Unlock instant, AI-driven research and patent intelligence for your innovation.

A single image cloud and fog removal method based on dual-tree complex wavelet transform

A dual-tree complex wavelet, single-image technology, applied in the field of remote sensing image processing, can solve the problem of inflexible threshold selection, and achieve the effect of good true color, less manual intervention, and strong adaptability

Active Publication Date: 2019-10-25
江苏智行未来汽车研究院有限公司
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the threshold selection is not flexible enough, and there are obvious cloud boundary after processing

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
  • A single image cloud and fog removal method based on dual-tree complex wavelet transform
  • A single image cloud and fog removal method based on dual-tree complex wavelet transform
  • A single image cloud and fog removal method based on dual-tree complex wavelet transform

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0073] use as image 3 A shown color remote sensing picture containing clouds and fog provides the application method of the present invention. The number of rows and columns of the image is 512, and the central part of the image contains clouds with high density. The scenery under the clouds is difficult to distinguish, while the scenery outside the cloud area is relatively clear.

[0074] First, according to formula (10), the number of decomposition layers n is calculated to be 7. Use qshift_06 as filter function pair image 3 Do 7-layer dual-tree complex wavelet transform, the integer l is taken as 5, the reconstruction diagrams of the low-level high-frequency sub-band, high-level high-frequency sub-band, and low-frequency sub-band are as follows Figure 4 , 5 , 6, it can be seen that the three sub-bands all contain a certain amount of clouds, but the high-level high-frequency sub-bands contain most of the clouds. Figure 4 It shows that the low-level high-frequency sub...

Embodiment 2

[0082] like Figure 13 A gray-scale remote sensing picture is shown, in which multiple regions contain thin clouds with different concentrations. The following describes the process of processing the picture by the method of the present invention.

[0083] First, the image size is 256×256, therefore, the number of decomposition layers n is 6. Use qshift_06 as filter function pair Figure 13 Do 6-layer dual-tree complex wavelet transform, the integer l is taken as 4, and the original image is divided into three parts: low-level high-frequency sub-bands, high-level high-frequency sub-bands, and low-frequency sub-bands.

[0084] Then, the Laplacian filter is used to process the low-level high-frequency subbands of layers 1 to 1 respectively to enhance the ground scene information. All processed low-level high-frequency sub-bands are combined, and the reconstructed image obtained is further normalized according to formula (6) through dual-tree complex wavelet inverse transformat...

Embodiment 3

[0090] like Figure 17 Shown is a remote sensing image with uniform cloud coverage, the process of processing the image using the method of the present invention will be described below.

[0091] First, the image size is 512×512, therefore, the number of decomposition layers n is 7. Use qshift_06 as filter function pair Figure 17Do 7-layer dual-tree complex wavelet transform, the integer l is taken as 5, and the original image is divided into three parts: low-level high-frequency sub-bands, high-level high-frequency sub-bands, and low-frequency sub-bands.

[0092] Then, the Laplacian filter is used to process the low-level high-frequency subbands of layers 1 to 1 respectively to enhance the ground scene information. All processed low-level high-frequency sub-bands are combined, and the reconstructed image obtained is further normalized according to formula (6) through dual-tree complex wavelet inverse transformation.

[0093] Next, use the inverse cloud thickness weighted ...

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

No PUM Login to View More

Abstract

The invention discloses a method for removing clouds and fog from a single image based on dual-tree complex wavelet transform. The dual-tree complex wavelet transform is used to decompose the original image, and allocate clouds and scenery to high-level and low-level high-frequency subbands and low-frequency subbands, respectively. The sub-bands are processed separately to remove clouds and fog. The steps of the present invention include: first performing multi-layer dual-tree complex wavelet transformation on the original image, dividing the decomposition coefficient into low-level high-frequency subbands, high-level high-frequency subbands and low-frequency subbands; using Laplacian filter to process layer by layer For low-level high-frequency subbands, double-tree complex wavelet inverse transform is performed, and then normalized to reconstruct the image to enhance ground scene information; high-level high-frequency subbands and low-frequency subbands are processed respectively according to the inverted cloud thickness weighting method. The invention processes a single remote sensing image containing uniform cloud coverage, dense clouds or thin clouds, while weakening the cloud coverage, improving the clarity of the scenery under the clouds, effectively retaining and highlighting the scenery information outside the cloud area, and has adaptive properties. The advantages are strong and less manual intervention.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image processing, and relates to an image enhancement method for removing clouds and fog from a single remote sensing image. Background technique [0002] Clouds and fog are common sources of interference in the process of satellite-to-earth remote sensing imaging. Cloud and fog interference often reduce the contrast of remote sensing images, blur the image, and lose part of the ground object information. Methods such as multispectral image declouding, multi-image stacking and multi-sensor image fusion declouding have been widely used in cloud and fog image processing, and the effect is obvious, but they all require multiple images from the same region and the same source, and the cost of data acquisition is high. , the cycle is long. In contrast, the single image cloud removal method only uses one remote sensing image, the cost of data collection is low, and the processing results can be ...

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
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06T5/00
CPCG06T2207/30181G06T2207/20064G06T2207/20024G06T2207/10032G06T5/73
Inventor 吴峰朱锡芳相入喜许清泉吴涛
Owner 江苏智行未来汽车研究院有限公司