Self-adaptive remote sensing image pixel missing value filling method and device, equipment and medium

Abstract

The invention provides a self-adaptive remote sensing image pixel missing value filling method and device, equipment and a medium. The method comprises the following steps: S1, constructing a time sequence data set by using images in the same research area according to a time reverse order; s2, performing marking according to the pixel quality band by band, and recording the position information of the pixel to be filled; s3, for each to-be-filled pixel, judging whether the pixel value of the image at the same position in the past unit time is valid or not band by band; s4, judging whether thevalues of all effective pixels at the moment are close to the value of the central pixel or not in a window threshold list at the pixel position; s5, judging whether the pixel at the position corresponding to the moment to be filled is valid or not; and S6, constructing a linear regression equation according to the effective pixel value within the range of the threshold list of the window at themoment to be filled and the effective pixel value meeting the condition moment, substituting the pixel value meeting the condition moment into the pixel value calculated by the linear regression equation, filling the pixel value back to the corresponding position, and marking that the pixel is successfully filled.

Description

technical field [0001] The present invention relates to the technical field of processing methods of time series remote sensing images, in particular to an adaptive remote sensing image pixel missing value filling method, device, equipment and medium. Background technique [0002] Due to the objective existence of clouds and fog and satellite detection and communication transmission failures, there are invalid and missing values ​​in optical remote sensing images. These invalid and missing values ​​need to be filled in the application. The existing methods mainly include: [0003] (1) If the entire row is missing, it can be filled with the average value of the upper and lower rows; [0004] (2) Missing individual pixels can be filled with the average value within a window; [0005] (3) Use time series to smooth the pixel value. [0006] Obviously, these methods have great limitations in application. [0007] The average value of method (1) and method (2) is only a simple...

AI Technical Summary

Problems solved by technology

[0008] Method (3) is often used when smoothing the vegetation index of the time series, and the reflectance value of the band does not change with time to present a peak-valley curve. This method cannot effectively fill the reflectance value of the multi-spectral band

Images