Target ranging method based on deep learning
A technology of target ranging and deep learning, applied in the field of artificial intelligence, can solve problems such as high cost and single target information, and achieve the effect of reducing hardware costs
Inactive Publication Date: 2018-12-18
QUANZHOU INST OF EQUIP MFG
View PDF4 Cites 26 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
The traditional stereo vision ranging method based on binocular cameras is expensive, and the target information obtained is relatively simple, so it cannot be well promoted
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
[0035] This implementation provides a ranging method based on deep learning, refer to Figure 1 to Figure 3 , a method for ranging based on deep learning according to an embodiment of the present invention includes the following steps.
[0036] Step S101, based on the establishment of target databases under different distances;
[0037] Step S102, building a target ranging model;
[0038] Step S103, designing a loss function of the target ranging model;
[0039] Step S104, designing a training method for the target ranging model;
[0040] Step S105, testing the trained target ranging model.
[0041] We will detail the deep learning based odometry method with the optimal implementation below.
[0042] First build a data collection platform. That is, the position of the binocular vision camera and the focal length of the camera are fixed, and the video information of the same target is collected at different positions from the camera. In this method, the farthest distance f...
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 a target ranging method based on deep learning. The method includes the following specific steps that (1), a target database based on different distances is established; (2), atarget ranging model is established; (3), loss functions of the target ranging model are designed; (4), a training method of the target ranging model is designed; and (5), the trained target rangingmodel is tested. According to the innovations of the target ranging method based on the deep learning, based on a video image target detection algorithm in the deep learning, a target ranging problemis transformed into a regression problem and integrated into a target detection algorithm model, and thus target detection and target ranging in an algorithm model are achieved. The target ranging method based on the deep learning has the advantages that through the algorithm, target category information and target distance information can be obtained only by ordinary monocular cameras so that traditional binocular cameras or laser radars can be replaced to obtain depth information of a target, and hardware cost is reduced.
Description
technical field [0001] The invention relates to a target ranging method based on deep learning, which belongs to the field of artificial intelligence. Background technique [0002] With the generation of big data and the improvement of computer computing power, artificial intelligence technology has been further developed. How to apply emerging technologies to traditional problems is an urgent problem to be solved. The traditional stereo vision ranging method based on binocular cameras is expensive, and the target information obtained is relatively simple, so it cannot be promoted very well. [0003] Deep learning (deep learning) is a branch of machine learning. It is an algorithm that attempts to use multiple processing layers that contain complex structures or consist of multiple nonlinear transformations to perform high-level abstraction on data; Data representation learning algorithms; observations (such as an image) can be represented in a variety of ways, such as a v...
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): G01C11/00
CPCG01C11/00G01C11/36
Inventor 魏宪张文涛兰海王泽荔郭杰龙孙威振
Owner QUANZHOU INST OF EQUIP MFG