Method, apparatus, storage medium and device for modeling lane line identification, and method, apparatus, storage medium and device for identifying lane line

• US 10,699,134 B2
• Filed: 12/31/2015
• Issued: 06/30/2020
• Est. Priority Date: 08/03/2015
• Status: Active Grant

• Current Assignee: Baidu Online Network Technology (Beijing) Co., Ltd (Baidu Incorporated)
• Original Assignee: Baidu Online Network Technology (Beijing) Co., Ltd (Baidu Incorporated)
• Inventors: He, Bei, Hao, Zhihui
• Primary Examiner(s): Entezari, Michelle M
• Application Number:
  US15/750,127
• Publication Number:
  US 20180225527A1
• Time in Patent Office: 1,643 Days
• Field of Search: None
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06F 18/21375 involving differential geometry, e.g. embedding of pattern manifold
• CPC Class Codes: G06F 18/2148 characterised by the process organisation or structure, e.g. boosting cascade
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06V 10/82 using neural networks
• CPC Class Codes: G06V 20/588 Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road

Abstract

Embodiments of the present disclosure disclose a lane line identification modeling method and apparatus, and a lane line identification method and apparatus. The lane line identification modeling method includes: identifying an image region of a lane line in an image based on two-dimensional filtering (S11); constructing model training data by using the identified image region (S12); and training a convolutional neural network-based lane line identification model by using the model training data (S13). The lane line identification modeling method and apparatus, and the lane line identification method and apparatus provided in the embodiments of the present disclosure can improve the accuracy of lane line detection.

Claims

1. A method for modeling a lane line identification, comprising:

identifying an image region of a lane line in an image based on two-dimensional filtering;

constructing model training data by using the identified image region; and

training a convolutional neural network-based lane line identification model by using the model training data;

wherein the method is performed by one or more processors;

wherein before the identifying an image region of a lane line in an image based on two-dimensional filtering, the method further comprises;

performing an inverse projection transformation on an original image to adjust an optical axis direction of the original image to be perpendicular to a ground surface;

wherein the performing an inverse projection transformation on an original image to adjust an optical axis direction of the original image to be perpendicular to a ground surface comprises;

performing the inverse projection transformation on the original image according to;

2. The method according to claim 1, wherein the identifying an image region of a lane line in an image based on two-dimensional filtering comprises:

performing filtering on the image by using filter kernels having different width parameters and height parameters, and selecting an image having a most distinguished image contour as a filtering result image;

binarizing the filtering result image to form a connected domain; and

performing linear boundary fitting on the connected domain by using an improved ransac algorithm.

3. The method according to claim 2, wherein the constructing model training data by using the identified image region comprises:

broadening the connected domain to form a region of interest on the image; and

using the image comprising the region of interest as the model training data.

4. A method for identifying a lane line, comprising:

identifying an image region of a lane line in an image based on two-dimensional filtering;

inputting the image, in which the image region of the lane line has been identified, into a convolutional neural network-based lane line identification model to obtain an output probability of the model; and

performing model reconstruction based on the output probability to identify the lane line in the input image;

wherein the method is performed by one or more processors;

wherein the performing model reconstruction based on the output probability to identify the lane line in the input image comprises;

performing model reconstruction according to;

5. The method according to claim 4, wherein before the identifying an image region of a lane line in an image based on two-dimensional filtering, the method further comprises:

performing an inverse projection transformation on an original image to adjust an optical axis direction of the original image to be perpendicular to a ground surface.

6. A non-transitory computer storage medium storing computer executable instructions, the computer executable instructions, when executed by a processor of a computer, causing the processor to execute a method for modeling a lane line identification, wherein the method comprises:

identifying an image region of a lane line in an image based on two-dimensional filtering;

constructing model training data by using the identified image region; and

training a convolutional neural network-based lane line identification model by using the model training data;

wherein before the identifying an image region of a lane line in an image based on two-dimensional filtering, the method further comprises;

performing an inverse projection transformation on an original image to adjust an optical axis direction of the original image to be perpendicular to a ground surface;

wherein the performing an inverse projection transformation on an original image to adjust an optical axis direction of the original image to be perpendicular to a ground surface comprises;

performing the inverse projection transformation on the original image according to;

7. A device, comprising:

one or more processors; and

a memory;

wherein one or more programs are stored in the memory, and when executed by the one or more processors, the one or more programs cause the one or more processors to perform operations comprising;

identifying an image region of a lane line in an image based on two-dimensional filtering;

constructing model training data by using the identified image region; and

training a convolutional neural network-based lane line identification model by using the model training data;

wherein before the identifying an image region of a lane line in an image based on two-dimensional filtering, the operations further comprise;

performing an inverse projection transformation on an original image to adjust an optical axis direction of the original image to be perpendicular to a ground surface;

wherein the performing an inverse projection transformation on an original image to adjust an optical axis direction of the original image to be perpendicular to a ground surface comprises;

performing the inverse projection transformation on the original image according to;

10. The device according to claim 7, wherein the identifying an image region of a lane line in an image based on two-dimensional filtering comprises:

performing filtering on the image by using filter kernels having different width parameters and height parameters, and selecting an image having a most distinguished image contour as a filtering result image;

binarizing the filtering result image to form a connected domain; and

performing linear boundary fitting on the connected domain by using an improved ransac algorithm.

11. The device according to claim 10, wherein the constructing model training data by using the identified image region comprises:

broadening the connected domain to form a region of interest on the image; and

using the image comprising the region of interest as the model training data.

8. A non-transitory computer storage medium storing computer executable instructions, the computer executable instructions, when executed by a processor of a computer, causing the processor to execute a method for identifying a lane line, wherein the method comprises:

identifying an image region of a lane line in an image based on two-dimensional filtering;

inputting the image, in which the image region of the lane line has been identified, into a convolutional neural network-based lane line identification model to obtain an output probability of the model; and

performing model reconstruction based on the output probability to identify the lane line in the input image;

wherein the performing model reconstruction based on the output probability to identify the lane line in the input image comprises;

performing the model reconstruction according to;

9. A device, comprising:

one or more processors; and

a memory;

wherein one or more programs are stored in the memory, and when executed by the one or more processors, the one or more programs cause the one or more processors to perform operations comprising;

identifying an image region of a lane line in an image based on two-dimensional filtering;

inputting the image, in which the image region of the lane line has been identified, into a convolutional neural network-based lane line identification model to obtain an output probability of the model; and

performing model reconstruction based on the output probability to identify the lane line in the input image;

wherein the performing model reconstruction based on the output probability to identify the lane line in the input image comprises;

performing the model reconstruction according to;

12. The device according to claim 9, wherein before the identifying an image region of a lane line in an image based on two-dimensional filtering, the operations further comprises:

performing an inverse projection transformation on an original image to adjust an optical axis direction of the original image to be perpendicular to a ground surface.