All-weather object recognition and lane detection methods for autonomous driving
A technology for lane line detection and object recognition, which is applied in neural learning methods, character and pattern recognition, biological neural network models, etc., can solve problems such as low efficiency, consumption, and multiple computing resources, so as to reduce repeated feature extraction and calculation volume effect
Active Publication Date: 2019-02-26
昆山星际舟智能科技有限公司
View PDF10 Cites 12 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the existing deep learning model of the visual perception module of deep automatic driving can only detect objects or lane lines alone. When objects or lane lines need to be checked at the same time, two models are required, which is not only inefficient, but also consumes more computing resources.
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0024] A preferred embodiment of the present invention will be described in detail below in conjunction with the accompanying drawings. However, the scope of protection of the present invention is not limited to the following examples, that is, any simple equivalent changes and modifications made based on the patent scope of the present invention and the content of the description are still within the scope of the patent of the present invention.
[0025] An all-weather object recognition and lane line detection method for automatic driving, comprising the following steps:
[0026] Step 1, collect sample pictures and build a training database: drive the test vehicle in different road scenes (such as urban roads, expressways, etc.), and use a near-infrared camera to collect image data during the test vehicle driving process, and collect at least 20,000 different scenes Save images of objects and objects to build a training database;
[0027] Step 2, image dataset labeling: Man...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses an all-weather object recognition and lane line detection method for automatic driving, which comprises the steps of collecting sample pictures and constructing a training database; Annotating image dataset; A method for construct a convolution neural network based on a training database; Training model steps; Model testing steps; The method integrates the results of the object recognition subnetwork and the lane line segmentation subnetwork, screens the results, removes the duplicate detection results, and finally obtains the class of the object in the current image, the rectangular frame coordinates and the lane line coordinates. The invention innovatively adopts a single depth learning model, At that same time, The model comprises a feature extraction sub-network, an object recognition sub-network and a lane line segmentation sub-network, and the feature extraction sub-network is shared by the other two sub-networks. The advantage of the invention lies in combining the advantages of detection and segmentation, and combining the two to reduce the repeated extraction of features and the computational load of the model.
Description
technical field [0001] The invention relates to an all-weather object recognition and lane line detection method for automatic driving. Background technique [0002] With the development of deep learning technology, there are more and more applications based on this technology, especially in the field of automatic driving. The main applications include: perception, fusion, decision-making, etc. However, the existing deep learning model of the visual perception module of deep automatic driving can only detect objects or lane lines alone. When objects or lane lines need to be checked at the same time, two models are required, which is not only inefficient, but also consumes more computing resources. . Contents of the invention [0003] In order to overcome the above defects, the present invention provides an all-weather object recognition and lane line detection method for automatic driving, which can simultaneously perform object recognition and lane line detection with on...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/00G06N3/04G06N3/08
CPCG06N3/08G06V20/588G06V20/56G06N3/045
Inventor 吴晓闯陆正达孙长亮
Owner 昆山星际舟智能科技有限公司