A method for producing high-precision maps based on real-world 3D models
A three-dimensional model, high-precision technology, applied in 3D modeling, image data processing, character and pattern recognition, etc., can solve the problems of low precision of DOM road network plane, inability to reflect elevation information, and more manpower consumption. Achieve the effect of breaking the difficult acquisition and low accuracy of the roll angle, improving the accuracy of image recognition and segmentation, and improving automation
Active Publication Date: 2021-02-23
TERRA DIGITAL CREATING SCI & TECH (BEIJING) CO LTD
View PDF13 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] The existing road network extraction methods still have quite a few shortcomings: 1. The plane accuracy of the road network extraction based on DOM is not high, and the elevation information cannot be reflected at all; 2. The road network extraction based on stereo pairs requires more manpower, less efficient
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
example I
[0077] Each pixel in the recognition result is assigned a semantic label and an instance ID;
[0078] The pixels with the same semantic label and instance ID are grouped into the same object, and the elements of the real-world 3D road network model are segmented.
[0079] In this embodiment, the resource pool contains a collection of all elements of the road, including graphic images and individual attributes; the elements of the segmented real-scene 3D road network model are matched one by one with the existing graphic image elements in the resource pool , the matching matching rule is that when the overlap ratio between the segmentation window predicted by the real-world 3D road network model and the original image marker window in the resource pool is greater than 0.5, the two can match, and the matching result is obtained and retained.
[0080] In this embodiment, in step S8, the road elements are classified into three types, which are respectively,
[0081] Lane model; i...
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 provides a method for producing a high-precision map based on a real-scene three-dimensional model, which specifically includes performing perspective transformation on the real-scene three-dimensional model to collect multi-view images, adopting a multi-view image recognition method for road network feature recognition, and adopting image panorama segmentation. The method segmented the recognition results, matched the elements of the segmented real-scene 3D road network model with the existing graphic image elements in the resource pool, vectorized the matching results and the segmented real-scene 3D road network model elements, and retained Carry out fine classification of road elements, obtain fine classification results and other steps, and finally obtain high-precision maps. The advantages are: it solves the problem that field collection is affected by external conditions such as weather, road conditions, and moving objects, and greatly reduces the cost of high-precision map production; the real-scene 3D model serves as an intuitive mapping of the real world, breaking the existing high-precision The map production mode also makes the accuracy and detailed expression of the map more perfect.
Description
technical field [0001] The invention relates to the field of high-precision map production, in particular to a method for producing a high-precision map based on a real-scene three-dimensional model. Background technique [0002] With the continuous development of high-resolution remote sensing images and aerial photography acquisition methods, the technology of road network extraction through aerial photography has also made great progress. At present, the road network extraction methods mainly include: road network extraction based on DOM and road network extraction based on stereo image pairs. [0003] The extraction of road network based on DOM includes the following contents: 1. Obtain aerial images; 2. Conduct aerial triangulation; 3. Generate DOM (digital orthophoto map) according to the results of the second step; 4. Manually collect road network according to DOM and sketch. [0004] The extraction of road network based on stereo image pairs includes the following ...
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 Patents(China)
IPC IPC(8): G06K9/00G06K9/34G06K9/62G06T17/00
CPCG06T17/00G06V20/588G06V20/56G06V10/267G06F18/214
Inventor 刘俊伟李嘉榆
Owner TERRA DIGITAL CREATING SCI & TECH (BEIJING) CO LTD