CN110969103B - A method for measuring the length of highway pavement disease based on PTZ camera - Google Patents

A method for measuring the length of highway pavement disease based on PTZ camera Download PDF

Info

Publication number
CN110969103B
CN110969103B CN201911148337.6A CN201911148337A CN110969103B CN 110969103 B CN110969103 B CN 110969103B CN 201911148337 A CN201911148337 A CN 201911148337A CN 110969103 B CN110969103 B CN 110969103B
Authority
CN
China
Prior art keywords
image
road surface
pixel
length
value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201911148337.6A
Other languages
Chinese (zh)
Other versions
CN110969103A (en
Inventor
杜豫川
邵春艳
刘成龙
潘宁
曹静
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tongji University
Original Assignee
Tongji University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tongji University filed Critical Tongji University
Priority to CN201911148337.6A priority Critical patent/CN110969103B/en
Publication of CN110969103A publication Critical patent/CN110969103A/en
Application granted granted Critical
Publication of CN110969103B publication Critical patent/CN110969103B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
    • G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Length Measuring Devices By Optical Means (AREA)
  • Traffic Control Systems (AREA)
  • Image Processing (AREA)
  • Image Analysis (AREA)

Abstract

本发明涉及一种基于PTZ相机的高速公路路面病害长度测量方法，具体包括以下步骤：步骤S1：获取不同视距下包含车道线的彩色的高速公路的路面图像；步骤S2：对获取的路面图像进行Harris角点特征提取，并对完成角点提取的路面图像进行直线段提取，获得对应的直线段图像；步骤S3：根据高速公路路面的直线段图像，计算不同视距下路面图像对应的像素精度；步骤S4：根据步骤S3中求出的像素精度，计算相应高速公路的路面图像中病害区域的图像病害长度，根据所述图像病害长度计算得到病害区域的物理长度，实现高速公路路面的病害长度测量。与现有技术相比，本发明具有高效测量路面病害长度、降低高速公路路面养护成本、提高路面病害维修时效性等优点。

Figure 201911148337

The present invention relates to a PTZ camera-based method for measuring the length of highway road surface diseases, which specifically includes the following steps: step S1: acquiring color highway road surface images including lane lines under different sight distances; step S2: comparing the acquired road surface images Perform Harris corner feature extraction, and extract straight line segments from the road surface image that has completed corner extraction to obtain a corresponding straight line segment image; Step S3: Calculate the pixels corresponding to the road surface image at different sight distances according to the straight line segment images of the highway pavement Accuracy; Step S4: According to the pixel accuracy obtained in step S3, calculate the image disease length of the diseased area in the road surface image of the corresponding expressway, and calculate the physical length of the diseased area according to the image disease length, so as to realize the disease of the expressway road surface. Length measurement. Compared with the prior art, the invention has the advantages of efficiently measuring the length of the road surface disease, reducing the maintenance cost of the expressway road surface, improving the timeliness of the road surface disease maintenance and the like.

Figure 201911148337

Description

一种基于PTZ相机的高速公路路面病害长度测量方法A method for measuring the length of highway pavement diseases based on PTZ camera

技术领域technical field

本发明涉及路面病害检测领域，尤其是涉及一种基于PTZ相机的高速公路路面病害长度测量方法。The invention relates to the field of road surface disease detection, in particular to a method for measuring the length of road surface disease on expressways based on a PTZ camera.

背景技术Background technique

我国国民经济的不断发展提升了交通运输在国民经济和社会中的显著地位，而作为交通运输主动脉的高速公路，更得到了快速的发展。随着高速公路修建量的增加，高速公路养护成为交通管理部门日渐关注的重点。路面病害检测可以很好地保障高速公路养护的实效性。目前，高速公路路面病害的严重程度主要由病害区域的长度评估。自动高效的路面病害检测技术在高速公路养护系统中起到重要作用，能够辅助高速公路养护人员实施正确的养护措施。而当小尺寸的检测病害得到及时修护，高速公路才能持续保持最佳的服务性能，同时提高路面的使用寿命，减少高速公路路面维护成本。因此，如何自动高效地对高速公路路面病害进行长度测量，成为目前路面病害检测领域具有重要现实意义的亟待解决问题之一。The continuous development of my country's national economy has enhanced the prominent position of transportation in the national economy and society, and the expressway, which is the main artery of transportation, has developed rapidly. With the increase of highway construction, highway maintenance has become the focus of the traffic management department. Pavement disease detection can well guarantee the effectiveness of highway maintenance. At present, the severity of highway pavement diseases is mainly assessed by the length of the diseased area. Automatic and efficient pavement disease detection technology plays an important role in the highway maintenance system, which can assist highway maintenance personnel to implement correct maintenance measures. When the small-sized detected diseases are repaired in time, the highway can continue to maintain the best service performance, while improving the service life of the pavement and reducing the maintenance cost of the highway pavement. Therefore, how to automatically and efficiently measure the length of highway pavement diseases has become one of the urgent problems of great practical significance in the field of pavement disease detection.

过去的30年间，已经提出不少高速公路路面病害检测方法，促进了路面病害长度测量方法的发展。现有路面病害长度测量方法主要可分为人工方法与自动方法。其中，人工测量方法基于有经验的专家，通过步行或检测车方式，采用先进的路面探测装置对路面病害进行评估并量化其尺寸；典型的方法包括罗盘测量方法、平面测量方法以及离线数据分析方法。这种基于人工测量方式的路面病害长度测量方法不仅耗时较多、精度低，更难于满足现有高速公路养护工作的高时效性需求。因此，近年来不少研究人员都提出了路面病害长度测量的自动化测量方法，例如基于高速公路巡检测车对采集到在多源路面病害数据融合，进行病害检测与长度测量，同时有研究提出了一种基于深度学习的高速公路路面病害检测技术，采用少量的标注样本实现了高精度的路面病害检测与尺寸测量。然而，目前已实现的高速公路路面病害自动检测技术均是基于移动式装置，如巡检测车、移动手机等，无法直接测量出病害与路面的距离信息，同时采集的图像缺乏完整性信息，进而无法满足路网级高速公路路面养护工作量需求。In the past 30 years, many highway pavement disease detection methods have been proposed, which has promoted the development of pavement disease length measurement methods. The existing pavement damage length measurement methods can be mainly divided into manual methods and automatic methods. Among them, the manual measurement method is based on experienced experts, using advanced road detection devices to evaluate and quantify the size of road defects by walking or testing vehicles; typical methods include compass measurement methods, plane measurement methods, and offline data analysis methods . This method of measuring the length of pavement disease based on manual measurement is not only time-consuming and low in accuracy, but also more difficult to meet the high timeliness requirements of existing highway maintenance work. Therefore, in recent years, many researchers have proposed automatic measurement methods for the length measurement of road surface diseases. For example, based on highway patrol vehicles, the data fusion of road surface disease data collected from multiple sources is used for disease detection and length measurement. At the same time, some studies have proposed A deep learning-based highway pavement disease detection technology uses a small number of labeled samples to achieve high-precision pavement disease detection and size measurement. However, the existing automatic detection technologies for highway road surface diseases are all based on mobile devices, such as inspection vehicles, mobile phones, etc., which cannot directly measure the distance information between the disease and the road surface, and the collected images lack completeness information. Unable to meet the workload demand of road network-level highway pavement maintenance.

发明内容SUMMARY OF THE INVENTION

本发明的目的就是为了克服上述现有技术存在的检测工具种类单一、无法满足路网级高速公路路面养护工作量需求的缺陷而提供一种基于PTZ相机的高速公路路面病害长度测量方法。The purpose of the present invention is to provide a PTZ camera-based highway pavement disease length measurement method in order to overcome the above-mentioned defects in the prior art that the detection tools have a single type and cannot meet the road network-level highway pavement maintenance workload requirements.

本发明的目的可以通过以下技术方案来实现：The object of the present invention can be realized through the following technical solutions:

一种基于PTZ相机的高速公路路面病害长度测量方法，具体包括以下步骤：A method for measuring the length of highway pavement disease based on PTZ camera, which specifically includes the following steps:

步骤S1：获取不同视距下包含车道线的彩色的高速公路的路面图像；Step S1: Acquire color images of highways with lane lines at different sight distances;

步骤S2：对步骤S1中获取的路面图像进行Harris角点特征提取，并对完成角点提取的路面图像进行直线段提取，获得高速公路路面的直线段图像；Step S2: perform Harris corner feature extraction on the road surface image obtained in step S1, and perform straight line segment extraction on the road surface image for which corner point extraction has been completed, to obtain a straight line segment image of the highway road surface;

步骤S3：根据步骤S2中的高速公路路面的直线段图像，计算不同视距下高速公路的路面图像对应的像素精度；Step S3: Calculate the pixel accuracy corresponding to the road surface image of the expressway under different sight distances according to the straight line segment image of the expressway road in step S2;

步骤S4：根据不同视距下高速公路的路面图像对应的像素精度，计算相应高速公路的路面图像中病害区域的图像病害长度，根据所述图像病害长度计算得到病害区域的物理长度，实现高速公路路面的病害长度测量。Step S4: Calculate the image disease length of the diseased area in the road surface image of the corresponding expressway according to the pixel accuracy corresponding to the road surface image of the expressway under different sight distances, and calculate the physical length of the diseased area according to the image disease length, so as to realize the expressway. Disease length measurement of pavement.

所述直线段的类型包括竖直线型、水平线型、左对角线型和右对角线型。The types of the straight line segments include vertical line style, horizontal line style, left diagonal line style and right diagonal line style.

所述竖直线型和水平线型的直线段为路面图像的角点确定的直线段的侯选区域中的中位线，所述左对角线型和右对角线型的直线段为路面图像的角点确定的直线段的侯选区域中的对角线。The straight line segment of the vertical line type and the horizontal line type is the median line in the candidate area of the straight line segment determined by the corner point of the road surface image, and the straight line segment of the left diagonal line type and the right diagonal line type is the road surface The corners of the image determine the diagonal in the candidate region of the straight line segment.

所述步骤S2具体包括：The step S2 specifically includes:

步骤S201：选择所述角点确定的直线段的侯选区域内，对角线或中位线上一点P作为中心点；Step S201: In the candidate area of the straight line segment determined by the corner point, a point P on the diagonal or median line is selected as the center point;

步骤S202：以步骤S201中选择的点P为中心，在半径为r的邻域边缘上，按逆时针方向依次确定四个边缘像素点C1、C2、C3和C4，其中C1、C3的连线与C2、C4的连线互相垂直，对所述四个像素点进行二值转换，具体转换公式如下：Step S202: Taking the point P selected in step S201 as the center, on the edge of the neighborhood with a radius of r, determine four edge pixels C 1 , C 2 , C 3 and C 4 in turn in a counterclockwise direction, where C 1 , the connection line of C 3 and the connection line of C 2 and C 4 are perpendicular to each other, and binary conversion is performed on the four pixel points. The specific conversion formula is as follows:

Figure BDA0002282844000000021
Figure BDA0002282844000000021

其中，Ci为边缘像素点，I(xi,yi)为边缘像素点对应的图像像素值，I(ximod4+1,yimod4+1)表示将边缘像素点循环右移时相邻位对应的图像像素值，T为设定的减少图像噪声干扰的边缘阈值；Among them, C i is the edge pixel point, I(x i , y i ) is the image pixel value corresponding to the edge pixel point, and I(x imod4+1 , y imod4+1 ) represents the adjacent edge pixel point when the edge pixel is cyclically shifted to the right The image pixel value corresponding to the bit, T is the set edge threshold to reduce image noise interference;

重复执行上述计算步骤，直至完成路面图像内所有直线段侯选区域内的像素二值转换，得到路面图像对应的二值图像；Repeat the above calculation steps until the pixel binary conversion in the candidate regions of all straight line segments in the road surface image is completed, and a binary image corresponding to the road surface image is obtained;

步骤S203：统计路面图像的二值图像中1值像素的数量，以单个1值像素为中心1值像素点建立大小为k×k的窗口，统计窗口内所有与中心1值像素点相连的1值像素点，得到单个1值像素点对应的1值像素集合，根据所述1值像素集合计算出单个1值像素点对应的特征值，特征值计算公式具体如下：Step S203: Count the number of 1-valued pixels in the binary image of the road surface image, establish a window of size k×k with a single 1-valued pixel as the center 1-valued pixel, and count all the 1-valued pixels connected to the central 1-valued pixel in the window. value pixel point, obtain a 1-value pixel set corresponding to a single 1-value pixel point, and calculate the eigenvalue corresponding to a single 1-value pixel point according to the 1-value pixel set. The eigenvalue calculation formula is as follows:

Figure BDA0002282844000000031
Figure BDA0002282844000000031

Figure BDA0002282844000000032
Figure BDA0002282844000000032

Figure BDA0002282844000000033
Figure BDA0002282844000000033

其中，λ为单个1值像素点对应的特征值，cx为窗口内所有1值像素点横坐标的平均值，cy为窗口内所有1值像素点纵坐标的平均值，c11、c22、c12和c12为过程变量；Among them, λ is the eigenvalue corresponding to a single 1-value pixel, c x is the average value of the abscissas of all 1-value pixels in the window, c y is the average value of the ordinates of all 1-value pixels in the window, c 11 , c 22 , c 12 and c 12 are process variables;

同时生成小特征值图像，具体生成公式为：At the same time, a small eigenvalue image is generated, and the specific generation formula is:

Figure BDA0002282844000000034
Figure BDA0002282844000000034

其中，T[ge(x,y)]为单个1值像素点对应的小特征值图像，ge(x,y)为窗口内的1值像素点，Fj为所有单个1值像素点对应的窗口1值像素点集合；Among them, T[g e (x, y)] is the small eigenvalue image corresponding to a single 1-value pixel, g e (x, y) is the 1-value pixel in the window, and F j is all single 1-value pixels. The corresponding set of window 1-valued pixels;

步骤S204：路面图像的二值图像中所有1值像素点执行步骤S203，生成最终的小特征值图像，并根据阈值化公式对所述小特征值图像进行阈值化，得到高速公路路面直线段图像。Step S204: Step S203 is performed on all 1-valued pixels in the binary image of the road surface image to generate the final small eigenvalue image, and the small eigenvalue image is thresholded according to the thresholding formula to obtain a straight line segment image of the highway road surface .

所述阈值化公式具体为：The thresholding formula is specifically:

Figure BDA0002282844000000035
Figure BDA0002282844000000035

其中，

Figure BDA0002282844000000036
为阈值化后的最终的小特征值图像，
Figure BDA0002282844000000037
为未进行阈值化的最终的小特征值图像上所有单个1值像素点对应的小特征值图像，t为设定阈值。in,
Figure BDA0002282844000000036
is the final small eigenvalue image after thresholding,
Figure BDA0002282844000000037
is the small eigenvalue image corresponding to all single 1-value pixels on the final small eigenvalue image without thresholding, and t is the set threshold.

所述步骤S3具体为：The step S3 is specifically:

步骤S301：建立相机倾斜角、相机视距、图像上物体像素长度以及对应实际物体长度间的对应关系，具体对应关系如下所示：Step S301: Establish a corresponding relationship between the camera tilt angle, the camera viewing distance, the pixel length of the object on the image, and the length of the corresponding actual object. The specific corresponding relationship is as follows:

Figure BDA0002282844000000041
Figure BDA0002282844000000041

其中，f为PTZ相机的焦距，α为PTZ相机中摄像机的倾斜角，li为路面图像中距离图像底部边界最近的直线段对应的像素长度，oli为li对应线段到图像中心像素的像素长度，si为与感光板平行平面的中心位置距离路面的物理距离，di为PTZ相机的视距；Among them, f is the focal length of the PTZ camera, α is the tilt angle of the camera in the PTZ camera , li is the pixel length corresponding to the line segment closest to the bottom boundary of the image in the road image , and ol i is the distance between the line segment corresponding to li and the center pixel of the image Pixel length, s i is the physical distance from the center of the plane parallel to the photosensitive plate to the road surface, and d i is the sight distance of the PTZ camera;

步骤S302：根据步骤S204获取的高速公路路面直线段图像，提取距离图像底部边界最近的直线段并细化，得到直线段集，同时计算每条直线段距离图像中心像素的像素距离，结合步骤S301中的对应关系计算出路面图像的像素精度及其对应的PTZ相机的视距；Step S302: According to the straight line segment image of the highway road obtained in step S204, extract the straight line segment closest to the bottom boundary of the image and refine it to obtain a straight line segment set, and calculate the pixel distance of each straight line segment from the center pixel of the image, combined with step S301 The correspondence in calculates the pixel accuracy of the road image and the corresponding line-of-sight of the PTZ camera;

步骤S303：设定需要采集的路面图像的视距间隔，在不同视距下执行步骤S302，获得在不同视距下路面图像的像素精度。Step S303: Set the line-of-sight interval of the road surface image to be collected, and perform step S302 under different line-of-sight distances to obtain the pixel accuracy of the road surface image under different line-of-sight distances.

所述路面图像的像素精度及其对应的PTZ相机的视距的计算公式具体如下：The calculation formula of the pixel accuracy of the road image and the corresponding line-of-sight of the PTZ camera is as follows:

Figure BDA0002282844000000042
Figure BDA0002282844000000042

Figure BDA0002282844000000043
Figure BDA0002282844000000043

其中，api为路面图像的像素精度，w为车道线规范标准宽度，m为直线段的数量。Among them, api is the pixel accuracy of the road surface image, w is the standard width of the lane line specification, and m is the number of straight line segments.

所述病害区域的物理长度的计算公式具体如下：The calculation formula of the physical length of the diseased area is as follows:

cs＝{ls·api|di-1≤ds＜di,i＝1,k，2k,…,l/2cs={ls·ap i |d i-1 ≤d s <d i ,i=1,k,2k,...,l/2

其中，cs为病害区域的物理长度，ls为图像病害长度，k为视距间隔，l为路侧PTZ相机的安装间隔，ds为PTZ相机的当前视距。Among them, cs is the physical length of the diseased area, ls is the image disease length, k is the line-of-sight interval, l is the installation interval of the roadside PTZ camera, and ds is the current line-of-sight of the PTZ camera.

与现有技术相比，本发明具有以下有益效果：Compared with the prior art, the present invention has the following beneficial effects:

1.本发明采用路侧监控系统配备的PTZ相机，基于相机成像原理，采用图像处理方法对高速公路路面图像上病害进行实际长度测量，减少动态环境对图像处理算法干扰，且可快速应用于国内高速公路的路侧监控系统，实现高效稳定的路面病害长度测量。1. The present invention adopts the PTZ camera equipped with the roadside monitoring system, and based on the camera imaging principle, adopts the image processing method to measure the actual length of the disease on the highway pavement image, reduces the interference of the dynamic environment on the image processing algorithm, and can be quickly applied to domestic The roadside monitoring system of the expressway realizes efficient and stable measurement of the length of road surface diseases.

2.本发明借助高速公路路侧现有的高清摄像机设备即PTZ相机，无需额外的硬件成本投入，可节省高速公路路面养护成本。2. The present invention uses the existing high-definition camera equipment on the roadside of the expressway, that is, the PTZ camera, without additional hardware cost investment, and can save the cost of expressway road maintenance.

3.本发明通过获取高速公路路面的彩色图像，计算不同相机视下图像的像素精度，通过计算不同视距下检测的病害图像区域的像素长度，测量病害的物理长度，实现了高速公路不同路段、不同视距下病害长度的统计，为高速公路路面养护人员提供科学参照，对超出养护需求范围的病害及时进行修补，及时获取高速公路最佳养护时期，保障了高速公路服务质量与使用性能。3. The present invention calculates the pixel accuracy of images under different camera views by acquiring color images of highway pavement, and measures the physical length of diseases by calculating the pixel lengths of disease image areas detected under different line-of-sight distances, thereby realizing different sections of the expressway. , The statistics of the disease length under different sight distances provide a scientific reference for the highway pavement maintenance personnel, repair the diseases beyond the maintenance demand range in time, obtain the best maintenance period of the highway in time, and ensure the service quality and performance of the highway.

附图说明Description of drawings

图1为本发明的流程示意图；Fig. 1 is the schematic flow chart of the present invention;

图2为本发明原理及应用场景的示意图；Fig. 2 is the schematic diagram of the principle and application scenario of the present invention;

图3为本发明角点确定的直线段类型的示意图；Fig. 3 is the schematic diagram of the straight line segment type determined by the corner point of the present invention;

图4为本发明直线段检测结果的示意图；Fig. 4 is the schematic diagram of the detection result of the straight line segment of the present invention;

图5为本发明病害长度测量结果的示意图。Fig. 5 is a schematic diagram of the measurement results of the disease length of the present invention.

具体实施方式Detailed ways

下面结合附图和具体实施例对本发明进行详细说明。本实施例以本发明技术方案为前提进行实施，给出了详细的实施方式和具体的操作过程，但本发明的保护范围不限于下述的实施例。The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented on the premise of the technical solution of the present invention, and provides a detailed implementation manner and a specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

如图1所示，一种基于PTZ相机的高速公路路面病害长度测量方法，具体包括以下步骤：As shown in Figure 1, a method for measuring the length of highway pavement diseases based on PTZ camera specifically includes the following steps:

步骤S1：获取不同视距下包含车道线的彩色的高速公路的路面图像；Step S1: Acquire color images of highways with lane lines at different sight distances;

步骤S2：对步骤S1中获取的路面图像进行Harris角点特征提取，并对完成角点提取的路面图像进行直线段提取，获得高速公路路面的直线段图像；Step S2: perform Harris corner feature extraction on the road surface image obtained in step S1, and perform straight line segment extraction on the road surface image for which corner point extraction has been completed, to obtain a straight line segment image of the highway road surface;

步骤S3：根据步骤S2中的高速公路路面的直线段图像，计算不同视距下高速公路的路面图像对应的像素精度；Step S3: Calculate the pixel accuracy corresponding to the road surface image of the expressway under different sight distances according to the straight line segment image of the expressway road in step S2;

步骤S4：根据不同视距下高速公路的路面图像对应的像素精度，计算相应高速公路的路面图像中病害区域的图像病害长度，根据所述图像病害长度计算得到病害区域的物理长度，实现高速公路路面的病害长度测量。Step S4: Calculate the image disease length of the diseased area in the road surface image of the corresponding expressway according to the pixel accuracy corresponding to the road surface image of the expressway under different sight distances, and calculate the physical length of the diseased area according to the image disease length, so as to realize the expressway. Disease length measurement of pavement.

如图3所示，直线段的类型包括竖直线型、水平线型、左对角线型和右对角线型，竖直线型和水平线型的直线段为路面图像的角点确定的直线段的侯选区域中的中位线，所述左对角线型和右对角线型的直线段为路面图像的角点确定的直线段的侯选区域中的对角线。As shown in Figure 3, the types of straight line segments include vertical line type, horizontal line type, left diagonal line and right diagonal line. The median line in the candidate area of the segment, the straight line segments of the left diagonal type and the right diagonal type are the diagonal lines in the candidate area of the straight segment determined by the corner points of the road surface image.

步骤S2具体包括：Step S2 specifically includes:

步骤S201：根据角点的定义，确定图像角点间存在图像中的直线段信息，即每确定图像中的一个角点，该角点附近至少存在两条方向不同的线段或边缘，基于高速公路路面图像提取的角点序列，进行直线段侯选区域确定，并选择侯选区域内对角线或中位线上一点P作为中心点；Step S201: According to the definition of the corner points, it is determined that there are straight line segment information in the image between the corner points of the image, that is, each time a corner point in the image is determined, there are at least two line segments or edges with different directions near the corner point. For the corner point sequence extracted from the road surface image, determine the candidate area of the straight line segment, and select a point P on the diagonal or median line in the candidate area as the center point;

步骤S202：以步骤S201中选择的点P为中心，在半径为r的邻域边缘上，按逆时针方向依次确定四个边缘像素点C1、C2、C3和C4，其中C1、C3的连线与C2、C4的连线互相垂直，对四个像素点进行二值转换，具体转换公式如下：Step S202: Taking the point P selected in step S201 as the center, on the edge of the neighborhood with a radius of r, determine four edge pixels C 1 , C 2 , C 3 and C 4 in turn in a counterclockwise direction, where C 1 , the connection line of C3 and the connection line of C2 and C4 are perpendicular to each other, and the binary conversion is performed on the four pixel points. The specific conversion formula is as follows:

Figure BDA0002282844000000061
Figure BDA0002282844000000061

其中，Ci为边缘像素点，I(xi,yi)为边缘像素点对应的图像像素值，I(ximod4+1,yimod4+1)表示将边缘像素点循环右移时相邻位对应的图像像素值，T为设定的减少图像噪声干扰的边缘阈值；Among them, C i is the edge pixel point, I(x i , y i ) is the image pixel value corresponding to the edge pixel point, and I(x imod4+1 , y imod4+1 ) represents the adjacent edge pixel point when the edge pixel is cyclically shifted to the right The image pixel value corresponding to the bit, T is the set edge threshold to reduce image noise interference;

重复执行上述计算步骤，直至完成路面图像内所有直线段侯选区域内的像素二值转换，得到路面图像对应的二值图像；Repeat the above calculation steps until the pixel binary conversion in the candidate regions of all straight line segments in the road surface image is completed, and a binary image corresponding to the road surface image is obtained;

步骤S203：统计路面图像的二值图像中1值像素的数量，对于1值像素pi(i＝1,2,…,J)，以单个1值像素为中心1值像素点建立大小为k×k的窗口，统计窗口内所有与中心1值像素点相连的1值像素点，得到单个1值像素点对应的1值像素集合Fi＝{pi(xi,yi)|i＝1,2,…,n}，根据1值像素集合计算出单个1值像素点对应的特征值，特征值计算公式具体如下：Step S203: Count the number of 1-valued pixels in the binary image of the road surface image, for the 1-valued pixel pi (i=1, 2, . For the window of k, count all the 1-valued pixels connected to the central 1-valued pixel in the window, and obtain the set of 1-valued pixels corresponding to a single 1-valued pixel F i ={pi (x i ,y i )| i =1 ,2,…,n}, calculate the eigenvalue corresponding to a single 1-value pixel point according to the 1-value pixel set, and the eigenvalue calculation formula is as follows:

Figure BDA0002282844000000062
Figure BDA0002282844000000062

Figure BDA0002282844000000063
Figure BDA0002282844000000063

Figure BDA0002282844000000071
Figure BDA0002282844000000071

其中，λ为单个1值像素点对应的特征值，cx为窗口内所有1值像素点横坐标的平均值，cy为窗口内所有1值像素点纵坐标的平均值，c11、c22、c12和c12为过程变量；Among them, λ is the eigenvalue corresponding to a single 1-value pixel, c x is the average value of the abscissas of all 1-value pixels in the window, c y is the average value of the ordinates of all 1-value pixels in the window, c 11 , c 22 , c 12 and c 12 are process variables;

同时生成小特征值图像，具体生成公式为：At the same time, a small eigenvalue image is generated, and the specific generation formula is:

Figure BDA0002282844000000072
Figure BDA0002282844000000072

其中，T[ge(x,y)]为单个1值像素点对应的小特征值图像，ge(x,y)为窗口内的1值像素点，Fj为所有单个1值像素点对应的窗口1值像素点集合；Among them, T[g e (x, y)] is the small eigenvalue image corresponding to a single 1-value pixel, g e (x, y) is the 1-value pixel in the window, and F j is all single 1-value pixels. The corresponding set of window 1-valued pixels;

步骤S204：路面图像的二值图像中所有1值像素点执行步骤S203，生成最终的小特征值图像，并根据阈值化公式对小特征值图像进行阈值化，得到高速公路路面直线段图像，如图4所示为采用本发明提出的直线段检测方法对京石高速公路路面图像进行直线段检测的结果，图4(a)对应的相机视距为157m，图4(b)对应的相机视距为227m。Step S204: Step S203 is performed on all the 1-valued pixels in the binary image of the road surface image to generate the final small eigenvalue image, and the small eigenvalue image is thresholded according to the thresholding formula to obtain an image of a straight line segment of the highway road surface, such as Figure 4 shows the result of detecting a straight line segment on the road image of the Jingshi Expressway using the straight line segment detection method proposed by the present invention. The distance is 227m.

阈值化公式具体为：The thresholding formula is specifically:

Figure BDA0002282844000000073
Figure BDA0002282844000000073

其中，

Figure BDA0002282844000000074
为阈值化后的最终的小特征值图像，
Figure BDA0002282844000000075
为未进行阈值化的最终的小特征值图像上所有单个1值像素点对应的小特征值图像，t为设定阈值。in,
Figure BDA0002282844000000074
is the final small eigenvalue image after thresholding,
Figure BDA0002282844000000075
is the small eigenvalue image corresponding to all single 1-value pixels on the final small eigenvalue image without thresholding, and t is the set threshold.

步骤S3具体为：Step S3 is specifically:

步骤S301：建立相机倾斜角、相机视距、图像上物体像素长度以及对应实际物体长度间的对应关系，具体对应关系如下所示：Step S301: Establish a corresponding relationship between the camera tilt angle, the camera viewing distance, the pixel length of the object on the image, and the length of the corresponding actual object. The specific corresponding relationship is as follows:

Figure BDA0002282844000000076
Figure BDA0002282844000000076

其中，f为PTZ相机的焦距，α为PTZ相机中摄像机的倾斜角，即成像感光板(CCD/CMOS)与路面垂线的夹角，li为路面图像中距离图像底部边界最近的直线段对应的像素长度，oli为li对应线段到图像中心像素的像素长度，si为与感光板平行平面的中心位置距离路面的物理距离，di为PTZ相机的视距；Among them, f is the focal length of the PTZ camera, α is the tilt angle of the camera in the PTZ camera, that is, the angle between the imaging photosensitive plate (CCD/CMOS) and the vertical line of the road surface, and li is the straight line segment in the road image that is closest to the bottom boundary of the image Corresponding pixel length, ol i is the pixel length of the line segment corresponding to l i to the center pixel of the image, s i is the physical distance from the center position of the parallel plane to the photosensitive plate and the road surface, and d i is the sight distance of the PTZ camera;

如图2所示，与感光板平行平面为虚拟平面，真实情况下难以直接计算得到，因此无法直接获取该虚拟平面与路面的切线位置，根据路面在图像上投影由图像底部至顶部呈非线性拉伸，因此选择高速公路路面图像上距离图像底部最近的直线段，即图2中获取图像区域内竖直虚线段对应的像素长度li近似作为与感光板平行平面与路面切线在图像上投影时与车道线在水平方向上的交集线段对应的像素长度；As shown in Figure 2, the plane parallel to the photosensitive plate is a virtual plane, which is difficult to calculate directly in reality. Therefore, the tangent position between the virtual plane and the road cannot be directly obtained. According to the projection of the road on the image, it is nonlinear from the bottom to the top of the image. Therefore, the straight line segment closest to the bottom of the image on the highway road surface image is selected, that is, the pixel length li corresponding to the vertical dashed line segment in the acquired image area in Figure 2 is approximately projected on the image as the parallel plane with the photosensitive plate and the tangent to the road surface is the pixel length corresponding to the intersection line of the lane line in the horizontal direction;

步骤S302：根据步骤S204获取的高速公路路面直线段图像，提取距离图像底部边界最近的直线段，如图2所示的直线段l1，l2，l3并细化，得到直线段集li(i＝1,2,…m)，同时计算每条直线段距离图像中心像素的像素距离oli(i＝1,2,…m)，结合步骤S301中的对应关系计算出路面图像的像素精度及其对应的PTZ相机的视距，路面图像的像素精度及其对应的PTZ相机的视距的计算公式具体如下：Step S302: According to the straight line segment image of the highway road obtained in step S204, extract the straight line segment closest to the bottom boundary of the image, such as straight line segments l 1 , l 2 , and l 3 as shown in FIG. 2 , and refine them to obtain a straight line segment set l i (i=1,2,...m), at the same time calculate the pixel distance ol i (i=1,2,...m) of each straight line segment from the center pixel of the image, and calculate the road surface image according to the corresponding relationship in step S301. The calculation formulas of the pixel accuracy and the corresponding PTZ camera's line-of-sight, the pixel accuracy of the road image and the corresponding PTZ camera's line-of-sight are as follows:

Figure BDA0002282844000000081
Figure BDA0002282844000000081

Figure BDA0002282844000000082
Figure BDA0002282844000000082

其中，api为路面图像的像素精度，w为车道线规范标准宽度，m为直线段的数量；Among them, api is the pixel accuracy of the road image, w is the standard width of the lane line specification, m is the number of straight line segments;

步骤S303：以k作为相机视距间隔，依次采集不同相机视距di(i＝k,i＝2k,…,l/2)对应的高速公路路面图像，在不同视距下执行步骤S302，统计di(i＝k,i＝2k,…,l/2)对应的路面图像的像素精度，得到不同相机视距下的像素精度AP＝{api|i＝1,k,2k,…,l/2}，如表1所示为京石高速公路某路段采集到的路面图像计算得到的路侧PTZ相机视距对应路面图像像素精度统计结果，表1具体内容如下所示：Step S303: Take k as the camera sight distance interval, sequentially collect highway road surface images corresponding to different camera sight distances d i (i=k, i=2k,...,l/2), and perform step S302 under different sight distances, Count the pixel accuracy of the road image corresponding to d i (i=k, i=2k,...,l/2), and obtain the pixel accuracy AP={ap i |i=1,k,2k,... ,l/2}, as shown in Table 1, the statistical results of the roadside PTZ camera line-of-sight corresponding to the pixel accuracy of the road surface image calculated from the road surface image collected from a certain section of the Jingshi Expressway. The specific content of Table 1 is as follows:

表1京石高速公路某路段相机视距与路面图像像素精度对照表Table 1 A comparison table of camera line-of-sight and road image pixel accuracy in a section of Jingshi Expressway

Figure BDA0002282844000000083
Figure BDA0002282844000000083

步骤S4具体为对给定视距ds下采集的包含病害的高速公路路面图像进行病害检测，得到病害区域，并计算病害区域对应的图像病害长度，根据图像病害长度计算得到病害区域的物理长度，病害区域的物理长度的计算公式具体如下：Step S4 is specifically performing disease detection on the highway road surface image containing the disease collected under the given sight distance d s , obtaining the disease area, and calculating the image disease length corresponding to the disease area, and calculating the physical length of the disease area according to the image disease length. , the calculation formula of the physical length of the diseased area is as follows:

cs＝{ls·api|di-1≤ds＜di,i＝1,k,2k,…,l/2cs={ls·ap i |d i-1 ≤d s <d i ,i=1,k,2k,...,l/2

其中，cs为病害区域的物理长度，ls为图像病害长度，k为视距间隔，l为路侧PTZ相机的安装间隔，ds为PTZ相机的当前视距；Among them, cs is the physical length of the diseased area, ls is the image disease length, k is the line-of-sight interval, l is the installation interval of the roadside PTZ camera, and ds is the current line-of-sight of the PTZ camera;

如图5所示为15米与40米的相机视距下采集的京石高速公路路面图像，其中图5(a)中病害区域的物理长度为148.59mm，图5(b)中病害区域的物理长度为476.95mm。As shown in Figure 5, the road surface images of Jingshi Expressway collected under the camera line-of-sight distance of 15 meters and 40 meters, in which the physical length of the diseased area in Figure 5(a) is 148.59mm, and the The physical length is 476.95mm.

此外，需要说明的是，本说明书中所描述的具体实施例，所取名称可以不同，本说明书中所描述的以上内容仅仅是对本发明结构所做的举例说明。凡依据本发明构思的构造、特征及原理所做的等小变化或者简单变化，均包括于本发明的保护范围内。本发明所属技术领域的技术人员可以对所描述的具体实例做各种各样的修改或补充或采用类似的方法，只要不偏离本发明的结构或者超越本权利要求书所定义的范围，均应属于本发明的保护范围。In addition, it should be noted that the names of the specific embodiments described in this specification may be different, and the above content described in this specification is only an example to illustrate the structure of the present invention. All minor changes or simple changes made according to the structure, features and principles of the present invention are included in the protection scope of the present invention. Those skilled in the art to which the present invention pertains can make various modifications or additions to the specific examples described or adopt similar methods, as long as they do not deviate from the structure of the present invention or go beyond the scope defined by the claims, all It belongs to the protection scope of the present invention.

Claims (5)

1. A method for measuring the length of a pavement defect of an expressway based on a PTZ camera is characterized by comprising the following steps:
step S1: acquiring road surface images of colored highways containing lane lines at different sight distances;
step S2: performing Harris corner feature extraction on the road surface image obtained in the step S1, and performing straight line segment extraction on the road surface image subjected to corner extraction to obtain a straight line segment image of the highway road surface;
step S3: calculating the pixel precision corresponding to the road surface image of the expressway under different visual ranges according to the straight-line segment image of the expressway road surface in the step S2;
step S4: calculating the image disease length of the disease area in the pavement image of the corresponding expressway according to the pixel precision corresponding to the pavement image of the expressway under different visual distances, and calculating the physical length of the disease area according to the image disease length to realize the disease length measurement of the expressway pavement;
the step S3 specifically includes:
step S301: establishing a corresponding relation among a camera inclination angle, a camera sight distance, an object pixel length on an image and a corresponding actual object length, wherein the specific corresponding relation is as follows:
Figure FDA0002953951730000011
where f is the focal length of the PTZ camera, α is the tilt angle of the camera in the PTZ camera, liThe pixel length, ol, corresponding to the straight line segment closest to the image bottom boundary in the road surface imageiIs 1iPixel length, s, from line segment to image center pixeliIs the physical distance from the center of the plane parallel to the plate to the road surface, diThe view distance of the PTZ camera is shown, and w is the standard width of the lane line specification;
step S302: according to the highway pavement straight-line segment image obtained in the step S204, extracting a straight-line segment closest to the bottom boundary of the image and thinning the straight-line segment to obtain a straight-line segment set, meanwhile, calculating the pixel distance of each straight-line segment from the central pixel of the image, and calculating the pixel precision of the pavement image and the sight distance of the corresponding PTZ camera by combining the corresponding relation in the step S301;
step S303: setting the sight distance interval of the road surface image to be acquired, and executing the step S302 under different sight distances to obtain the pixel precision of the road surface image under different sight distances;
the calculation formula of the pixel precision of the road surface image and the sight distance of the PTZ camera corresponding to the pixel precision is as follows:
Figure FDA0002953951730000012
Figure FDA0002953951730000021
wherein apiThe pixel precision of the road surface image is shown, w is the standard width of the lane line specification, and m is the number of straight line segments;
the above-mentionedStep S4 is embodied as a step for a given viewing distance dsDetecting the diseases of the collected highway pavement images containing the diseases to obtain a disease area, calculating the length of the image diseases corresponding to the disease area, and calculating the physical length of the disease area according to the length of the image diseases, wherein the calculation formula of the physical length of the disease area is as follows:
cs＝{ls·api|di-1≤ds＜di,i＝1,k,2k,…,l/2}
wherein cs is the physical length of the disease area, ls is the image disease length, k is the view distance interval, l is the installation interval of the roadside PTZ camera, dsIs the current line of sight of the PTZ camera.
2. The PTZ camera-based highway pavement damage length measuring method according to claim 1, wherein the types of the straight line segments comprise a vertical line segment, a horizontal line segment, a left diagonal line segment and a right diagonal line segment.
3. The PTZ camera-based highway pavement damage length measuring method according to claim 2, wherein the straight line segments of the vertical line type and the horizontal line type are median lines in candidate areas of the straight line segments determined by the corner points of the pavement image, and the straight line segments of the left diagonal line type and the right diagonal line type are diagonal lines in candidate areas of the straight line segments determined by the corner points of the pavement image.
4. The method for measuring the length of the pavement damage of the expressway based on the PTZ camera as claimed in claim 3, wherein the step S2 specifically comprises the following steps:
step S201: selecting a point P on a diagonal line or a median line in a candidate area of a straight line segment determined by the corner points as a central point;
step S202: with the point P selected in the step S201 as the center, four edge pixel points C are sequentially determined in the counterclockwise direction on the neighborhood edge with the radius r1、C2、C3And C4In which C is1、C3Of (2) a connection lineAnd C2、C4The connecting lines are mutually vertical, and binary conversion is carried out on the four pixel points, wherein the specific conversion formula is as follows:
Figure FDA0002953951730000022
wherein, CiAs edge pixels, I (x)i,yi) Image pixel values, I (x), corresponding to edge pixel pointsimod4+1,yimod4+1) Representing image pixel values corresponding to adjacent positions of the edge pixel points when the edge pixel points are circularly shifted to the right, wherein T is a set edge threshold value for reducing image noise interference;
repeatedly executing the calculation steps until pixel binary conversion in all the straight-line segment candidate areas in the road surface image is completed, and obtaining a binary image corresponding to the road surface image;
step S203: counting the number of 1-value pixels in a binary image of a pavement image, establishing a window with the size of k multiplied by k by taking a single 1-value pixel as a central 1-value pixel, counting all 1-value pixels connected with the central 1-value pixel in the window to obtain a 1-value pixel set corresponding to the single 1-value pixel, and calculating a characteristic value corresponding to the single 1-value pixel according to the 1-value pixel set, wherein the characteristic value calculation formula is as follows:
Figure FDA0002953951730000031
Figure FDA0002953951730000032
Figure FDA0002953951730000033
wherein, λ is the characteristic value corresponding to a single 1-value pixel point, cxIs the average value of the abscissa of all 1-value pixel points in the window, cyIs the average value of the vertical coordinates of all 1-value pixel points in the window, c11、c22、c12And c21Is a process variable;
and simultaneously generating a small characteristic value image, wherein a specific generation formula is as follows:
Figure FDA0002953951730000034
wherein, T [ g ]e(x,y)]For small feature value images, g, corresponding to a single 1-value pixel pointe(x, y) are 1-valued pixels in a window, FjA window 1 value pixel point set corresponding to all single 1 value pixel points;
step S204: and (3) executing step S203 on all 1-value pixel points in the binary image of the road surface image to generate a final small characteristic value image, and thresholding the small characteristic value image according to a thresholding formula to obtain the straight-line-section image of the highway road surface.
5. The method for measuring the length of the highway pavement damage based on the PTZ camera as claimed in claim 4, wherein the thresholding formula is specifically as follows:
Figure FDA0002953951730000035
wherein,
Figure FDA0002953951730000036
for the final small eigenvalue image after thresholding,
Figure FDA0002953951730000037
and t is a set threshold value, wherein the small characteristic value images correspond to all the single 1-value pixel points on the final small characteristic value image which is not subjected to thresholding.
CN201911148337.6A 2019-11-21 2019-11-21 A method for measuring the length of highway pavement disease based on PTZ camera Active CN110969103B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911148337.6A CN110969103B (en) 2019-11-21 2019-11-21 A method for measuring the length of highway pavement disease based on PTZ camera

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911148337.6A CN110969103B (en) 2019-11-21 2019-11-21 A method for measuring the length of highway pavement disease based on PTZ camera

Publications (2)

Publication Number Publication Date
CN110969103A CN110969103A (en) 2020-04-07
CN110969103B true CN110969103B (en) 2021-05-11

Family

ID=70031162

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911148337.6A Active CN110969103B (en) 2019-11-21 2019-11-21 A method for measuring the length of highway pavement disease based on PTZ camera

Country Status (1)

Country Link
CN (1) CN110969103B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114018932B (en) * 2021-11-02 2023-05-30 西安电子科技大学 Measuring Method of Pavement Disease Index Based on Rectangular Calibration Object

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102509291A (en) * 2011-10-31 2012-06-20 东南大学 Pavement disease detecting and recognizing method based on wireless online video sensor
CN103345542A (en) * 2013-06-05 2013-10-09 西安理工大学 Method for marking bridge disease based on three-dimensional visualization technology
CN106053475A (en) * 2016-05-24 2016-10-26 浙江工业大学 Tunnel disease full-section dynamic rapid detection device based on active panoramic vision
CN106296691A (en) * 2016-08-11 2017-01-04 梁晓东 Pavement disease recognition methods based on graphical analysis
CN107633516A (en) * 2017-09-21 2018-01-26 武汉武大卓越科技有限责任公司 A kind of method and apparatus for identifying surface deformation class disease
CN108733053A (en) * 2018-04-23 2018-11-02 上海圭目机器人有限公司 A kind of Intelligent road detection method based on robot
CN109870456A (en) * 2019-02-01 2019-06-11 上海智能交通有限公司 A kind of road surface health status rapid detection system and method

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103955923B (en) * 2014-04-18 2017-03-01 南京理工大学 A kind of quickly pavement disease detection method based on image
CN104021574B (en) * 2014-07-04 2016-12-07 武汉武大卓越科技有限责任公司 Pavement disease automatic identifying method
US10480939B2 (en) * 2016-01-15 2019-11-19 Fugro Roadware Inc. High speed stereoscopic pavement surface scanning system and method
CN106677037B (en) * 2016-11-21 2020-11-27 同济大学 Portable Asphalt Pavement Disease Detection Method and Device Based on Machine Vision
CN106778541B (en) * 2016-11-28 2019-08-13 华中科技大学 A kind of identification in the multilayer beam China and foreign countries beam hole of view-based access control model and localization method
CN107462223B (en) * 2017-08-08 2020-06-16 长安大学 A kind of automatic measuring device and measuring method of vehicle sight distance before road turning
CN107609486A (en) * 2017-08-16 2018-01-19 中国地质大学（武汉） To anti-collision early warning method and system before a kind of vehicle
CN109934452A (en) * 2019-01-21 2019-06-25 上海同济检测技术有限公司 Road Comfort Evaluation method based on multi-source data

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102509291A (en) * 2011-10-31 2012-06-20 东南大学 Pavement disease detecting and recognizing method based on wireless online video sensor
CN103345542A (en) * 2013-06-05 2013-10-09 西安理工大学 Method for marking bridge disease based on three-dimensional visualization technology
CN106053475A (en) * 2016-05-24 2016-10-26 浙江工业大学 Tunnel disease full-section dynamic rapid detection device based on active panoramic vision
CN106296691A (en) * 2016-08-11 2017-01-04 梁晓东 Pavement disease recognition methods based on graphical analysis
CN107633516A (en) * 2017-09-21 2018-01-26 武汉武大卓越科技有限责任公司 A kind of method and apparatus for identifying surface deformation class disease
CN108733053A (en) * 2018-04-23 2018-11-02 上海圭目机器人有限公司 A kind of Intelligent road detection method based on robot
CN109870456A (en) * 2019-02-01 2019-06-11 上海智能交通有限公司 A kind of road surface health status rapid detection system and method

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
吉林省高速公路路面路基病害同步检测的技术研究;刘若川;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20160315(第3期);第C034-198页 *

Also Published As

Publication number Publication date
CN110969103A (en) 2020-04-07

Similar Documents

Publication Publication Date Title
CN110910359B (en) FPC circuit detection method based on feature tracking and skeleton feature matching
CN115082377A (en) Building surface crack geometric parameter measuring method and system based on unmanned aerial vehicle
CN110929690B (en) Remote sensing image road network extraction method based on deep neural network
CN111383204A (en) Video image fusion method, fusion device, panoramic monitoring system and storage medium
CN111145157A (en) Road network data automatic quality inspection method based on high-resolution remote sensing image
Liang et al. Research on concrete cracks recognition based on dual convolutional neural network
CN116071721A (en) Transformer-based high-precision map real-time prediction method and system
CN114298972A (en) Method for calculating road pavement damage degree
CN113011283A (en) Non-contact type rail sleeper relative displacement real-time measurement method based on video
CN116109616A (en) Pavement crack detection and small-surface element fitting detection method based on YOLOv5
CN110969103B (en) A method for measuring the length of highway pavement disease based on PTZ camera
CN111080640B (en) Hole detection method, device, equipment and medium
CN114111602B (en) Bridge surface crack width calculation method based on image technology
CN115330684A (en) Underwater structure apparent defect detection method based on binocular vision and line structured light
CN110956640B (en) A Method of Edge Point Detection and Registration in Heterogeneous Images
CN110969576B (en) Highway pavement image splicing method based on roadside PTZ camera
CN110942026B (en) Deep learning-based capsule robot drain pipe disease detection method and system
CN116758507B (en) Pavement quality analysis method, device and program based on disease image acquisition and segmentation
CN118196028A (en) An improved YOLOv8 method for extracting rural cement pavement defects
CN104881669A (en) Method and system for extracting local area detector based on color contrast
CN116152190A (en) A road damage target detection system combining deep learning and image analysis
Xing et al. A lightweight detection method of pavement potholes based on binocular stereo vision and deep learning
CN107014303B (en) A Track Creep Monitoring Method Based on Digital Image Processing
Pan et al. Embedded U-Net: combines multiple feature fusion encode and subpixel reconstruction for microcracks salient object detection
CN111429430A (en) Calculation method of dimension mapping relationship of tunnel inspection vehicle based on machine vision

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant