Scene classification method for high-resolution remote sensing images based on hierarchical multi-feature fusion
A multi-feature fusion and high-resolution technology, applied in the field of remote sensing image processing, can solve the problem of unsatisfactory classification accuracy and scene classification accuracy, and achieve high classification accuracy
Inactive Publication Date: 2019-02-19
HOHAI UNIV +1
View PDF5 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] Although the above existing technologies have their own characteristics, there is a common problem that the classification accuracy of some specific categories of scenes is relatively high, while the classification accuracy of other categories of scenes is not ideal.
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
[0027] The technical scheme of the present invention is described in detail below in conjunction with accompanying drawing:
[0028] For remote sensing images, a single feature can only describe some of its attributes, and lacks sufficient distinguishing information. When the image types are relatively similar (such as figure 1 In the case of dense residential areas, medium-density residential areas, and sparse residential areas (shown from left to right in the middle), it is usually not possible to achieve good classification results, and because some categories of images in remote sensing images have obvious color features and some categories of images The local features are more obvious, and if only a single feature is used, the classification accuracy will be low. In addition, the description based on global features and local features has its own advantages; global features are the global information describing the image, which can reflect the overall structure of the enti...
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 high-resolution remote sensing image scene classification method based on layered multi-feature fusion. The present invention clusters and analyzes the training sample image set according to the combination of global features, divides the training sample image set into a subset suitable for global feature expression and a subset suitable for local feature expression, and then uses the two subsets to train A global feature classifier and a local streamlined feature classifier; when testing, the classifier that is most suitable for the test sample is selected from the two classifiers according to the cluster analysis results to classify the test sample. The present invention has extremely high classification accuracy, and the overall average classification accuracy of the present invention can reach 96% through experimental verification, which is higher than the existing typical classification methods.
Description
technical field [0001] The invention relates to remote sensing image processing technology, in particular to a high-resolution remote sensing image scene classification method based on layered multi-feature fusion. Background technique [0002] In recent years, high-resolution earth observation technology has developed rapidly, and has played an important role in many fields such as land survey, urban planning and military affairs. Compared with low- and medium-resolution remote sensing images, high-resolution remote sensing images can provide detailed ground object information, but the spatial structure distribution of various ground objects is more complex. With the improvement of the resolution of remote sensing images, the geometric information and texture information of spatial objects are more obvious, but at the same time, it also brings about the enhancement of spectral differences of similar objects and the decrease of spectral heterogeneity of different objects. T...
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): G06K9/62
CPCG06F18/232G06F18/2413G06F18/2411G06F18/214
Inventor 蔡阳李士进蒋亚平陈德清王伶俐袁俐新付静郑展王继民余宇峰朱海晨王声特
Owner HOHAI UNIV