Remote sensing image object level change detection method

Abstract

The invention discloses a remote sensing image object level change detection method. The method specifically comprises the steps of image preliminary processing, mean filtering obtaining calculation and image analysis based on a convolutional neural network. The invention provides a self-adaptive k value calculation method, and provides a remote sensing image object level change detection method.The preliminary graph is acquired on the basis of the deep full convolutional neural network model as a whole, and the image is finally subjected to binarization processing on the basis of a corresponding algorithm, so that digital processing of the image is realized, later change detection can be conveniently carried out on the image on the basis of the digital processing of the image, and the detection precision and accuracy are improved.

Description

Technical field [0001] The present invention relates to the field of image processing, in particular to a method for detecting changes in the level of remote sensing image objects, which is mainly used for grayscale and binarization processing of traditional remote sensing images. Background technique [0002] The emergence of remote sensing technology has changed the way humans observe the earth; the acquisition of reproducible remote sensing data has given humans a new way to periodically understand the dynamic changes of the wide-area surface environment. With the development of remote sensing technology, the spatial resolution of optical remote sensing images has been improved from the original 100-meter level to the meter level or even higher, and the revisit period has been greatly shortened. With the development of image technology, geographic object-oriented image analysis methods that target the use of spatial information in remote sensing images have been rapidly develo...

AI Technical Summary

Problems solved by technology

With the improvement of remote sensing images from medium and low resolution to sub-meter high resolution, the image characteristics of ground objects have changed significantly, and the information has become more abundant. It is difficult to fully describe many complex ground objects on high-resolution images with a single spectral feature

However, the data analysis and processing capabilities of remote sensing in complex environments have not been improved accordingly, and most of the ever-growing massive remote sensing data have not been effectively analyzed and utilized.

Images