed into'rectangle' road surface;Road surface image after perspective transformation is spliced according to corresponding mileage;Optimize the boundary of spliced image, so that image splicing boundary is more natural;Finally, complete and clear road surface image is output.The application can achieve the requirement of real-time collection and real-time imaging in efficiency, is easy to realize, and cost is relatively low;It can be used for highway, urban road and various road surface splicing imaging;Road surface condition can be more intuitively and clearly observed, and informationization, visual management of road surface maintenance is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of road surface detection, and is a kind of mileage control road surface image rapid splicing method, in particular to the splicing of road surface image by using mileage control information, which is a low-cost and effective method, and is widely used in various road surface disease detection, road surface flatness detection and other fields of expressway, urban road, etc. BACKGROUND

[0002] As an important infrastructure supporting economic, political and cultural development, urban road is not only an important part of modern transportation system, but also closely related to the improvement of people's living standards. According to the "2020 Traffic and Transportation Industry Development Statistics Bulletin", by the end of 2020, the total mileage of highways in China was 519.81 million kilometers, of which the mileage of expressways was 161,000 kilometers.

[0003] With the increase of highway mileage, and the influence of natural factors (light temperature, temperature sudden change, humidity content, etc.) and increasing traffic load, a large number of roads in use have cracks, pits, pumping and subsidence and other diseases. These road diseases not only seriously affect the bearing capacity, transportation capacity and driving safety of the road, but also lead to increasingly serious road maintenance and management problems. Therefore, how to quickly and clearly observe the road surface conditions in a short time has become a problem to be solved.

[0004] The previous manual inspection method is low in efficiency, high in risk coefficient, and consumes manpower and material resources. This method affects traffic during detection. With the expansion of road network, the manual inspection method cannot meet the demand. Intelligent and efficient detection equipment can overcome the defects of manual inspection, making it possible to detect road diseases and maintain and manage roads on a large scale and with high timeliness. Digital images can more intuitively and clearly observe the road surface conditions, and data can be collected without affecting traffic, realizing the informatization, comprehensiveness and visualization management of road maintenance.

[0005] Mileage control road surface image splicing mainly refers to superimposing image sequences according to mileage information to splice a complete road surface image. Since the imaging unit is at a certain angle with the road surface, according to the perspective principle, the road surface part in the collected image presents a "trapezoidal" shape. The field of view needs to be adjusted to the top of the road surface, so as to convert the "trapezoidal" road surface into a "rectangular" road surface. Then, according to the mileage information corresponding to each picture, the splicing boundary is optimized, and finally a complete road surface image is formed. SUMMARY

[0006] The present application aims to overcome the above-mentioned defects of manual inspection and provide a kind of mileage control road surface image rapid splicing method.

[0007] The technical solution of the present application is a kind of mileage control road surface image fast splicing method, the steps are as follows:

[0008] Step (1): the road surface image is obtained by a high-definition camera located at the top of the car tail, the camera is at a certain angle with the ground, an encoder is installed on the wheel, and the mileage information is obtained by the rotation of the wheel, and the encoder parameters have the following relationship:

[0009]

[0010] Wherein, l is the circumference of the encoder, d is the pulse interval, q is the number of single pulse, and p is the interval pulse number, when the circumference of the encoder is 0.2 meters, the number of single pulse is 500, and the interval pulse number is 25, the pulse interval is 0.01 meters, which means that the pulse signal is sent every 0.01 meters; Step (2): according to the mileage information, the camera collects the road surface image, and the initial mileage and the pulse interval have the following relationship:

[0011] s=d*c

[0012] Wherein, s is the initial mileage, c is the cumulative pulse number in the driving process, the camera collects the ith road surface image I i , the cumulative pulse number of the ith road surface image is c i , and the initial mileage position s i of the ith road surface image is d*c i , wherein the road surface image number i=1, 2,..., N, and N is the total number of road surface images;

[0013] Step (3): the perspective transformation is carried out on the road surface image I i to obtain I' i , and the calculation formula of the point (x', y') on the two-dimensional image after transformation is as follows:

[0014]

[0015] Wherein, (x, y) is the coordinate point of the original graph, a 11 , a 12 , a 13 , a 21 , a 22 , a 23 , a 31 , a 32 are known parameters;

[0016] Step (4): the road surface image I' i after perspective transformation is according to the corresponding mileage s iThe stitching is performed, v is an imaging range of the road surface image, when a corresponding mileage interval of adjacent road surface images is less than or equal to v, an overlapping range w of adjacent images is i =v-(s i+1 -s i ) and w i ≥0, when the corresponding mileage interval of adjacent road surface images is greater than v, the overlapping range w of adjacent images is i <0, that is, the adjacent images do not overlap, wherein i=1, 2,..., N is the number of road surface images;

[0017] Step (5): the adjacent image overlapping parts are weighted and superimposed, so that the image stitching boundary is more natural, and the weighting formula is as follows:

[0018] I″ i =α*I′ i +β*I′ i+1 ,i=1,2,...,N-1

[0019] wherein I′ i , I′ i+1 are adjacent images, I″ i is an adjacent image overlapping part, and α and β are proportion weights of I′ i and I′ i+1 , respectively, and α and β satisfy the formula:

[0020]

[0021] wherein ξ is an adjustable parameter and 0<ξ<1, k is a mileage coordinate of the overlapping part I″ i , and The significant advantages of the present application are: 1. high efficiency, which can avoid the problems of time-consuming, influence on traffic operation and high risk of manual inspection and quickly present the complete road surface condition; 2. easy to implement, according to the mileage information to splice the road surface image, without additional other information, convenient and fast; 3. low cost, compared with other road surface detection vehicles, the installation and collection implementation process is simple, and the cost is low; 4. high robustness, which can be widely applied to various road surface images such as expressways and urban roads. BRIEF DESCRIPTION OF DRAWINGS

[0022] Figure 1 is a flowchart of the present application;

[0023] Figure 2 is a rear view of a high-definition camera road surface collection vehicle;

[0024] Figure 3 is a perspective transformation effect drawing, (a) before perspective transformation, (b) after perspective transformation;

[0025] Figure 4For the adjacent two road surface image splicing effect picture, (a) before splicing, (b) after splicing;

[0026] Figure 5 For the adjacent two road surface image splicing boundary optimization effect picture, (a) before optimization, (b) after optimization.

CONCRETE EMBODIMENT

[0027] The application will be described in further detail below with reference to the drawings.

[0028] In combination Figure 1 , the application is a kind of mileage control road surface image fast splicing method, mainly including in the collection car tail top installation high-definition camera, collection road surface image, processing image, splicing, optimization boundary process. First, install high-definition camera on the top of the car tail, the camera is kept at a certain angle with the ground, and an encoder is installed on the wheel to obtain mileage information through the rotation of the wheel. Secondly, according to the mileage information, the camera collects the road surface image. Then, the collected road surface image is processed by perspective transformation, and the field of view is adjusted to the top of the road surface, so that the "trapezoidal" road surface is converted into a "rectangular" road surface. Then, the road surface image after perspective transformation is spliced according to the corresponding mileage. Finally, the image splicing boundary is optimized to obtain a clear and intuitive complete road surface image. The specific steps are as follows:

[0029] First, install a high-definition camera on the top of the detection car tail

[0030] The road surface image is obtained by the high-definition camera located at the top of the car tail, and the camera is kept at a certain angle with the ground. An encoder is installed on the wheel to obtain mileage information through the rotation of the wheel. The encoder parameters have the following relationship:

[0031]

[0032] Wherein, l is the circumference of the encoder, d is the pulse interval, q is the number of pulses per revolution, and p is the number of interval pulses. When the circumference of the encoder is 0.2 meters, the number of pulses per revolution is 500, and the number of interval pulses is 25, the pulse interval is 0.01 meters, which means that a pulse signal is sent every 0.01 meters. The high-definition camera rear view image is shown in Figure 2 .

[0033] Second, according to the mileage information, control the camera to collect the road surface image, the initial mileage and the pulse interval have the following relationship:

[0034] s=d*c

[0035] Wherein, s is the initial mileage, c is the cumulative pulse number in the driving process, and the camera collects the i-th road surface image I i , i=1, 2,..., N, the corresponding cumulative pulse number is ci i = 1, 2,..., N, starting mileage position s i = d * c i i = 1, 2,..., N, N is the total number of road surface images.

[0036] Third, the road surface image I i is subjected to perspective transformation to obtain I' i The calculation formula of the point (x', y') on the transformed two-dimensional image is as follows:

[0037]

[0038] Where (x, y) is the coordinate point of the original image, a 11 , a 12 , a 13 , a 21 , a 22 , a 23 , a 31 , a 32 are known parameters, a 33 = 1, and the effect diagrams before and after the perspective transformation are shown in Figure 3 .

[0039] Fourth, the road surface image I' i i = 1, 2,..., N after perspective transformation is spliced according to the corresponding mileage s i i = 1, 2,..., N, and v is the imaging range of the road surface image. When the distance between the corresponding mileages of adjacent road surface images is less than or equal to v, the overlapping range w i = v - (s i+1 - s i ) and w i ≥ 0. When the distance between the corresponding mileages of adjacent road surface images is greater than v, the overlapping range w i < 0, i.e. the adjacent images do not overlap, and the splicing effect diagram of the adjacent two road surface images is shown in Figure 4 .

[0040] Fifth, the overlapping part of the adjacent images is subjected to weighted superposition to make the image splicing boundary more natural, and the weighted formula is as follows:

[0041] I" i = a * I' i + b * I' i+1 i = 1, 2,..., N - 1

[0042] Where I' i , I' i+1 are adjacent images, I" i is the overlapping part of the adjacent images, and a, b are I' i , I'i+1 The proportion weight of the two is α, β, which satisfies the formula:

[0043]

[0044] Wherein, ξ is an adjustable parameter and 0 < ξ < 1, k is the mileage coordinate of the overlapping part I" i The road surface image is stitched. The effect diagram of the stitching boundary of the adjacent two road surface images before and after the optimization is shown in Figure 5 .

[0045] The specific embodiments described in the specification can have different names, and the above description in the specification is only an example of the structure of the present application. Any changes or simple changes made according to the structure, features and principles of the present application are included in the protection scope of the present application. Those skilled in the art can make various modifications or supplements to the described specific examples or use similar methods, as long as they do not deviate from the structure of the present application or exceed the scope defined by the present claims, which shall belong to the protection scope of the present application.

Claims

1. A method for rapid image stitching of mileage-controlled road surfaces, characterized in that: Step (1): The road surface image is obtained by a high-definition camera located at the top of the rear of the vehicle. The camera is positioned at a certain angle to the ground. Encoders are installed on the wheels to obtain mileage information through the rotation of the wheels. The encoder parameters are related as follows: Where l is the encoder circumference, d is the pulse interval, q is the number of pulses per revolution, and p is the number of pulses per interval. When the encoder circumference is 0.2 meters, the number of pulses per revolution is 500, and the number of pulses per interval is 25, the pulse interval is 0.01 meters, which means that a pulse signal is sent every 0.01 meters. Step (2): Based on the mileage information, control the camera to collect road surface images. The initial mileage and pulse interval have the following relationship: s=d*c Where s is the initial mileage, c is the cumulative number of pulses during the journey, and the camera captures the i-th road image I. i At that time, the cumulative pulse count of the i-th road surface image is c. i The starting mileage position s of the i-th road surface image i =d*c i Where the road surface image number is i = 1, 2, ..., N, and N is the total number of road surface images; Step (3): Process the road surface image I i Perform perspective transformation to obtain I′ i The formula for calculating the point (x′, y′) on the transformed 2D image is as follows: Where (x,y) are the coordinates of the original graph, and a 11 a 12 a 13 a 21 a 22 a 23 a 31 a 32 The parameters are known. Step (4): Perspective transformed road surface image I′ i According to the corresponding mileage s i The images are stitched together, where v represents the imaging range of the road surface images. When the mileage distance between adjacent road surface images is less than or equal to v, the overlap range of the adjacent images is w. i =v-(s i+1 -s i And w i ≥0, when the mileage interval between adjacent road surface images is greater than v, the overlap range of adjacent images is w. i <0, meaning adjacent images do not overlap, where i = 1, 2, ..., N are the road surface image numbers; Step (5): Weighted superposition of overlapping parts of adjacent images to make the image stitching boundary more natural. The weighting formula is as follows: I″ i =α*I′ i +β*I′ i+1 ,i=1,2,...,N-1 Among them, I′ i 、I′ i+1 For adjacent images, I″ i The overlapping portion of adjacent images is represented by α and β, respectively, where I′ i 、I′ i+1 The weights, α and β, satisfy the formula: Where ξ is an adjustable parameter and 0 < ξ < 1, and k is the overlapping part I″. i Mileage coordinates and

Images