Unlock instant, AI-driven research and patent intelligence for your innovation.
Remote sensing object detection method based on few samples
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
A target detection and sample technology, which is applied in the field of remote sensing target detection, can solve problems such as few labeled samples, lack of generalization ability of the detection model, and a small number of samples
Active Publication Date: 2018-02-02
UNIVERSITY OF CHINESE ACADEMY OF SCIENCES
View PDF6 Cites 29 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
However, in the field of remote sensing applications, the acquisition of target samples has great limitations
First of all, it is very time-consuming and laborious to manually label large-scale image data; secondly, for some specific targets in remote sensing scenes, there are only a few samples in large-scale image data, and it is not enough for complete training. data set
[0004] However, in remote sensing target detection, there is a problem of a small number of samples. It is difficult to directly obtain a large number of samples in remote sensing images. The reason is that the targets in remote sensing ima
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 more
Image
Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Embodiment approach
[0062] According to a preferred embodiment of the present invention, step 2 includes the following sub-steps:
[0063] Step 2.1, analyze the distribution characteristics of the target samples in the sample data set obtained in step 1, and use the target samples and target categories to perform three-dimensional modeling to obtain the three-dimensional model of the target;
[0064] Wherein, the target category refers to the category of the target to be detected, such as vehicle category, aircraft category or other categories; the distribution characteristics include the geometric statistical characteristics of the target such as the sum of pixel values, angle, and aspect ratio.
[0065] Step 2.2, merging the three-dimensional model of the target obtained in step 2.1 with the remote sensing background to form a virtual data set;
[0066] Step 2.3, performing noise addition processing on the virtual data set formed in step 2.2 to obtain an auxiliary data set.
[0067] According ...
Embodiment
[0112] Adopt method of the present invention to carry out the detection of remote sensing vehicle target, its specific embodiment is as follows:
[0113] Collect 300 remote sensing vehicle sample data and remote sensing backgrounds whose positions and contours have been marked. Through the data enhancement method described in step 1.3, such as Figure 4 As shown in Fig. 1, after rotation, flipping, scale transformation, and contrast transformation, the extended data of 1500 vehicle samples is finally obtained.
[0114] According to the angle and aspect ratio distribution obtained from the extended vehicle sample, the general vehicle 3D model is adjusted in the 3D modeling software SketchUp, specifically, as Figure 5 As shown, on the basis of the top view angle, the length and width of the model are stretched and compressed to obtain a three-dimensional model suitable for the angle and aspect ratio distribution of this embodiment.
[0115] In the 3D modeling software, combin...
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 remote sensing object detection method based on a few samples, and the method comprises the steps: firstly carrying out the processing of collected samples, and obtaining anauxiliary data set; secondly carrying out the similarity verification of the auxiliary data set and a target domain; thirdly training DPM (Deformable Part based Model) parameters according to the auxiliary data set, obtaining a target classifier, and carrying out the target detection, wherein the parameters in the model are divided during the training of the DPM parameters, and then the transfer learning is carried out to obtain the target classifier. The method employs a target clustering mode during the training of a source domain data mixed model, can effectively reduce the training errorsand can reduce the training complexity. Moreover, a structural parameter dividing method is employed, so the method can improve the detection performances of an adaptive model.
Description
technical field [0001] The invention relates to the field of remote sensing target detection, in particular to a remote sensing target detection method based on a small number of samples. Background technique [0002] Remote sensing target detection is the process of accurately locating or identifying one or more types of objects in aerial images and videos through machine learning methods and computer vision algorithms. It is an advanced task in computer vision and is often used as the research basis for activity analysis, event recognition, scene understanding and object tracking. With the development of high-resolution remote sensing imaging technology in China, remote sensing target detection will play an increasingly important role in the future. [0003] However, the traditional target detection technology needs to label a large number of samples from the application data for the training of the detection classifier, so that the algorithm can approach the real distrib...
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
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to View More 
Login to View More 
