WO2013013563A1 - 路面积水积冰检测方法及装置 - Google Patents
路面积水积冰检测方法及装置 Download PDFInfo
- Publication number
- WO2013013563A1 WO2013013563A1 PCT/CN2012/078042 CN2012078042W WO2013013563A1 WO 2013013563 A1 WO2013013563 A1 WO 2013013563A1 CN 2012078042 W CN2012078042 W CN 2012078042W WO 2013013563 A1 WO2013013563 A1 WO 2013013563A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- road
- road surface
- brightness
- camera
- Prior art date
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 27
- 238000000034 method Methods 0.000 claims abstract description 45
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 43
- 238000003384 imaging method Methods 0.000 claims description 27
- 238000012545 processing Methods 0.000 claims description 21
- 238000005457 optimization Methods 0.000 claims description 11
- 238000001914 filtration Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 2
- 238000012986 modification Methods 0.000 abstract description 6
- 230000004048 modification Effects 0.000 abstract description 6
- 238000012360 testing method Methods 0.000 abstract description 3
- 230000010287 polarization Effects 0.000 description 18
- 239000011159 matrix material Substances 0.000 description 16
- 239000013598 vector Substances 0.000 description 11
- 230000009466 transformation Effects 0.000 description 8
- 238000010586 diagram Methods 0.000 description 7
- 230000003287 optical effect Effects 0.000 description 7
- 238000009825 accumulation Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 230000006872 improvement Effects 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012805 post-processing Methods 0.000 description 2
- PXFBZOLANLWPMH-UHFFFAOYSA-N 16-Epiaffinine Natural products C1C(C2=CC=CC=C2N2)=C2C(=O)CC2C(=CC)CN(C)C1C2CO PXFBZOLANLWPMH-UHFFFAOYSA-N 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000037406 food intake Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/21—Polarisation-affecting properties
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J1/00—Photometry, e.g. photographic exposure meter
- G01J1/02—Details
- G01J1/04—Optical or mechanical part supplementary adjustable parts
- G01J1/0407—Optical elements not provided otherwise, e.g. manifolds, windows, holograms, gratings
- G01J1/0429—Optical elements not provided otherwise, e.g. manifolds, windows, holograms, gratings using polarisation elements
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J1/00—Photometry, e.g. photographic exposure meter
- G01J1/10—Photometry, e.g. photographic exposure meter by comparison with reference light or electric value provisionally void
- G01J1/16—Photometry, e.g. photographic exposure meter by comparison with reference light or electric value provisionally void using electric radiation detectors
- G01J1/1626—Arrangements with two photodetectors, the signals of which are compared
Definitions
- the present invention relates to the field of optical communications, and in particular to a road area water ice detecting method and apparatus.
- BACKGROUND OF THE INVENTION A road condition monitor of the related art mainly uses a principle that infrared rays or microwaves have different absorption rates on dry roads and ice and snow roads, and the road surface state detection is adopted. The detection mode has the following characteristics: limited sampling space, high price, and inability to follow the road The layout of the section is large and it is difficult to reflect the true condition of the entire section.
- the detection method is to distinguish different road conditions by using the maximum full width and spatial frequency distribution of the monitored image.
- this method requires an artificial light source as an aid, and can only be used at night, and the position of the light source and the CCD camera is fixed, resulting in a small detection range and difficulty in popularization.
- Embodiments of the present invention provide a road area water ice detecting method and apparatus to solve at least the above problems.
- a road area water ice detecting method including: acquiring a first road image image captured by a first camera through a horizontal polarizing plate on a road surface in an imaging range; acquiring a second camera through a second road image image captured by the vertical polarizer on the road surface in the imaging range, wherein the image plane of the image sensor (CCD) of the second camera is coplanar with the photographing plane of the CCD of the first camera, and the second camera and the second camera The distance between the cameras does not exceed a predetermined value; determining whether the difference between the brightness of the first road image image and the brightness of the second road image image is greater than a preset brightness range, and if so, determining that there is water on the road surface in the imaging range Or ice accretion.
- CCD image sensor
- a road area water ice detecting device including: a first acquiring module, configured to acquire a first road surface captured by a first camera through a horizontal diaphragm to a road surface in an imaging range; An image acquisition screen is configured to acquire a second road image image captured by the second camera through the vertical polarizer on the road surface in the imaging range, and a shooting plane of the image sensor (CCD) of the second camera and the first camera The shooting plane of the CCD is coplanar, and the distance between the second camera and the first camera does not exceed a predetermined value; the determining module is configured to determine whether the difference between the brightness of the first road image image and the brightness of the second road image image is greater than a predetermined The set brightness range, if yes, it is determined that there is water or ice accumulating on the road surface in the imaging range.
- a first acquiring module configured to acquire a first road surface captured by a first camera through a horizontal diaphragm to a road surface in an imaging range
- An image acquisition screen is configured
- the horizontally polarized plate and the vertical polarizing plate are respectively disposed in front of the first camera and the second camera to detect the road area water ice, which solves the complicated road surface detection method in the prior art.
- the detection range is small, and the implementation is difficult, and the effect is stable, the detection range is wide, and the application value is large.
- FIG. 1 is a flow chart of a road area water ice detecting method according to an embodiment of the present invention
- FIG. 2 is a flow chart of road area water ice detecting according to a preferred embodiment of the present invention
- 4 is a schematic view showing a state of reflection of a water accumulated ice road surface and a normal road surface according to an embodiment of the present invention
- FIG. 5 is a schematic view showing a state of total reflected light of a road surface according to an embodiment of the present invention
- FIG. 6 is a double view according to a preferred embodiment of the present invention.
- FIG. 1 is a flow chart of a road area water ice detecting method according to an embodiment of the present invention
- FIG. 2 is a flow chart of road area water ice detecting according to a preferred embodiment of the present invention
- 4
- FIG. 7 is a schematic diagram showing changes in road surface light under a horizontal polarizer according to an embodiment of the present invention
- FIG. 8 is a schematic diagram showing changes in road light under a vertical polarizer according to an embodiment of the present invention
- FIG. 10 is a schematic structural view of a road area water ice detecting device according to a preferred embodiment of the present invention
- FIG. 11 is a schematic view of a road area water product according to a preferred embodiment of the present invention
- Step S102 Acquire a first road image image captured by the first camera through the horizontal polarizer on the road surface in the imaging range.
- the first camera may generate a first road image image according to the first reflected light transmitted through the horizontal polarizer, and then upload the first road image image to the designated image processing device.
- Step S104 acquiring a second road image image captured by the second camera through the vertical polarizing plate on the road surface in the imaging range, wherein the imaging plane of the image sensor (CCD) of the second camera is shared with the imaging plane of the CCD of the first camera. And the distance between the second camera and the first camera does not exceed a predetermined value.
- the second camera may generate a second road image image according to the second reflected light transmitted through the vertical polarizing plate, and then upload the second road image image to a designated image processing device, for example, a PC.
- step S106 it is determined whether the difference between the brightness of the first road surface image screen and the brightness of the second road surface image screen is greater than a brightness range set in advance, and if so, it is determined that there is water or ice accumulating on the road surface in the imaging range.
- the first road image captured by the first camera through the horizontal vibration plate and the second road surface captured by the second camera through the vertical polarizer on the road surface in the imaging range are obtained.
- image registration can be performed on the first road image screen and the second road image screen.
- the first road image image and the second road image image may be image-registered using the SURF algorithm.
- the image optimization processing operation may be performed on the first road image image and the second road image image.
- the image optimization processing operations mainly include: , filtering and binarization. Then, it is determined whether the difference between the brightness of the first road image image and the brightness of the second road image image is greater than a brightness range set in advance, and if so, it can be determined that there is water or ice accumulating on the road surface in the imaging range. For example, in a practical application, the first picture brightness value of the first road surface image image after the image optimization processing operation may be acquired, and then the second picture brightness value of the second road surface image picture after the image optimization processing operation may be acquired.
- the difference between the first picture brightness value and the second picture brightness value is greater than a preset threshold, and if yes, determining the first road image image The difference between the brightness and the brightness of the second road image image is greater than a preset brightness range. Otherwise, it is determined that the difference between the brightness of the first road image image and the brightness of the second road image image is less than a preset brightness range. It can be known from the above that when the difference between the brightness of the first road image image and the brightness of the second road image image is greater than the preset brightness range, there is water or ice accumulation on the current monitored road surface, and vice versa.
- the current monitored road surface is a normal road surface, and there is no accumulated water or ice accretion.
- the method can be applied to a nighttime environment or a particularly dark environment. When it is necessary to detect the road surface at night (or a particularly dark environment), whether there is water or ice accumulation, a separate setting can be set.
- the light source to be measured is a light source with an incident angle of 53 degrees. Although the light emitted by the light source is different from the natural light, it does not affect the reflection of water or ice accumulated on the road surface, and is then passed through the horizontal polarizer and the vertical polarizer. Camera ingestion. The above method will be described in detail below by taking a binocular camera as an example.
- FIG. 2 is a flow chart of road area water ice accumulation detection in accordance with a preferred embodiment of the present invention.
- the polarizing plates in the vertical direction and the horizontal direction are respectively placed on the front end of the binocular camera.
- two polarizing plates can be placed on the front end of the binocular camera, and images are acquired and image registration is performed.
- the post-processing measurement uses the polarization characteristics of the light to detect the accumulation of ice on the road surface by the polarized light reflected from the road surface.
- two cameras of the same type can be erected on the road surface, the camera parameters are set the same, the front polarizing plates are placed horizontally and vertically, the distance between the cameras is as small as possible, and the image planes are coincident to ensure that the two cameras are photographed.
- the overlapping area of the image is as large as possible.
- the polarizing plates in the vertical direction and the horizontal direction can be respectively placed in front of the binocular camera while acquiring images and performing measurement of image registration and post-processing. It is well known that the human visual system is able to perceive the frequency and intensity characteristics of light in the form of color and brightness, but it is not directly perceptible to another fundamental feature of light, the polarization characteristics.
- the traditional image processing and understanding process is based on the signal processing of intensity and frequency domain. It can only make some preliminary analysis and judgment on the contour and category of the target, but can not distinguish the material and detail features of the target.
- an electromagnetic wave light has the polarization characteristics of the transverse wave, and the difference in the conductivity and smoothness of the material of the reflecting surface will cause a difference in the polarization characteristics of the light.
- a natural light can be decomposed into two mutually equal, equal-width, and incoherent linearly polarized light. For an electric vector incident on a plane, it can be decomposed into a component perpendicular to the incident surface and a component parallel to the incident surface. According to the Fresnel reflection model, the two components of the reflected electrical vector can be expressed as: ( 1 ); sin ( - ⁇ 2 )
- the component drops sharply, especially at 2 o'clock, the incident angle is Brewster's angle
- an artificial light source can be used as an aid to enable the light source to detect the surface of the road at a 53 degree angle of incidence (ie, Brewster's angle).
- a 53 degree angle of incidence ie, Brewster's angle.
- FIG. 4 As shown in FIG. 4, when the road surface is covered with water or ice, the surface is smooth, and when the light is irradiated, the above-mentioned polarized light is obtained after specular reflection, and the normal road surface is diffusely reflected. When there is no water or ice on the road surface, the surface is rough and the incident angle is not fixed. Therefore, the reflected light has vibration in all directions and has no statistical property.
- the aging pavement tends to be smooth, it may exhibit a certain degree of polarization, but the degree of polarization of the reflected light relative to the accumulated water or ice accumulating surface is negligible, so it can be filtered as noise during processing.
- the reflection of the normal road surface is unpolarized light, and the accumulated water ice road surface is reflected as horizontally linearly polarized light, and the superimposed partial polarized light enters the camera, because the polarization direction is horizontal, therefore, Brightness is higher in horizontally polarized images and lower in vertically polarized images.
- the SURF algorithm can be used to obtain the feature point pairs of the original images of camera A and camera B, and the approximate nearest neighbor algorithm (BBF) method is used to perform rough matching on the feature point pairs, and the structure is based on the perspective transformation matrix H.
- the image mapping model then uses RANSAC to further reject the erroneous matches and uses the least squares method to find the exact H matrix between the images to obtain the registered image.
- the image quality obtained is often poor due to the existence of complex scenes such as highways, and the road surface image is caused by factors such as large brightness variation, large dimensional change, large rotation and interference of moving objects.
- the difficulty of registration At this time, the road image registration method is required to be robust to these unfavorable factors.
- the extracted feature (SIFT) algorithm is robust due to its large rotation, size scaling, chromatic aberration, vision and illumination changes. Sex, especially suitable for complex image registration.
- problems such as large data volume and long calculation time have adversely affected the real-time performance of image registration.
- the SURF algorithm is introduced into the process of road surface image registration.
- the SURF is similar in concept to SIFT, focusing on the spatial distribution of gradient information, and inherits the advantages of SIFT algorithm for rotation, scale scaling, brightness change invariance, and robustness to viewing angle variation, image blur, and noise.
- the SURF algorithm makes full use of the box filter and the integral image in the feature point detection and description vectors to speed up the calculation and reduce the dimension of the local image description vector, which greatly accelerates the detection of the feature points.
- the generation of the description vector can significantly improve the speed of the image registration method.
- the extraction of feature points based on the SURF algorithm may include the following two steps. Step 1. Detection of local features of SURF.
- Step 2 Construction of the SURF descriptor.
- the structure of the SURF descriptor is divided into two parts: the main direction and the generation of the feature vector. These two parts are performed on the scale ⁇ where the feature point is located. For example, taking the feature point as the center, the calculated radius is 6 ⁇ inside the circle X and y.
- the Haar wavelet response coefficient in the direction is obtained by summing the coefficients in the x and y directions in a sector of 60 degrees, and then constructing a new vector, rotating the fan to traverse the entire circle, and selecting the direction of the longest vector as the main direction.
- the BBF method is used for rough matching to quickly converge the feature pairs, and the random sampling consistency algorithm (RANSAC) method is used to filter the matching. Point pairs get fine matching point pairs.
- RBSAC random sampling consistency algorithm
- the approximate nearest neighbor (BBF) algorithm is an improvement of the KD-Tree algorithm. Most of the time of the KD-Tree algorithm is used to query nodes, but only a small number of nodes satisfy the nearest neighbor condition, BBF can effectively solve the problem. .
- the BBF adopts a priority queue to make the search sequentially from the node to the queried node in the near and far distance.
- the KD-Tree It defines the number of leaf nodes in the KD-Tree, and defines the maximum number of searches, so it can quickly find the nearest neighbor and The next nearest neighbor point greatly improves the search efficiency. For a certain feature vector, first calculate the distance between all the feature vectors in the image to be searched and the vector, and then find the ratio of the nearest neighbor to the next nearest neighbor. If the ratio is less than a preset threshold, the nearest neighbor is considered to be better. match. For example, in a preferred embodiment of the present invention, the test results are statistically found, and the rough matching obtained when the threshold is set to 0.7 is correct.
- the relationship between two or more images formed by the three-dimensional scene can be completely transformed by the image transformation model. description. In the actual shooting process, when the captured 3D scene is far away (far larger than the focal length;), it can be considered that the perspective transformation model is satisfied.
- the correspondence between two images / ( , , can be represented by a 3x3 Planar Perspective Transform:
- H is a 3 * 3 full rank matrix, called the plane perspective transformation matrix, referred to as the perspective transformation matrix, also known as the homography matrix (Homography), according to the homogeneous
- H is a 3 * 3 full rank matrix, called the plane perspective transformation matrix, referred to as the perspective transformation matrix, also known as the homography matrix (Homography), according to the homogeneous
- Homography homography matrix
- the nature of the coordinates h33 can be normalized to 1, that is, the degree of freedom of H is 8, where hl l, hl2, h21, h22 are scaling and rotation factors, and hl3 and h23 are horizontal and vertical translation factors, h31, H32 is an affine transformation factor.
- the Random Sampling Consensus Algorithm (RANSAC) is the most widely used robust estimation method in the field of computer vision.
- the method uses the rough matching data based on the BBF method as input, and uses equations (3) and (4) as geometric constraints to further use RANSAC. Eliminate the wrong match and get the exact value of the matrix H.
- the RANSAC can be implemented by the following steps. Step 1. Extract 4 pairs from the coarse matching data and use it as the initial interior point calculation matrix 11. Step 2. Fit the initial matrix H with the remaining coarse data and calculate the sum of the distances t ⁇ . between them. Wherein, the distance t is defined by the Mahalanobis distance, If the sum of the distances ⁇ is greater than the selected threshold, it is discarded as an outer point. If it is smaller than the threshold, it is added to the inner point set, and then the homography matrix is updated by the least squares method.
- Step 3 Repeat 1 and 2 to select the largest set of inner point sets as the correct matching point pair.
- the model estimation result is the matrix 11 between the images.
- the horizontally polarized light reflected on the stagnant water surface is filtered by the vertically placed polarizer as shown in FIG. Therefore, in the picture acquired by the camera, the accumulated iced road surface exhibits a small light intensity in the picture acquired by the camera.
- the reflected light of the normal road surface is the same in the images acquired by the camera in the above two states. Therefore, by comparing the images acquired by the binocular camera, it can be determined that the portion where the brightness changes is the accumulated ice and ice road surface. This detects the accumulation of ice on the road surface.
- the method has the advantages of simple modification, low cost, no influence on the original system and integration into the original monitoring network.
- FIG. 9 is a schematic structural diagram of a road area water ice detecting device according to an embodiment of the present invention.
- the device is used to implement the road area water ice detecting method provided by the foregoing method embodiment.
- the device mainly includes: An acquisition module 10, a second acquisition module 20, and a determination module 30.
- the first acquisition module 10 is configured to acquire a first road image image captured by the first camera through the horizontal vibration plate on the road surface in the imaging range;
- the second acquisition module 20 is connected to the first acquisition module 10, and is configured to acquire The second camera passes the vertical polarizing plate to the second road image image captured on the road surface of the imaging range, the imaging plane of the image sensor (CCD) of the second camera is coplanar with the imaging plane of the CCD of the first camera, and the second camera The distance from the first camera does not exceed a predetermined value;
- the determination module 30 is connected to the second acquisition module 20, and is configured to determine whether the difference between the brightness of the first road image image and the brightness of the second road image image is greater than a preset The brightness range, if yes, determines that there is water or ice accretion on the road surface in the imaging range.
- FIG. 10 is a schematic structural diagram of a road area water ice detecting device according to a preferred embodiment of the present invention.
- the determining module of the device may further include a registration module 32 and a processing module 34.
- the registration module 32 is configured to perform image registration on the acquired first road image image and the second road image image by using the SURF algorithm.
- the processing module 34 is configured to register the first road surface after the registration module is registered.
- the image picture and the second road image picture perform image optimization processing operations.
- the image optimization processing operations that the processing module can take include, but are not limited to: differential, filtering, and binarization.
- the road area water ice detecting device can be inspected by the method described in the above method embodiment, and will not be described again.
- a schematic diagram of the installation structure of the road area water ice detecting device is as shown in FIG. 11 , and two cameras of the same type are mounted on the road surface, the camera parameters are set to be the same, and the front polarization is set.
- the sheets are rotated to be horizontal and vertical, respectively, to constitute a horizontal polarizing plate and a vertical polarizing plate.
- try to ensure that the distance between the two cameras is small enough (of course, preferably, the binocular camera in the above embodiment can be used), and the image planes coincide to ensure that the overlapping area of the images captured by the two cameras is as large as possible. .
- the horizontal polarizing plate and the vertical polarizing plate are respectively disposed in front of the first camera and the second camera, and the water content of the road area is detected, thereby solving the existing
- the road surface detection method in the technology is complicated, the detection range is small, and the implementation difficulty is large.
- the method adopts a method of respectively setting a horizontal polarizing plate and a vertical polarizing plate in front of the first camera and the second camera to detect the water content of the road area.
- the utility model solves the problems that the road surface detection method in the prior art is complicated, the detection range is small, and the implementation difficulty is large.
- the method can be directly applied to the networked surveillance camera, or directly used on the existing road test camera, and has the advantages of simple modification, low cost, no influence on the original system and integration into the original monitoring network, and stable operation.
- the detection range is wide, the detection is accurate, and it has great application value.
- the above modules or steps of the present invention can be implemented by a general-purpose computing device, which can be concentrated on a single computing device or distributed over a network composed of multiple computing devices. Alternatively, they may be implemented by program code executable by the computing device, such that they may be stored in the storage device by the computing device and, in some cases, may be different from the order herein.
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
一种路面积水积冰检测方法及装置。其中,该方法包括:获取第一摄像机通过水平偏振片对处于摄像范围的路面拍摄到的第一路面图像画面;获取第二摄像机通过垂直偏振片对处于摄像范围的路面拍摄到的第二路面图像画面,其中第二摄像机的图像传感器(CCD)的拍摄平面与第一摄像机的CCD的拍摄平面共面,且第二摄像机与第一摄像机之间的距离不超过预定值;判断第一路面图像画面的亮度与第二路面图像画面的亮度的差别是否大于预先设置的亮度范围,如果是,则判定处于摄像范围的路面存在积水或积冰。该方法及装置可以在现有路测摄像机上推广使用,并具有改装简单、改装费用低且对原系统无影响的优点。
Description
路面积水积冰检测方法及装置 技术领域 本发明涉及光通信领域, 具体而言, 涉及一种路面积水积冰检测方法及装置。 背景技术 相关技术的路况监测仪主要是利用红外线或微波在干燥路面、 冰雪路面的吸收率 不同的原理进行路面状态检测, 采用这种检测方式存在以下特点: 取样空间受限、 价 格昂贵、 不能沿路段大范围布设以及难以反映整条路段的真实状况。 目前, Hiroshi FUKUI Junichi TAKAGI等人提出使用图像处理的方法 (参见: Hiroshi Fukui, Junichi Takagi, Yoshiro Murata and MasashiTakeuchi . An Image Processing Method To Detect RoadSurface Condition Using Optical Spatial Frequency [J]. lEEEConference on Intelligent Transportation System, 1997: 1005 - 1009. )来检测路面状态, 该检测方法是利用监测图 像的最大全宽和空间频率分布区分不同的路面状况。 但是, 这种方法需要人工光源作 为辅助, 只能在夜间使用, 且光源和 CCD摄像机位置固定, 导致探测范围小, 难以推 广使用。 目前, Yuukou Horita和 Keiji Shibata等人还提出一种使用光学系统将垂直偏振图 像和水平偏振图像成像在摄像机的不同位置 (参见: Yuukou Horita, Keiji Shibata, Kei Maeda, Yuji Hayashi. Omni-directional Polarization Image Sensor Based on an Omni-directional Camera and a Polarization Filter [C]. Sixth IEEE International Conference on Advanced Video and Signal Based Surveillance, 2009:280-285. ), 但是, 该光学系统在 使用时需要特殊的光学设备 (例如, 该光学系统需要使用分光镜, 且对分光镜的要求 非常严格), 而且成像范围和效果也不如人意, 难以应用在现有监控场景下; 如果使用 半透半反射片将光线分为均等的两部分,则需要复杂的光学系统和双倍数量的摄像机, 也难于在大范围内普及使用。 由此可见, 相关技术中的路面积水积冰检测方法存在检测范围小、 检测装置复杂 导致的高成本或者检测方法复杂的问题, 而针对以上的各个缺陷, 尚未提出一种完善 的技术方案。
发明内容 本发明实施例提供了一种路面积水积冰检测方法及装置, 以至少解决上述问题之
根据本发明的一个实施例, 提供了一种路面积水积冰检测方法, 包括: 获取第一 摄像机通过水平偏振片对处于摄像范围的路面拍摄到的第一路面图像画面; 获取第二 摄像机通过垂直偏振片对处于摄像范围的路面拍摄到的第二路面图像画面, 其中, 第 二摄像机的图像传感器 (CCD) 的拍摄平面与第一摄像机的 CCD 的拍摄平面共面, 且第二摄像机与第一摄像机之间的距离不超过预定值; 判断第一路面图像画面的亮度 与第二路面图像画面的亮度的差别是否大于预先设置的亮度范围, 如果是, 则判定处 于摄像范围的路面存在积水或积冰。 根据本发明的另一实施例, 提供了一种路面积水积冰检测装置, 包括: 第一获取 模块, 设置为获取第一摄像机通过水平振片对处于摄像范围的路面拍摄到的第一路面 图像画面; 第二获取模块, 设置为获取第二摄像机通过垂直偏振片对处于摄像范围的 路面拍摄到的第二路面图像画面, 第二摄像机的图像传感器 (CCD) 的拍摄平面与第 一摄像机的 CCD的拍摄平面共面,且第二摄像机与第一摄像机之间的距离不超过预定 值; 判定模块, 设置为判断第一路面图像画面的亮度与第二路面图像画面的亮度的差 别是否大于预先设置的亮度范围, 如果是, 则判定处于摄像范围的路面存在积水或积 冰。 通过本发明的实施例, 采用在第一摄像机、 第二摄像机前分别设置水平偏振片、 垂直偏振片的方式, 对路面积水积冰进行检测, 解决了现有技术中的路面检测方法较 为复杂、 检测范围小、 实施难度大问题, 进而达到了工作稳定、 检测范围广、 应用价 值大的效果。 附图说明 此处所说明的附图用来提供对本发明的进一步理解, 构成本申请的一部分, 本发 明的示意性实施例及其说明用于解释本发明, 并不构成对本发明的不当限定。 在附图 中: 图 1是根据本发明实施例的路面积水积冰检测方法流程图; 图 2是根据本发明优选实施例的路面积水积冰检测流程图; 图 3是根据本发明实施例的反射光偏振度与入射角的关系图;
图 4是根据本发明实施例的积水积冰路面与正常路面的反射状况示意图; 图 5是根据本发明实施例的路面反射光总的状态示意图; 图 6是根据本发明优选实施例的双目图像配准流程图; 图 7是根据本发明实施例的水平偏振片下路面光线的变化情况示意图; 图 8是根据本发明实施例的垂直偏振片下路面光线的变化情况示意图; 图 9是根据本发明实施例的路面积水积冰检测装置结构示意图; 图 10是根据本发明优选实施例的路面积水积冰检测装置结构示意图; 图 11是根据本发明优选实施例的路面积水积冰检测装置的安装位置示意图。 具体实施方式 下文中将参考附图并结合实施例来详细说明本发明。 需要说明的是, 在不冲突的 情况下, 本申请中的实施例及实施例中的特征可以相互组合。 图 1是根据本发明实施例的路面积水积冰检测方法流程图, 如图 1所示, 该方法 主要包括以下步骤 (步骤 S102-步骤 S106)。 步骤 S102,获取第一摄像机通过水平偏振片对处于摄像范围的路面拍摄到的第一 路面图像画面。 在本发明实施例中, 第一摄像机可以根据透过水平偏振片的第一反射光线生成第 一路面图像画面, 然后将第一路面图像画面上传至指定的图像处理设备。 步骤 S104,获取第二摄像机通过垂直偏振片对处于摄像范围的路面拍摄到的第二 路面图像画面, 其中, 第二摄像机的图像传感器 (CCD) 的拍摄平面与第一摄像机的 CCD的拍摄平面共面, 且第二摄像机与第一摄像机之间的距离不超过预定值。 在本发明实施例中, 第二摄像机可以根据透过垂直偏振片的第二反射光线生成第 二路面图像画面,然后将第二路面图像画面上传至指定的图像处理设备,例如, PC机。 步骤 S106,判断第一路面图像画面的亮度与第二路面图像画面的亮度的差别是否 大于预先设置的亮度范围, 如果是, 则判定处于摄像范围的路面存在积水或积冰。
在本发明实施例中, 在获取第一摄像机通过水平振片对处于摄像范围的路面拍摄 到的第一路面图像画面和第二摄像机通过垂直偏振片对处于摄像范围的路面拍摄到的 第二路面图像画面之后,可以对第一路面图像画面合第二路面图像画面进行图像配准。 优选地,可以使用 SURF算法对第一路面图像画面和第二路面图像画面进行图像配准。 在本发明实施例的一个优选实施方式中, 为了使比较更为准确, 还可以先对第一 路面图像画面、 第二路面图像画面进行图像优化处理操作, 其中, 图像优化处理操作 主要有: 差分、 滤波及二值化。 然后再判断第一路面图像画面的亮度与第二路面图像 画面的亮度的差别是否大于预先设置的亮度范围, 如果是, 则可以判定处于摄像范围 的路面存在积水或积冰。 例如, 在实际应用中, 可以先获取经图像优化处理操作后的 第一路面图像画面的第一画面亮度值, 再获取经图像优化处理操作后的第二路面图像 画面的第二画面亮度值。 在得到第一画面亮度值和第二画面亮度值后, 可以进一步判 断第一画面亮度值和第二画面亮度值的差值是否大于预先设置的阈值, 如果是, 则判 定第一路面图像画面的亮度与第二路面图像画面的亮度的差别大于预先设置的亮度范 围, 否则, 判定第一路面图像画面的亮度与第二路面图像画面的亮度的差别小于预先 设置的亮度范围。 通过以上判断可以得知, 当第一路面图像画面的亮度与第二路面图 像画面的亮度的差别大于预先设置的亮度范围时, 当前的被监测路面存在积水或者积 冰, 反之, 则意味着当前的被监测路面是正常路面, 不存在积水或积冰。 需要说明的是, 该方法完全可以应用在夜间环境或光线特别暗的环境中, 当需要 检测夜间的路面 (或者, 光线特别暗的环境) 是否存在积水或积冰时, 可以单独设置 一个与待测路面成 53度入射角的光源, 该光源发出的光线虽然和自然光有所差别,但 是, 并不影响其被路面的积水或积冰反射后, 再通过水平偏振片和垂直偏振片被摄像 机摄取。 下面以双目相机为例, 对上述方法进行详细说明。 图 2是根据本发明优选实施例的路面积水积冰检测流程图。 在图 2中, 将垂直方 向和水平方向的偏振片分别前置于双目摄像机的前端, 例如, 可以将两个偏振片套置 在双目摄像机的前端, 同时采集图像并进行图像配准和后处理的测量, 利用光的偏振 特性, 由路面反射光的偏振光检测路面的积水积冰情况。 具体地, 可以在路面上方架 设两台同型号相机, 相机参数设定相同, 前置的偏振片分别水平和垂直放置, 相机间 距离尽可能小, 且像平面重合, 以保证两台相机所拍摄图像的重叠区域尽可能大。 优选地, 垂直方向和水平方向的偏振片可以分别前置于双目摄像机, 同时采集图 像并进行图像配准和后处理的测量。
公知地,人类的视觉系统能够通过色彩和亮度的形式感受到光的频率和强度特征, 但对于光的另一基本特征——偏振特性却无法直接感知。 因此, 传统的图像处理和理 解过程都是基于强度和频率域的信号处理, 只能对目标的轮廓、 类别等做一些初步的 分析和判断, 而不能辨别目标的材质和细节特征。 光作为一种电磁波, 具有横波的偏振特性, 反射面材料的导电特性和光滑程度的 差异将引起光偏振特性的差异。 一束自然光可以分解为两束振动方向相互垂直的、 等 幅的, 且不相干的线偏振光。 对入射某一平面的电矢量 而言, 可以分解为垂直于入 射面的分量 和平行于入射面的分量 ^。 根据 Fresnel反射模型, 反射电矢量的两个分量可以表示为:
( 1 ); sin ( - θ2 )
E
sin (3 + θη)
(2); 其中 , 表示入射角和折射角,其关系可以由 Fresnel公式算出,显见, 和 强度不相等, 且 > E" , 反射光表现出平行于路面的线偏振特性。 当入射角增加时,
π
E + Θ = Ε
分量急剧下降, 尤其是 2时, 入射角为布儒斯特角,
. E
¾ = 91
tan2 (/¾ + ft ) sin2 (ft + ft )
(4); 请参考图 3, 以水为例, 其折射率为 1.3333, 绘制出图 2所示的反射光偏振度一 入射角关系图, 显见, 对任意入射角, 其反射光均具有偏振度。 当入射角小于 20度或 者大于 80度时, 检测效果会受一定的干扰影响, 因此, 可以根据当前地理位置的太阳
光的入射角判断其对应的偏振度。 例如, 对于中国的地理位置, 正常太阳高度对应的 入射角具有较大的偏振度, 自然光入射角小于 20度的情况很少出现, 即使出现这种情 况也是正午的时候, 路面基本不会出现结冰等恶劣路况。在入射角大于 80度(多为晨 昏时刻或无自然光源的夜间时刻) 时, 可以使用人工光源作为辅助, 使光源对路面成 53度入射角 (即: 布儒斯特角) 实现检测。 请参考图 4, 如图 4所示, 当路面覆盖有积水或积冰的时候, 其表面光滑, 当有 光线照射时会发生镜面反射后得到上述的偏振光, 而正常路面发生漫反射。 当路面没 有积水或冰时, 表面粗糙, 入射角不固定, 故反射光具有各个方向的振动, 从统计特 性上不具有偏振性。 虽然老化路面趋于光滑, 可能会表现出一定的偏振度, 但相对积 水或积冰面反射光的偏振度可忽略不计, 所以在处理时可当作噪声滤除。 请参考图 5, 如图 5所示, 正常路面的反射为非偏振光, 积水积冰路面反射为水 平方向线偏振光, 叠加为部分偏振光进入摄像机, 由于其偏振方向为水平, 所以, 水 平偏振图像中亮度较高, 而垂直偏振图像中亮度较低。 正常路面的亮度在两种偏振状 态下几乎没有区别, 亮度较低的部分可以判断为有低摩擦系数的介质存在。 请参考图 6, 上述方法中可以使用 SURF算法获得相机 A和相机 B原始图像的特 征点对,再采用近似最近邻算法 (BBF)方法对特征点对进行粗匹配,并构造基于透视变 换矩阵 H的图像映射模型, 然后利用 RANSAC进一步剔除错误的匹配, 并使用最小 二乘法求出图像之间精确的 H矩阵, 从而得到配准后的图像。 在实际应用中, 经常由于高速公路等复杂场景的存在导致获取的图像质量较差, 并且由于图像间亮度变化较大、 尺寸变化较大、 旋转较大及运动物体干扰等因素, 造 成了路面图像配准的困难。 此时, 就要求路面图像配准方法对于这些不利因素具有较 强的鲁棒性, 提取特征 (SIFT) 算法由于其对较大旋转、 尺寸缩放、 色差、 视觉以及 光照变化所体现出来的鲁棒性, 特别适用于复杂图像配准。 但是, 同时也存在着数据 量大、 计算耗时长等问题使其对于图像配准的实时性产生不利影响。 在本发明实施例的一个优选实施方式中,将 SURF算法引入路面图像配准的过程。
SURF在理念上与 SIFT相似, 注重梯度信息的空间分布, 并且继承了 SIFT算法对旋 转、 尺度缩放、 亮度变化保持不变性, 以及对视角变化、 图像模糊、 噪声的鲁棒性的 优点。 最为重要的是, SURF 算法在特征点检测、 描述向量中都充分利用方框滤波器 和积分图像来加快计算速度, 并减少了局部图像描述向量的维数, 极大地加快了对特 征点的检测和描述向量的产生, 能够显著地提高图像配准方法的速度。 具体地, 基于 SURF 算法的特征点的提取可以包括以下 2个步骤。
步骤 1、 SURF局部特征的检测。 SURF 特征点的检测可以基于 Hessian矩阵实现, 其中, Hessian矩阵 Η(χ,σ)在尺度为 σ的 x 点是被定义为:
( 1 ); 其中, 在式 (1 ) 中: ·χ, σ)是高斯滤波波二阶导 ^^(σ)同 / = (xj)卷积的结 au
果, 其中, ξ{ ) = 1^-Ιβ {^^ Lxy (x, a) , ^ (χ, σ)与 χ, σ)相类似; 步骤 2、 SURF描述符的构造。 SURF描述符的构造分为主方向分配和生成特征向 量两部分, 这两部分都是在特征点所在的尺度 σ上进行的, 例如, 以特征点为圆心, 计算半径为 6σ圆内 X和 y方向上的 Haar小波响应系数, 在 60度的扇形区域内求 x 和 y方向上的系数之和, 进而构建一个新向量, 转动扇形遍历整个圆, 选择最长向量 的方向为主方向。 需要说明的是, 由于 SURF算法的鲁棒性能极大地降低对双目相机的控制性能要 求以及提高配准的成功率, 所以, 即使在存在天气恶劣、 车流大的复杂环境中, 也能 完成路面图像配准任务。 SURF算法的快速性则有利于路面图像配准的实时性的实现, 这正是动态更新有无路面积水或积冰的关键。 例如,在本发明实施例中,在用 SURF算法获得相机 A与相机 B的图像特征点后, 采用 BBF方法进行粗匹配以快速收敛特征对,并结合随机采样一致性算法 (RANSAC ) 方法过滤匹配点对获取精匹配点对。 这里需要说明, 近似最近邻 (BBF) 算法是对 KD-Tree算法的改进, KD-Tree算法 的大部分时间都用来查询节点, 但只有小部分节点满足最近邻条件, BBF可以有效解 决该问题。 BBF采用一个优先级队列使搜索依次从节点与被查询节点距离由近及远的 顺序进行, 它限定 KD-Tree中叶子节点数, 限定了搜索的最大次数, 所以能够快速的 找到最近邻点和次近邻点, 极大提高了搜索效率。 而对于某一特征向量, 首先计算待 搜索图像中所有的特征向量与该向量的距离, 然后求最近邻与次近邻的比值, 如果比 值小于预先设定的阈值, 认为该最近邻是较好的匹配。 例如, 在本发明实施例的一个 优选实施方式中, 对测试结果进行统计, 发现将阈值设为 0. 7时得到的粗匹配的正确
对于图像变换模型是指两幅二维图像之间的坐标变换关系而言, 在某种约束的摄 像机运动条件下, 三维场景形成的两幅或多幅图像之间的关系可以完全由图像变换模 型描述。 在实际拍摄过程中, 当所拍摄的三维场景很远时 (远大于焦距;), 都可以认为 近似满足透视变换模型。例如, 两幅图像 /( , 之间的对应关系可以由一个 3x3的平面透视变换矩阵 (Planar Perspective Transform)来表示:
( 2); 其中, 在式 (2 ) 中, H是一个 3 *3的满秩矩阵, 称为平面透视变换矩阵, 简称为 透视变换矩阵, 又称单应性矩阵 (Homography) , 根据齐次坐标的性质 h33可以归一化 为 1, 即 H的自由度为 8, 其中 hl l、 hl2、 h21、 h22是缩放、 旋转因子, hl3、 h23分 别是水平、 竖直方向的平移因子, h31、 h32是仿射变换因子。 进一步, 由式 (2)可以得到两个方程:
ιχ + y + - ixy - yy' - / = o ( 4 ) ; 其中, 矩阵 H共有 8个未知参数, 使用线性方法求解, 至少需要 4对特征点的坐 标, 联立 8个方程, 使用 SVD分解, 可以求得矩阵 H。 基于 BBF方法的粗匹配含有 误匹配,直接用这些点难以求得精确的矩阵 H,文中使用 RANSAC算法剔除匹配质量 较差的点得到精确解。 随机采样一致性算法(RANSAC )是计算机视觉领域中应用最广的稳健估计方法, 该方法使用基于 BBF方法的粗匹配数据作为输入, 以式 (3 ), (4 )作为几何约束, 利 用 RANSAC进一步剔除错误的匹配, 得到矩阵 H的精确值。 具体地, RANSAC可以 通过以下步骤实现。 步骤 1、 从粗匹配数据中抽取 4对, 并用其作为初始内点计算矩阵11。 步骤 2、 用剩下的粗配数据来拟合初始矩阵 H, 并计算它们之间的距离 t^.的和。 其中, 以马氏距离定义距离 t ,
如果距离^的和大于选取的阈值, 则作为外点舍弃, 如果小于阈值, 则添加到内 点集中, 再运用最小二乘法更新单应矩阵。 如此反复迭代, 直到内点集不再扩充。 步骤 3、重复 1和 2, 选取内点集最大的一组作为正确的匹配点对, 此时模型估计 结果就是图像间的矩阵11。 采用上述实施例提供的方法在第一摄像机、 第二摄像机前分别设置水平偏振片、 垂直偏振片的方式, 实现对路面积水积冰进行检测, 如图 7所示, 积水积冰路面反射 的水平偏振光通过水平放置的偏振片(即水平偏振镜), 在摄像机获取的画面中, 积水 积冰路面表现出明显的亮度。 积水积冰路面反射的水平偏振光被垂直放置的偏振片滤 除, 如图 8所示。 因此, 在摄像机获取的画面中, 此时积水积冰路面在摄像机获取的 画面中表现出很小的光强。 通常, 正常路面的反射光在上述两种状态下, 在摄像机获 取的画面中相同, 所以, 通过比较双目摄像机获取的画面, 就可以判定存在亮度变化 的部分则为积水积冰路面, 由此来检测路面的积水积冰情况。 该方法具有改装简便、 费用低, 对原系统无影响并可融入原监控网络的优点。 图 9是根据本发明实施例的路面积水积冰检测装置结构示意图, 该装置用于实施 上述方法实施例提供的路面积水积冰检测方法, 如图 9所示, 该装置主要包括: 第一 获取模块 10、 第二获取模块 20以及判定模块 30。 其中, 第一获取模块 10, 设置为获 取第一摄像机通过水平振片对处于摄像范围的路面拍摄到的第一路面图像画面; 第二 获取模块 20,连接至第一获取模块 10, 设置为获取第二摄像机通过垂直偏振片对处于 摄像范围的路面拍摄到的第二路面图像画面, 第二摄像机的图像传感器 (CCD) 的拍 摄平面与第一摄像机的 CCD的拍摄平面共面,且第二摄像机与第一摄像机之间的距离 不超过预定值; 判定模块 30, 连接至第二获取模块 20, 设置为判断第一路面图像画面 的亮度与第二路面图像画面的亮度的差别是否大于预先设置的亮度范围, 如果是, 则 判定处于摄像范围的路面存在积水或积冰。 图 10是根据本发明优选实施例的路面积水积冰检测装置结构示意图, 如图 10所 示, 该装置的判断模块还可以包括配准模块 32和处理模块 34。 其中, 配准模块 32, 设置为使用 SURF算法对获取到的第一路面图像画面和第二路面图像画面进行图像配 准; 处理模块 34, 设置为对经配准模块配准后的第一路面图像画面和第二路面图像画 面进行图像优化处理操作。 其中, 处理模块可以采取的图像优化处理操作包括但不限 于: 差分、 滤波及二值化。
另外,该路面积水积冰检测装置可以采用上述方法实施例所描述的方法进行检 具体不再赘述。 在本发明实施例的一个优选实施方式中, 该路面积水积冰检测装置的安装结构示 意图如图 11所示, 在路面上方架设两台同型号相机, 相机参数设定相同, 前置的偏振 片分别旋转至水平、 垂直以构成水平偏振片、 垂直偏振片。 同时, 尽量保证两个相机 之间的距离足够小(当然, 优选地, 可以采用上述实施例中的双目相机), 且像平面重 合, 以保证两台相机所拍摄图像的重叠区域尽可能大。 采用上述实施例提供的路面积水积冰检测装置, 可以采用在第一摄像机、 第二摄 像机前分别设置水平偏振片、 垂直偏振片的方式, 对路面积水积冰进行检测, 解决了 现有技术中的路面检测方法较为复杂、 检测范围小、 实施难度大问题。 从以上的描述中, 可以看出, 本发明实现了如下技术效果: 该方法采用在第一摄 像机、 第二摄像机前分别设置水平偏振片、 垂直偏振片的方式, 对路面积水积冰进行 检测, 解决了现有技术中的路面检测方法较为复杂、 检测范围小、 实施难度大问题。 同时, 该方法可以直接应用在联网监控的摄像机上, 或者直接在现有路测摄像机上推 广使用, 具有改装简便、 费用低, 对原系统无影响并可融入原监控网络的优点, 并且 工作稳定、 检测范围广、 检测准确, 具有极大的应用价值。 显然, 本领域的技术人员应该明白, 上述的本发明的各模块或各步骤可以用通用 的计算装置来实现, 它们可以集中在单个的计算装置上, 或者分布在多个计算装置所 组成的网络上, 可选地, 它们可以用计算装置可执行的程序代码来实现, 从而, 可以 将它们存储在存储装置中由计算装置来执行, 并且在某些情况下, 可以以不同于此处 的顺序执行所示出或描述的步骤, 或者将它们分别制作成各个集成电路模块, 或者将 它们中的多个模块或步骤制作成单个集成电路模块来实现。 这样, 本发明不限制于任 何特定的硬件和软件结合。 以上所述仅为本发明的优选实施例而已, 并不用于限制本发明, 对于本领域的技 术人员来说, 本发明可以有各种更改和变化。 凡在本发明的精神和原则之内, 所作的 任何修改、 等同替换、 改进等, 均应包含在本发明的保护范围之内。
Claims
1. 一种路面积水积冰检测方法, 包括:
获取第一摄像机通过水平偏振片对处于摄像范围的路面拍摄到的第一路面 图像画面;
获取第二摄像机通过垂直偏振片对所述处于摄像范围的路面拍摄到的第二 路面图像画面, 其中, 所述第二摄像机的图像传感器 CCD 的拍摄平面与所述 第一摄像机的所述 CCD 的拍摄平面共面, 且所述第二摄像机与所述第一摄像 机之间的距离不超过预定值;
判断所述第一路面图像画面的亮度与所述第二路面图像画面的亮度的差别 是否大于预先设置的亮度范围, 如果是, 则判定所述处于摄像范围的路面存在 积水或积冰。
2. 根据权利要求 1所述的方法, 其中, 判断所述第一路面图像画面的亮度与所述 第二路面图像画面的亮度的差别是否大于预先设置的亮度范围之前, 包括: 对所述第一路面图像画面和所述第二路面图像画面进行图像优化处理操 作。
3. 根据权利要求 2 所述的方法, 其中, 判断所述第一路面图像画面的亮度与所述 第二路面图像画面的亮度的差别是否大于所述预先设置的亮度范围, 包括: 获取经所述图像优化处理操作后的所述第一路面图像画面的第一画面亮度 值;
获取经所述图像优化处理操作后的所述第二路面图像画面的第二画面亮度 值;
判断所述第一画面亮度值和所述第二画面亮度值的差值是否大于预先设置 的阈值, 如果是, 则判定所述第一路面图像画面的亮度与所述第二路面图像画 面的亮度的差别大于所述预先设置的亮度范围, 否则, 判定所述第一路面图像 画面的亮度与所述第二路面图像画面的亮度的差别小于所述预先设置的亮度范 围。
4. 根据权利要求 2所述的方法, 其中, 所述图像优化处理操作包括: 差分、 滤波 及二值化。
根据权利要求 1至 4中任一项所述的方法, 其中, 在获取第一摄像机通过水平 振片对处于摄像范围的路面拍摄到的第一路面图像画面和第二摄像机通过垂直 偏振片对所述处于摄像范围的路面拍摄到的第二路面图像画面之后, 所述方法 还包括: 对所述第一路面图像画面和所述第二路面图像画面进行图像配准。 根据权利要求 5所述的方法, 其中, 使用快速提取特征 SURF算法对所述第一 路面图像画面和所述第二路面图像画面进行图像配准。 根据权利要求 6所述的方法, 其中, 检测夜间的路面是否存在积水或积冰时, 所述方法还包括:
设置与待测路面成 53度入射角的光源。 一种路面积水积冰检测装置, 包括:
第一获取模块, 设置为获取第一摄像机通过水平振片对处于摄像范围的路 面拍摄到的第一路面图像画面;
第二获取模块, 设置为获取第二摄像机通过垂直偏振片对所述处于摄像范 围的路面拍摄到的第二路面图像画面, 所述第二摄像机的图像传感器 CCD 的 拍摄平面与所述第一摄像机的所述 CCD 的拍摄平面共面, 且所述第二摄像机 与所述第一摄像机之间的距离不超过预定值;
判定模块, 设置为判断所述第一路面图像画面的亮度与所述第二路面图像 画面的亮度的差别是否大于预先设置的亮度范围, 如果是, 则判定所述处于摄 像范围的路面存在积水或积冰。 根据权利要求 8所述的装置, 其中, 所述判断模块包括:
配准模块, 设置为使用 SURF算法对获取到的所述第一路面图像画面和所 述第二路面图像画面进行图像配准。 根据权利要求 9所述的装置, 其中, 所述判断模块还包括:
处理模块, 设置为对经所述配准模块配准后的所述第一路面图像画面和所 述第二路面图像画面进行图像优化处理操作。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201110208960.3A CN102901489B (zh) | 2011-07-25 | 2011-07-25 | 路面积水积冰检测方法及装置 |
CN201110208960.3 | 2011-07-25 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2013013563A1 true WO2013013563A1 (zh) | 2013-01-31 |
Family
ID=47573855
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2012/078042 WO2013013563A1 (zh) | 2011-07-25 | 2012-07-02 | 路面积水积冰检测方法及装置 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN102901489B (zh) |
WO (1) | WO2013013563A1 (zh) |
Families Citing this family (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2016086380A1 (zh) * | 2014-12-04 | 2016-06-09 | 深圳市大疆创新科技有限公司 | 一种物体检测方法、装置及遥控移动设备、飞行器 |
JP6361631B2 (ja) * | 2015-10-29 | 2018-07-25 | Smk株式会社 | 車載センサ、車両用灯具及び車両 |
CN105809131B (zh) * | 2016-03-08 | 2019-10-15 | 宁波裕兰信息科技有限公司 | 一种基于图像处理技术进行车位积水检测的方法及系统 |
US10183677B2 (en) * | 2016-09-20 | 2019-01-22 | Ford Global Technologies, Llc | Ice and snow detection systems and methods |
CN108701356A (zh) * | 2017-06-29 | 2018-10-23 | 深圳市大疆创新科技有限公司 | 一种检测方法、检测设备以及飞行器 |
CN107909070A (zh) * | 2017-11-24 | 2018-04-13 | 天津英田视讯科技有限公司 | 一种道路积水检测的方法 |
CN108288063A (zh) * | 2018-01-09 | 2018-07-17 | 交通运输部公路科学研究所 | 路面的气象状态确定方法、装置及系统 |
JPWO2020202695A1 (zh) * | 2019-04-03 | 2020-10-08 | ||
CN110146897A (zh) * | 2019-05-24 | 2019-08-20 | 北京海益同展信息科技有限公司 | 积水检测装置及机器人 |
CN110411366B (zh) * | 2019-07-31 | 2021-01-05 | 北京领骏科技有限公司 | 一种道路积水深度的检测方法及电子设备 |
CN111308494B (zh) * | 2019-12-11 | 2022-05-24 | 中国科学院长春光学精密机械与物理研究所 | 一种物体表面结冰检测系统 |
CN114202573B (zh) * | 2022-02-18 | 2022-04-29 | 南京路健通工程技术有限公司 | 一种旅游区道路用提示方法及装置 |
CN114808823B (zh) * | 2022-04-28 | 2024-06-14 | 湖北佰思图汽车有限公司 | 一种清扫车快速清理路面积液的智能控制方法及系统 |
CN115356271A (zh) * | 2022-10-20 | 2022-11-18 | 长春理工大学 | 基于偏振探测的金属物表面冰雪探测装置及方法 |
CN116311028B (zh) * | 2023-01-09 | 2023-11-24 | 无锡市德宁节能科技有限公司 | 一种基于物联网的护栏控制方法和系统 |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH02161337A (ja) * | 1988-12-14 | 1990-06-21 | Nagoya Denki Kogyo Kk | 路面状態検知装置 |
JP2004301708A (ja) * | 2003-03-31 | 2004-10-28 | Nagoya Electric Works Co Ltd | 車両用路面状態検出装置、車両用路面状態検出方法および車両用路面状態検出装置の制御プログラム |
JP2005043240A (ja) * | 2003-07-23 | 2005-02-17 | Mitsubishi Electric Corp | 路面状態検出センサ |
JP2005308437A (ja) * | 2004-04-19 | 2005-11-04 | Quest Engineer:Kk | 雪検知システム |
JP2006058122A (ja) * | 2004-08-19 | 2006-03-02 | Nagoya Electric Works Co Ltd | 路面状態判別方法およびその装置 |
JP2007064888A (ja) * | 2005-09-01 | 2007-03-15 | Tokai Rika Co Ltd | 路面状態検出装置 |
JP2009025198A (ja) * | 2007-07-20 | 2009-02-05 | Denso It Laboratory Inc | 路面状態検出装置および路面状態検出方法 |
CN101610357A (zh) * | 2008-06-18 | 2009-12-23 | 株式会社理光 | 摄像装置及路面状态判别方法 |
JP2011038827A (ja) * | 2009-08-07 | 2011-02-24 | Kitami Institute Of Technology | 路面状態検出方法および路面状態検出装置 |
-
2011
- 2011-07-25 CN CN201110208960.3A patent/CN102901489B/zh not_active Expired - Fee Related
-
2012
- 2012-07-02 WO PCT/CN2012/078042 patent/WO2013013563A1/zh active Application Filing
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH02161337A (ja) * | 1988-12-14 | 1990-06-21 | Nagoya Denki Kogyo Kk | 路面状態検知装置 |
JP2004301708A (ja) * | 2003-03-31 | 2004-10-28 | Nagoya Electric Works Co Ltd | 車両用路面状態検出装置、車両用路面状態検出方法および車両用路面状態検出装置の制御プログラム |
JP2005043240A (ja) * | 2003-07-23 | 2005-02-17 | Mitsubishi Electric Corp | 路面状態検出センサ |
JP2005308437A (ja) * | 2004-04-19 | 2005-11-04 | Quest Engineer:Kk | 雪検知システム |
JP2006058122A (ja) * | 2004-08-19 | 2006-03-02 | Nagoya Electric Works Co Ltd | 路面状態判別方法およびその装置 |
JP2007064888A (ja) * | 2005-09-01 | 2007-03-15 | Tokai Rika Co Ltd | 路面状態検出装置 |
JP2009025198A (ja) * | 2007-07-20 | 2009-02-05 | Denso It Laboratory Inc | 路面状態検出装置および路面状態検出方法 |
CN101610357A (zh) * | 2008-06-18 | 2009-12-23 | 株式会社理光 | 摄像装置及路面状态判别方法 |
JP2011038827A (ja) * | 2009-08-07 | 2011-02-24 | Kitami Institute Of Technology | 路面状態検出方法および路面状態検出装置 |
Also Published As
Publication number | Publication date |
---|---|
CN102901489A (zh) | 2013-01-30 |
CN102901489B (zh) | 2016-09-07 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2013013563A1 (zh) | 路面积水积冰检测方法及装置 | |
JP7330376B2 (ja) | 偏光によるセンサシステム及び撮像システムの増強のための方法 | |
CN107635129B (zh) | 一种三维三目摄像装置及深度融合方法 | |
JP3955616B2 (ja) | 画像処理方法、画像処理装置及び画像処理プログラム | |
JP4563513B2 (ja) | 画像処理装置及び擬似立体画像生成装置 | |
CN102646272A (zh) | 基于局部方差和加权相结合的小波气象卫星云图融合方法 | |
Krishna et al. | Passive polarimetric imagery-based material classification robust to illumination source position and viewpoint | |
CN110425983B (zh) | 一种基于偏振多光谱的单目视觉三维重建测距方法 | |
EP2561482A1 (en) | Shape and photometric invariants recovery from polarisation images | |
Liu et al. | High quality depth map estimation of object surface from light-field images | |
CN108548603A (zh) | 一种非共轴四通道偏振成像方法及系统 | |
Simon et al. | A simple and effective method to detect orthogonal vanishing points in uncalibrated images of man-made environments | |
CN109325912A (zh) | 基于偏振光光场的反光分离方法及标定拼合系统 | |
Hajebi et al. | Structure from infrared stereo images | |
CN109490867A (zh) | 水面目标偏振遥感探测能力评价方法 | |
Yang et al. | A method of removing reflected highlight on images based on polarimetric imaging | |
Raisanen et al. | Simulation of practical single-pixel wire-grid polarizers for superpixel stokes vector imaging arrays | |
CN111343368B (zh) | 基于偏振的散射介质深度恢复方法及装置 | |
CN107292859B (zh) | 基于光学相关器的混沌介质偏振图像获取方法 | |
Gribben et al. | Structured light 3D measurement of reflective objects using multiple DMD projectors | |
Zhao et al. | 3D reconstruction and dehazing with polarization vision | |
Atkinson | Two-source surface reconstruction using polarisation | |
Bartlett et al. | Anomaly detection with varied ground sample distance utilizing spectropolarimetric imagery collected using a liquid crystal tunable filter | |
CN118864300A (zh) | 基于全偏振信息处理的空间目标偏振去雾装置及方法 | |
Ren et al. | Image registration method for full-Stokes-vector division-of-aperture polarimetric camera |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 12817480 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 12817480 Country of ref document: EP Kind code of ref document: A1 |