CN107909010B - Road obstacle detection method and device - Google Patents
Road obstacle detection method and device Download PDFInfo
- Publication number
- CN107909010B CN107909010B CN201711031619.9A CN201711031619A CN107909010B CN 107909010 B CN107909010 B CN 107909010B CN 201711031619 A CN201711031619 A CN 201711031619A CN 107909010 B CN107909010 B CN 107909010B
- Authority
- CN
- China
- Prior art keywords
- image
- obstacle
- road
- area
- detected
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/245—Aligning, centring, orientation detection or correction of the image by locating a pattern; Special marks for positioning
-
- 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
- G06V10/20—Image preprocessing
- G06V10/34—Smoothing or thinning of the pattern; Morphological operations; Skeletonisation
Abstract
The invention discloses a road obstacle detection method and a device, and the method comprises the following steps: acquiring a road image through binocular equipment, and processing the road image to obtain a disparity map of the road image; performing image segmentation based on road surface information on the disparity map to obtain a binary map with an obstacle region as a foreground and other regions as backgrounds; performing morphological operation on the binary image to obtain a template of the area to be detected of the suspected obstacle; carrying out information fusion on the template of the area to be detected and the monocular image to obtain a gray level image of the area to be detected, wherein the monocular image is one of the road images; and inputting the gray level image to a preset obstacle detection model to judge obstacles on the road image, and confirming the suspected obstacles according to the judgment result. According to the invention, the respective advantages of binocular parallax information and monocular image information are fused, and target identification and target classification based on machine learning are carried out, so that the barrier detection result can be rapidly and accurately obtained.
Description
Technical Field
The invention relates to the technical field of digital image processing, in particular to a road obstacle detection method and device.
Background
With the development of sensor technology and machine vision technology, more and more technologies for target detection based on a binocular parallax algorithm are emerging, and obstacle information obtained by analyzing parallax images is applied to the fields of robots and intelligent automobiles.
However, in the prior art, distance detection is mainly performed through binocular parallax, a special-shaped obstacle cannot be stably identified in the aspect of target detection, and a target detection result is poor in robustness and is prone to being interfered by the environment to cause detection errors.
Therefore, the binocular parallax information and the monocular image information are fused, the binocular parallax information is processed, and the monocular image information is fused, so that the problem that in the prior art, the special-shaped obstacles cannot be well identified, and the identification result is poor in robustness is solved.
Disclosure of Invention
The invention mainly aims to disclose a road obstacle detection method and a road obstacle detection device, and aims to solve the problem that in the prior art, a special-shaped obstacle cannot be well detected, and the robustness of a detection result is poor.
In order to achieve the above purpose, according to one aspect of the present invention, a road obstacle detection method is disclosed, and the following technical solutions are adopted:
a method of road obstacle detection comprising: acquiring a road image through binocular equipment, and processing the road image to obtain a disparity map of the road image; carrying out image segmentation based on road surface information on the disparity map to obtain a binary map with an obstacle region as a foreground and other regions as backgrounds; performing morphological operation on the binary image to obtain a template of a to-be-detected area of the suspected obstacle; performing information fusion on the template of the area to be detected and a monocular image to obtain a gray level image of the area to be detected, wherein the monocular image is one of the road images; and inputting the gray level image to a preset obstacle detection model to judge obstacles on the road image, and confirming the suspected obstacles according to a judgment result.
Further, the inputting the gray image to a preset obstacle detection model to judge an obstacle of the road image, and the confirming the suspected obstacle according to a judgment result includes: performing feature analysis on the gray level image through an obstacle recognition model in the preset obstacle detection model, and performing target recognition and marking on the suspected obstacle according to an analysis result to obtain a marked target; and classifying the marked targets through an obstacle classification model in the preset obstacle detection model.
Further, the method for acquiring the preset obstacle detection model includes: acquiring a plurality of groups of road images as samples through binocular equipment, and generating corresponding disparity maps by taking disparity information of each group of images as known input quantity; carrying out image segmentation based on road surface information on the disparity map to obtain a binary map with a suspected obstacle area as a foreground and other areas as backgrounds; performing morphological operation on the binary image to obtain a template of a to-be-detected area of the suspected obstacle; performing information fusion on the template of the area to be detected and a monocular image, performing connected domain statistics on the template of the area to be detected to obtain a minimum external rectangle of each connected region, and obtaining a gray level image of the area to be detected from the position of the rectangle on the monocular image, wherein the monocular image is one of the road images; screening and concentrating the obstacles in the gray level image to form an obstacle identification model; and performing machine training learning on the preset obstacle recognition sample set to obtain the obstacle classification model.
Further, the image segmentation based on the road surface information on the disparity map to obtain a binary map including a suspected obstacle area as a foreground and other areas as a background includes: obtaining a road surface model without obstacles in an imaging range through the road surface information; and inputting the road surface model to carry out image segmentation on the parallax image to obtain a binary image with a suspected obstacle area as a foreground and other areas as a background.
Further, the obtaining of the road image through the binocular device and the processing of the road image to obtain the disparity map of the road image includes: and acquiring parallax information of the road image, displaying the parallax information in an image matrix form, representing different parallax values by different colors, and visually displaying the digitized parallax information in an image form to generate the parallax map.
According to another aspect of the present invention, there is provided a road obstacle detection device, and the following technical solution is adopted:
this road obstacle detection device includes: the processing module is used for acquiring a road image through binocular equipment and processing the road image to obtain a disparity map of the road image; the first segmentation module is used for carrying out image segmentation based on road surface information on the parallax map to obtain a binary map with an obstacle area as a foreground and other areas as backgrounds; the first operation module is used for carrying out morphological operation on the binary image to obtain a template of the area to be detected of the suspected obstacle; the first fusion module is used for carrying out information fusion on the template of the area to be detected and a monocular image to obtain a gray level image of the area to be detected, wherein the monocular image is one of the road images; and the judging module is used for inputting the gray level image to a preset obstacle detection model to judge obstacles on the road image and confirming the suspected obstacles according to a judging result.
Further, the judging module comprises: the analysis module is used for performing feature analysis on the gray level image through an obstacle recognition model in the preset obstacle detection model, performing target recognition on the suspected obstacle according to an analysis result, and marking to obtain a marked target; and the classification module is used for classifying the marked targets through an obstacle classification model in the preset obstacle detection model.
Further, the classification module includes: the generating module is used for acquiring a plurality of groups of road images as samples through binocular equipment and generating corresponding disparity maps by taking the disparity information of each group of images as known input quantity; the second segmentation module is used for carrying out image segmentation based on road surface information on the parallax map to obtain a binary map with a suspected obstacle area as a foreground and other areas as a background; the second operation module is used for carrying out morphological operation on the binary image to obtain a template of the area to be detected of the suspected obstacle; the second fusion module is used for performing information fusion on the template of the area to be detected and the monocular image, performing connected domain statistics on the template of the area to be detected to obtain a minimum external rectangle of each connected region, and obtaining a gray level image of the area to be detected from the position of the rectangle on the monocular image, wherein the monocular image is one of the road images; the screening module is used for screening and concentrating the obstacles in the gray level image to form the obstacle identification model; and the training module is used for performing machine training learning on the preset obstacle recognition sample set to obtain the obstacle classification model.
Further, the first segmentation module comprises: the first acquisition module is used for acquiring a road surface model without obstacles in an imaging range through the road surface information; and the input module is used for inputting the road surface model to carry out image segmentation on the parallax map so as to obtain a binary map with a suspected obstacle area as a foreground and other areas as a background.
Further, the processing module comprises: and the second acquisition module is used for acquiring the parallax information of the road image, displaying the parallax information in an image matrix form, representing different parallax values by using different colors, and visually displaying the digitized parallax information in an image form to generate the parallax map.
On the basis of image segmentation of a binocular vision parallax map based on a road model, monocular information is fused, and then target identification and target classification based on machine learning are carried out, so that rapid and accurate obstacle detection is realized. The scheme of the invention particularly focuses on the interference of the lane line to the obstacle, so that the obstacle detection result is more robust.
Drawings
In order to more clearly illustrate the embodiments of the present application or technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present invention, and other drawings can be obtained by those skilled in the art according to the drawings.
Fig. 1 is a flowchart of a road obstacle detection method according to an embodiment of the present invention;
FIG. 2 is a left side view of a road image according to an embodiment of the present invention;
FIG. 3 is a right side view of a road image according to an embodiment of the present invention;
fig. 4 is a gray scale representation of a disparity map corresponding to disparity information according to an embodiment of the present invention;
FIG. 5 is a gray scale rendering of a pavement model according to an embodiment of the present invention;
fig. 6 is a gray scale representation of a binary image of a disparity map and a road surface model segmentation result according to an embodiment of the present invention;
fig. 7 is a template of a region to be detected after binary image morphological operation according to an embodiment of the present invention;
fig. 8 is a schematic diagram illustrating a fusion result of the template of the region to be detected and the monocular image according to the embodiment of the present invention;
fig. 9 is a schematic diagram illustrating a road obstacle recognition result according to an embodiment of the present invention;
fig. 10 is a structural view of a road obstacle detecting apparatus according to an embodiment of the present invention.
Detailed Description
The embodiments of the invention will be described in detail below with reference to the drawings, but the invention can be implemented in many different ways as defined and covered by the claims.
Fig. 1 is a flowchart of a road obstacle detection method according to an embodiment of the present invention.
Referring to fig. 1, a road obstacle detection method includes:
s101: acquiring a road image through binocular equipment, and processing the road image to obtain a disparity map of the road image;
s103: carrying out image segmentation based on road surface information on the disparity map to obtain a binary map with an obstacle region as a foreground and other regions as backgrounds;
s105: performing morphological operation on the binary image to obtain a template of a to-be-detected area of the suspected obstacle;
s107: performing information fusion on the template of the area to be detected and a monocular image to obtain a gray level image of the area to be detected, wherein the monocular image is one of the road images;
s109: and inputting the gray level image to a preset obstacle detection model to judge obstacles on the road image, and confirming the suspected obstacles according to a judgment result.
In step S101, a road image is acquired through binocular equipment, and the road image is processed to obtain a disparity map of the road image.
Specifically, a road image may be acquired by a binocular camera, as shown in fig. 2 to 3, left and right views of the road image acquired by the binocular camera are used to acquire disparity information of the left and right views, the disparity information is given in the form of an image pixel matrix as a known input quantity, different disparity values are represented by different colors, and the digitized disparity information is visually displayed in the form of an image to obtain a disparity map corresponding to the disparity information, as shown in fig. 4, the disparity map is an image corresponding to the different disparity values by colors, and fig. 4 is a gray scale map corresponding to the disparity values with colors.
In step S103, image segmentation based on road surface information is performed on the disparity map, and a binary map including an obstacle region as a foreground and other regions as a background is obtained.
More specifically, the disparity map is subjected to image segmentation based on road surface information. The road surface model is obtained from road surface information, and as shown in fig. 5, the road surface model is parallax information generated in a case where there is no near obstacle in the binocular vision imaging range. Corresponding to the disparity map in step S101, at the same position, if the disparity information of the disparity map does not match the road surface disparity information, it is considered to be a possible obstacle region here. According to the idea, the disparity map in step S101 is compared with the road surface model pixel by pixel, a region that may be an obstacle is set as a foreground, other regions are set as backgrounds, and the obtained binary image is a segmentation result, which is shown in fig. 6.
In step S105, morphological operations are performed on the binary image to obtain a template of the area to be detected of the suspected obstacle.
More specifically, the process of generating, according to the formula,
and (3) performing morphological operation on the binary image to obtain a template of the area to be detected of the suspected obstacle, which is specifically shown in fig. 7.
In step S107, information fusion is performed on the template of the region to be detected and a monocular image to obtain a grayscale image of the region to be detected, where the monocular image is one of the road images.
More specifically, connected domain statistics is performed on the template of the region to be detected in step S105 to obtain the minimum circumscribed rectangle of each connected region, and a grayscale image of the region to be detected can be obtained from the positions of the rectangles on the monocular image, as shown in fig. 8. Wherein the monocular image selects the left view of the corresponding road image because the position where the object is imaged is different in the left and right views.
In step S109, the grayscale image is input to a preset obstacle detection model to determine an obstacle in the road image, and the suspected obstacle is confirmed according to the determination result.
More specifically, a preset obstacle detection model generated based on machine learning is utilized to judge obstacles of a gray level image of a region to be detected, target recognition and marking are carried out on suspected obstacles, and target classification is carried out on marked targets.
In the technical scheme of the embodiment, on the basis that the binocular vision parallax map is subjected to image segmentation based on the road model, monocular information is fused, and then target identification and target classification based on machine learning are performed, so that rapid and accurate obstacle detection is realized, and the obstacle detection result is more robust.
Preferentially, the step of judging the obstacle by using the gray-scale image and a preset obstacle identification model, and the step of identifying the suspected obstacle according to a judgment result comprises the steps of: performing feature analysis on the gray level image through the preset obstacle recognition model, performing target recognition on the suspected obstacle according to an analysis result, and marking to obtain a marked target; and classifying the marked target through the preset obstacle classification model.
In this embodiment, a specific implementation manner of step S109 is given, and the preset obstacle identification model includes an obstacle identification model and an obstacle classification model. After the obstacle is identified and marked, the marked target is classified according to the obstacle classification sample set, so that the identification result is more available.
Preferably, the method for acquiring the preset obstacle detection model includes: acquiring a plurality of groups of road images as samples through binocular equipment, and generating corresponding disparity maps by taking disparity information of each group of images as known input quantity; carrying out image segmentation based on road surface information on the disparity map to obtain a binary map with a suspected obstacle area as a foreground and other areas as backgrounds; performing morphological operation on the binary image to obtain a template of a to-be-detected area of the suspected obstacle; performing information fusion on the template of the area to be detected and a monocular image, performing connected domain statistics on the template of the area to be detected to obtain a minimum external rectangle of each connected region, and obtaining a gray level image of the area to be detected from the position of the rectangle on the monocular image, wherein the monocular image is one of the road images; and screening and concentrating the obstacles in the gray level image to form the obstacle identification model. And classifying the identified marked targets to form an obstacle classification model.
The embodiment specifically provides an acquisition method of a preset obstacle detection model, which is similar to an obstacle detection process, except that machine learning training is performed on a large number of sample images to screen and concentrate obstacles to form an obstacle identification model and an obstacle classification model when the preset obstacle detection model is acquired.
Fig. 10 is a structural view of a road obstacle detecting apparatus according to an embodiment of the present invention.
Referring to fig. 10, a road obstacle detecting apparatus includes: the processing module 100 is configured to acquire a road image through binocular equipment, and process the road image to obtain a disparity map of the road image; the first segmentation module 102 is configured to perform image segmentation based on road surface information on the disparity map to obtain a binary map including an obstacle region as a foreground and other regions as a background; the first operation module 104 is configured to perform morphological operation on the binary image to obtain a template of a to-be-detected region of the suspected obstacle; the first fusion module 106 is configured to perform information fusion on the template of the region to be detected and a monocular image to obtain a grayscale image of the region to be detected, where the monocular image is one of the road images; and the judging module 108 is configured to input the grayscale image to a preset obstacle detection model to judge an obstacle of the road image, and confirm the suspected obstacle according to a judgment result.
Optionally, the determining module 108 includes: an analysis module (not shown in the figure) for performing feature analysis on the gray level image through an obstacle recognition model in the preset obstacle detection model, and performing target recognition and marking on the suspected obstacle according to an analysis result to obtain a marked target; and a classification module (not shown) for classifying the labeled target through an obstacle classification model in the preset obstacle detection model.
Optionally, the classification module comprises: the generating module (not shown in the figure) is used for acquiring a plurality of groups of road images as samples through binocular equipment and generating corresponding disparity maps by taking the disparity information of each group of images as known input quantity; a second segmentation module (not shown in the figure) for performing image segmentation based on road surface information on the disparity map to obtain a binary map including a suspected obstacle area as a foreground and other areas as a background; a second operation module (not shown in the figure) for performing morphological operation on the binary image to obtain a template of the area to be detected of the suspected obstacle; a second fusion module (not shown in the figure) for performing information fusion on the template of the region to be detected and the monocular image, performing connected domain statistics on the template of the region to be detected to obtain a minimum external rectangle of each connected region, and obtaining a grayscale image of the region to be detected from the position of the rectangle on the monocular image, wherein the monocular image is one of the road images; a screening module (not shown in the figure) for screening and concentrating the obstacles in the gray level image to form the obstacle identification model; and the training module (not shown in the figure) is used for performing machine training learning on the preset obstacle recognition sample set to obtain the obstacle classification model.
Optionally, the first segmentation module comprises: a first obtaining module (not shown) for obtaining a road surface model without obstacles in an imaging range through the road surface information; and an input module (not shown) for inputting the road surface model to perform image segmentation on the disparity map to obtain a binary map including a suspected obstacle area as a foreground and other areas as a background.
Optionally, the processing module 100 includes: and a second obtaining module (not shown in the figure) for obtaining the parallax information of the road image, displaying the parallax information in an image matrix form, representing different parallax values with different colors, and visually displaying the digitized parallax information in an image form to generate the parallax map.
On the basis of image segmentation of a binocular vision parallax map based on a road model, monocular information is fused, and then target identification and target classification based on machine learning are carried out, so that rapid and accurate obstacle detection is realized. The scheme of the invention particularly focuses on the interference of the lane line to the obstacle, so that the obstacle detection result is more robust.
On the basis of image segmentation of a binocular vision parallax map based on a road model, monocular information is fused, and then target identification and target classification based on machine learning are carried out, so that rapid and accurate obstacle detection is realized. The scheme for training and learning of the invention particularly focuses on the interference of the background environment on the obstacle, so that the obstacle detection result is more robust.
While certain exemplary embodiments of the present invention have been described above by way of illustration only, it will be apparent to those of ordinary skill in the art that the described embodiments may be modified in various different ways without departing from the spirit and scope of the invention. Accordingly, the drawings and description are illustrative in nature and should not be construed as limiting the scope of the invention.
Claims (4)
1. A method of road obstacle detection, comprising:
acquiring a road image through binocular equipment, and processing the road image to obtain a disparity map of the road image;
carrying out image segmentation based on road surface information on the disparity map to obtain a binary map with an obstacle region as a foreground and other regions as backgrounds;
performing morphological operation on the binary image to obtain a template of a to-be-detected area of the suspected obstacle;
performing information fusion on the template of the area to be detected and a monocular image to obtain a gray level image of the area to be detected, wherein the monocular image is one of the road images;
inputting the gray level image to a preset obstacle detection model to judge obstacles on the road image, and confirming the suspected obstacles according to a judgment result;
wherein, the inputting the gray image to a preset obstacle detection model to judge the obstacle of the road image, and the confirming the suspected obstacle according to the judgment result comprises:
performing feature analysis on the gray level image through an obstacle recognition model in the preset obstacle detection model, and performing target recognition and marking on the suspected obstacle according to an analysis result to obtain a marked target;
classifying the marked targets through an obstacle classification model in the preset obstacle detection model; the method for acquiring the preset obstacle detection model comprises the following steps:
acquiring a plurality of groups of road images as samples through binocular equipment, and generating corresponding disparity maps by taking disparity information of each group of images as known input quantity;
carrying out image segmentation based on road surface information on the disparity map to obtain a binary map with a suspected obstacle area as a foreground and other areas as backgrounds;
performing morphological operation on the binary image to obtain a template of a to-be-detected area of the suspected obstacle;
performing information fusion on the template of the area to be detected and a monocular image, performing connected domain statistics on the template of the area to be detected to obtain a minimum external rectangle of each connected region, and obtaining a gray level image of the area to be detected from the position of the rectangle on the monocular image, wherein the monocular image is one of the road images;
screening and concentrating the obstacles in the gray level image to form an obstacle identification model;
performing machine training learning on the preset obstacle recognition sample set to obtain the obstacle classification model;
the method for obtaining the road image through the binocular equipment and processing the road image comprises the following steps of:
and acquiring parallax information of the road image, displaying the parallax information in an image matrix form, representing different parallax values by different colors, and visually displaying the digitized parallax information in an image form to generate the parallax map.
2. The method according to claim 1, wherein the image segmentation based on road surface information is performed on the disparity map, and obtaining a binary map including a suspected obstacle area as a foreground and other areas as a background includes:
obtaining a road surface model without obstacles in an imaging range through the road surface information;
and inputting the road surface model to carry out image segmentation on the parallax image to obtain a binary image with a suspected obstacle area as a foreground and other areas as a background.
3. A road obstacle detection device, comprising:
the processing module is used for acquiring a road image through binocular equipment and processing the road image to obtain a disparity map of the road image;
the first segmentation module is used for carrying out image segmentation based on road surface information on the parallax map to obtain a binary map with an obstacle area as a foreground and other areas as backgrounds;
the first operation module is used for carrying out morphological operation on the binary image to obtain a template of the area to be detected of the suspected obstacle;
the first fusion module is used for carrying out information fusion on the template of the area to be detected and a monocular image to obtain a gray level image of the area to be detected, wherein the monocular image is one of the road images;
the judging module is used for inputting the gray level image to a preset obstacle detection model to judge obstacles on the road image and confirming the suspected obstacles according to a judging result;
wherein, the judging module comprises:
the analysis module is used for performing feature analysis on the gray level image through an obstacle recognition model in the preset obstacle detection model, performing target recognition on the suspected obstacle according to an analysis result, and marking to obtain a marked target;
the classification module is used for classifying the marked targets through an obstacle classification model in the preset obstacle detection model;
wherein the classification module comprises:
the generating module is used for acquiring a plurality of groups of road images as samples through binocular equipment and generating corresponding disparity maps by taking the disparity information of each group of images as known input quantity;
the second segmentation module is used for carrying out image segmentation based on road surface information on the parallax map to obtain a binary map with a suspected obstacle area as a foreground and other areas as a background;
the second operation module is used for carrying out morphological operation on the binary image to obtain a template of the area to be detected of the suspected obstacle;
the second fusion module is used for performing information fusion on the template of the area to be detected and the monocular image, performing connected domain statistics on the template of the area to be detected to obtain a minimum external rectangle of each connected region, and obtaining a gray level image of the area to be detected from the position of the rectangle on the monocular image, wherein the monocular image is one of the road images;
the screening module is used for screening and concentrating the obstacles in the gray level image to form the obstacle identification model;
the training module is used for performing machine training learning on the preset obstacle recognition sample set to obtain the obstacle classification model; wherein the processing module comprises:
and the second acquisition module is used for acquiring the parallax information of the road image, displaying the parallax information in an image matrix form, representing different parallax values by using different colors, and visually displaying the digitized parallax information in an image form to generate the parallax map.
4. The road obstacle detection device according to claim 3, wherein the first division module includes:
the first acquisition module is used for acquiring a road surface model without obstacles in an imaging range through the road surface information;
and the input module is used for inputting the road surface model to carry out image segmentation on the parallax map so as to obtain a binary map with a suspected obstacle area as a foreground and other areas as a background.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711031619.9A CN107909010B (en) | 2017-10-27 | 2017-10-27 | Road obstacle detection method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711031619.9A CN107909010B (en) | 2017-10-27 | 2017-10-27 | Road obstacle detection method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107909010A CN107909010A (en) | 2018-04-13 |
| CN107909010B true CN107909010B (en) | 2022-03-18 |
Family
ID=61842025
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711031619.9A Active CN107909010B (en) | 2017-10-27 | 2017-10-27 | Road obstacle detection method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107909010B (en) |
Families Citing this family (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10242264B1 (en) * | 2018-04-25 | 2019-03-26 | Imagry (Israel) Ltd. | System and method for training a machine-learning model to identify real-world elements |
| CN108648211B (en) * | 2018-05-16 | 2020-05-01 | 思百达物联网科技(北京)有限公司 | Small target detection method, device, equipment and medium based on deep learning |
| CN109270524B (en) * | 2018-10-19 | 2020-04-07 | 禾多科技（北京）有限公司 | Unmanned-vehicle-based multi-data fusion obstacle detection device and detection method thereof |
| CN109298415B (en) * | 2018-11-20 | 2020-09-22 | 中车株洲电力机车有限公司 | Method for detecting obstacles on track and road |
| CN109600546A (en) * | 2018-11-26 | 2019-04-09 | 维沃移动通信（杭州）有限公司 | A kind of image-recognizing method and mobile terminal |
| CN109766784A (en) * | 2018-12-21 | 2019-05-17 | 北京理工大学 | Capsule robot interaction control method based on monocular image |
| CN110335322B (en) * | 2019-07-09 | 2024-03-01 | 成都理工大学 | Road recognition method and road recognition device based on image |
| CN110502983B (en) * | 2019-07-11 | 2022-05-06 | 平安科技（深圳）有限公司 | Method and device for detecting obstacles in expressway and computer equipment |
| CN111489345B (en) * | 2020-04-13 | 2023-08-15 | 中国科学院高能物理研究所 | Training method, device, equipment and storage medium of region segmentation model |
| CN112810669A (en) * | 2020-07-17 | 2021-05-18 | 周慧 | Intercity train operation control platform and method |
| CN111932581B (en) * | 2020-08-11 | 2023-09-26 | 沈阳帝信人工智能产业研究院有限公司 | Safety rope detection method, device, electronic equipment and readable storage medium |
| CN113128347B (en) * | 2021-03-24 | 2024-01-16 | 北京中科慧眼科技有限公司 | Obstacle target classification method and system based on RGB-D fusion information and intelligent terminal |
| CN113378628B (en) * | 2021-04-27 | 2023-04-14 | 阿里云计算有限公司 | Road obstacle area detection method |
| CN113793330B (en) * | 2021-11-10 | 2022-03-18 | 北京中科慧眼科技有限公司 | Method and system for detecting road surface flatness area |
| CN115423865A (en) * | 2022-07-29 | 2022-12-02 | 松灵机器人(深圳)有限公司 | Obstacle detection method, obstacle detection device, mowing robot, and storage medium |
| CN115393813B (en) * | 2022-08-18 | 2023-05-02 | 中国人民公安大学 | Road identification method, device, equipment and storage medium based on remote sensing image |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5128874A (en) * | 1990-01-02 | 1992-07-07 | Honeywell Inc. | Inertial navigation sensor integrated obstacle detection system |
| CN101852609A (en) * | 2010-06-02 | 2010-10-06 | 北京理工大学 | Ground obstacle detection method based on binocular stereo vision of robot |
| CN103033817A (en) * | 2012-11-25 | 2013-04-10 | 中国船舶重工集团公司第七一○研究所 | Obstruction automatic recognition system for collision preventing of large-scale autonomous underwater vehicle (AUV) |
| CN103955920A (en) * | 2014-04-14 | 2014-07-30 | 桂林电子科技大学 | Binocular vision obstacle detection method based on three-dimensional point cloud segmentation |
| CN104021388A (en) * | 2014-05-14 | 2014-09-03 | 西安理工大学 | Reversing obstacle automatic detection and early warning method based on binocular vision |
| CN105550665A (en) * | 2016-01-15 | 2016-05-04 | 北京理工大学 | Method for detecting pilotless automobile through area based on binocular vision |
| CN105740802A (en) * | 2016-01-28 | 2016-07-06 | 北京中科慧眼科技有限公司 | Disparity map-based obstacle detection method and device as well as automobile driving assistance system |
| CN105761306A (en) * | 2016-01-29 | 2016-07-13 | 珠海汇迪科技有限公司 | Road surface model based on field depth image or point cloud and establishment method thereof |
| CN106599832A (en) * | 2016-12-09 | 2017-04-26 | 重庆邮电大学 | Method for detecting and recognizing various types of obstacles based on convolution neural network |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103679707A (en) * | 2013-11-26 | 2014-03-26 | 西安交通大学 | Binocular camera disparity map based road obstacle detection system and method |
| CN103714533A (en) * | 2013-12-09 | 2014-04-09 | 广西科技大学 | Method for automatically detecting obstacles based on binocular vision |
| CN106228110B (en) * | 2016-07-07 | 2019-09-20 | 浙江零跑科技有限公司 | A kind of barrier and drivable region detection method based on vehicle-mounted binocular camera |
| CN106650701B (en) * | 2017-01-05 | 2020-01-14 | 华南理工大学 | Binocular vision-based obstacle detection method and device in indoor shadow environment |
| CN106650708B (en) * | 2017-01-19 | 2023-08-11 | 南京航空航天大学 | Automatic driving obstacle vision detection method and system |
-
2017
- 2017-10-27 CN CN201711031619.9A patent/CN107909010B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5128874A (en) * | 1990-01-02 | 1992-07-07 | Honeywell Inc. | Inertial navigation sensor integrated obstacle detection system |
| CN101852609A (en) * | 2010-06-02 | 2010-10-06 | 北京理工大学 | Ground obstacle detection method based on binocular stereo vision of robot |
| CN103033817A (en) * | 2012-11-25 | 2013-04-10 | 中国船舶重工集团公司第七一○研究所 | Obstruction automatic recognition system for collision preventing of large-scale autonomous underwater vehicle (AUV) |
| CN103955920A (en) * | 2014-04-14 | 2014-07-30 | 桂林电子科技大学 | Binocular vision obstacle detection method based on three-dimensional point cloud segmentation |
| CN104021388A (en) * | 2014-05-14 | 2014-09-03 | 西安理工大学 | Reversing obstacle automatic detection and early warning method based on binocular vision |
| CN105550665A (en) * | 2016-01-15 | 2016-05-04 | 北京理工大学 | Method for detecting pilotless automobile through area based on binocular vision |
| CN105740802A (en) * | 2016-01-28 | 2016-07-06 | 北京中科慧眼科技有限公司 | Disparity map-based obstacle detection method and device as well as automobile driving assistance system |
| CN105761306A (en) * | 2016-01-29 | 2016-07-13 | 珠海汇迪科技有限公司 | Road surface model based on field depth image or point cloud and establishment method thereof |
| CN106599832A (en) * | 2016-12-09 | 2017-04-26 | 重庆邮电大学 | Method for detecting and recognizing various types of obstacles based on convolution neural network |
Non-Patent Citations (6)
| Title |
|---|
| A Global Optimization Algorithm for Real-Time On-Board Stereo Obstacle Detection Systems;Susumu Kubota 等;《2007 IEEE Intelligent Vehicles Symposium》;20070813;7-12 * |
| A Stereovision Method for Obstacle Detection and Tracking in Non-Flat Urban Environments;QIAN YU 等;《Autonomous Robots》;20050930;141-157 * |
| 低照度立体视觉及显著目标的距离测定;万一龙;《中国优秀硕士学位论文全文数据库（信息科技辑）》;20140715;第2014年卷(第07期);第3章 * |
| 基于加权Hough变换和时域相关性分析的双目视频障碍物检测技术研究;黄俊;《中国优秀硕士学位论文全文数据库（工程科技Ⅱ辑）》;20140815;第2014年卷(第08期);第3章、第5章 * |
| 基于双目视觉的道路障碍物识别与检测方法研究;李靖 等;《江苏科技信息》;20170430;第2017年卷(第11期);52-55 * |
| 基于双目视觉的障碍物检测算法[;林川 等;《电视技术》;20141231;第38卷(第5期);186-189 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107909010A (en) | 2018-04-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107909010B (en) | Road obstacle detection method and device | |
| CN109034047B (en) | Lane line detection method and device | |
| CN108364010B (en) | License plate recognition method, device, equipment and computer readable storage medium | |
| CN105488454B (en) | Front vehicles detection and ranging based on monocular vision | |
| CN108898047B (en) | Pedestrian detection method and system based on blocking and shielding perception | |
| CN105913093B (en) | A kind of template matching method for Text region processing | |
| CN107909009B (en) | Obstacle detection method and device based on road surface learning | |
| CN107424142B (en) | Weld joint identification method based on image significance detection | |
| KR101403876B1 (en) | Method and Apparatus for Vehicle License Plate Recognition | |
| CN104700072B (en) | Recognition methods based on lane line historical frames | |
| CN107992819B (en) | Method and device for determining vehicle attribute structural features | |
| EP2775349A1 (en) | A method for determining an in-focus position and a vision inspection system | |
| CN106951898B (en) | Vehicle candidate area recommendation method and system and electronic equipment | |
| CN105718931B (en) | System and method for determining clutter in acquired images | |
| JP2015215877A (en) | Object detection method from stereo image pair | |
| CN111126393A (en) | Vehicle appearance refitting judgment method and device, computer equipment and storage medium | |
| JP2009163682A (en) | Image discrimination device and program | |
| CN104966283A (en) | Imaging layered registering method | |
| CN113569679B (en) | Method, device and system for measuring elongation at break | |
| CN110675442A (en) | Local stereo matching method and system combined with target identification technology | |
| CN110969135B (en) | Vehicle logo recognition method in natural scene | |
| Kurnianggoro et al. | Camera and laser range finder fusion for real-time car detection | |
| CN105718929A (en) | All-weather unknown environment high precision rapid circular object positioning method and system | |
| CN202771439U (en) | Traffic sign automatic identification system based on MATLAB | |
| CN109753858A (en) | A kind of road barricade object detecting method and device based on binocular 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 |