A Sample Equalization Method for Non-uniform Classes Oriented to Cloud Masks
A cloud mask and non-uniform technology, which is applied in the field of sample equalization of non-uniform categories, can solve problems such as sample imbalance, and achieve the effect of solving sample imbalance, improving accuracy, and enhancing small category samples
Active Publication Date: 2022-03-18
ZHEJIANG UNIV OF TECH
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0008] Aiming at the problem of sample imbalance caused by non-uniform categories, the present invention proposes a sample equalization method, so as to ensure that the classifier can accurately identify different types of clouds
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
[0054] In order to make the purpose, technical solution and advantages of the present invention more clear, the present invention will be further described below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present invention, and are not intended to limit the present invention, that is, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Therefore, based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.
[0055] Assuming that an image set to be masked has been acquired, and some areas on the image are covered by clouds, resulting in missing data, the clouds in this area need to be masked. In this embodiment, a total of 24 scenes of Landsat 5TM images, 20 scenes...
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
A sample equalization method for non-uniform categories of cloud masks, comprising: step 1: data preprocessing and sample acquisition; step 2: sample grouping; step 3: cloud mask model training; step 4: classifier mask and evaluation; step 5: iterative training; step 6: cloud mask; step 7: mask data post-processing. The sample equalization method adopted in the present invention effectively solves the sample imbalance caused by uneven cloud categories in remote sensing images, and realizes effective identification and segmentation of various types of clouds in the images, thereby improving the accuracy of cloud masks; selectively Input samples, which can highlight the impact of missing or wrongly detected samples, thereby enhancing small category samples, thereby effectively adjusting the features extracted by the deep learning model, and solving the problem of small category cloud missed detection or false detection.
Description
technical field [0001] The invention relates to a method for equalizing samples of non-uniform categories, in particular to a method for equalizing samples of non-uniform categories in the cloud mask training process based on deep learning. Background technique [0002] With the development of remote sensing technology, remote sensing images are widely used in crop mapping, forest monitoring and other fields. However, due to the influence of clouds in the imaging process of remote sensing images, there is data loss or distortion in the corresponding areas of the acquired images, which affects the accuracy of information extraction results and its application significance. The problem of data loss or distortion caused by clouds is particularly prominent in cloudy and rainy areas such as southern China. Using images partially obscured by clouds can improve data availability and enhance its support capabilities. In applying partially missing imagery, the cloud masking process,...
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): G06V10/774G06V10/764G06K9/62G06N20/00
CPCG06N20/00G06F18/241G06F18/214
Inventor 吴炜高星宇范菁沈瑛夏列钢葛炜炜
Owner ZHEJIANG UNIV OF TECH