Remote sensing image situation identification method based on SVM and step-by-step grid search
A technology of remote sensing image and recognition method, which is applied in the field of image processing and intent resolution, can solve the problems of low recognition accuracy and efficiency, and achieve the effects of improving detection accuracy, fast parameter search, and high accuracy of situation recognition
Pending Publication Date: 2022-07-15
XIDIAN UNIV
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0009] The purpose of the present invention is to overcome the defects in the above-mentioned prior art, and propose a remote sensing image situation recognition method based on SVM and step-by-step grid search, which is used to solve the problems of low recognition accuracy and efficiency in the prior art. technical problem
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
[0043] The present invention will be described in further detail below with reference to the accompanying drawings and specific embodiments.
[0044] refer to figure 1 , the present invention comprises the steps:
[0045] (1) Construct a remote sensing image target detection network model, the structure is as follows figure 2 shown:
[0046] (1a) Construct a remote sensing image target detection model including cascaded feature extraction sub-network, region generation sub-network, ROI Align pooling layer and classification and positioning sub-network; wherein, the feature extraction sub-network includes multiple cascaded feature extraction module, each feature extraction module contains multiple convolutional layers-ReLU layers connected in sequence and a max pooling layer; the region generation sub-network includes sequentially connected convolutional layers-ReLU layer, parallel connected first classification sub-network and The first positioning module and proposal laye...
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 remote sensing image situation recognition method based on an SVM (Support Vector Machine) and a step-by-step grid search algorithm, which mainly solves the technical problems of low recognition intention accuracy and slow SVM hyper-parameter search in the prior art, and comprises the following implementation steps: 1) constructing a remote sensing image target detection network model; 2) obtaining a training sample set and a test sample set; (3) an SVM classification model is applied, and hyper-parameters of the SVM classification model are trained; and 4) obtaining a situation identification result of the remote sensing image. According to the method, the contribution of a sample image with a poor detection effect to a loss function is increased, the detection effect of the whole sample image is improved, then the detected target number serves as one-dimensional data, other six-dimensional features are obtained from a situation map, a sample feature vector is manufactured, the hyper-parameter search range of an SVM classification model is determined firstly, and then the detection result is obtained. And then searching the hyper-parameter optimal solution of the SVM classification model, thereby improving the accuracy of the recognition intention of the SVM classification model and the hyper-parameter search speed.
Description
technical field [0001] The invention belongs to the technical field of image processing, and relates to a remote sensing image situation recognition method, in particular to a remote sensing image situation recognition method based on SVM and step-by-step grid search, which can be used in the field of intention discrimination. Background technique [0002] Remote sensing images refer to non-contact, long-distance detection images obtained by sensors. According to different sensors, it is divided into optical remote sensing images and synthetic aperture radar (Synthetic Aperture Radar, SAR) images. Compared with natural images, the target individuals of remote sensing images are smaller, the similarity between certain categories is large, and the target features are difficult to extract and are different from those extracted from natural images, resulting in poor feature extraction. The features of remote sensing images are divided into three categories: one is intrinsic fea...
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): G06V20/10G06V10/22G06V10/764G06V10/774G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/047G06N3/045G06F18/214G06F18/2411G06F18/2415
Inventor 李阳阳刘冠龙史雯熙刘光远魏强杨子琛张施淮焦李成
Owner XIDIAN UNIV