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

Method of Image Change Detection Using Deep Convolutional Neural Networks Combined with Morphology

A neural network and deep convolution technology, applied in the field of deep convolutional neural network combined with morphological detection of image changes, can solve the problems of intractability, low detection accuracy, poor visual effect, etc., to achieve high accuracy and robustness, The effect of high detection accuracy and simple method

Active Publication Date: 2022-02-22
NANJING INST OF TECH +1
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Purpose of the invention: The purpose of the present invention is to provide a method for detecting image changes with a deep convolutional neural network combined with morphology, which solves the problems of large noise, difficult processing, low detection accuracy, and poor visual effects in existing methods for detecting image changes

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
  • Method of Image Change Detection Using Deep Convolutional Neural Networks Combined with Morphology
  • Method of Image Change Detection Using Deep Convolutional Neural Networks Combined with Morphology
  • Method of Image Change Detection Using Deep Convolutional Neural Networks Combined with Morphology

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0020] The present invention will be further described below in conjunction with the accompanying drawings.

[0021] like figure 1 As shown, the method of combining deep convolutional neural network with morphology to detect image changes includes the following steps:

[0022] (1) Segment the registered remote sensing images of 2015 and 2017. Since the input image size of the improved SegNet network is an 8-channel image of 224×224, the images of 2015 and 2017 are respectively divided into 224×224 size. In order to make reasonable use of data resources, the original image is segmented by partial overlapping sliding, which can increase the amount of training data after segmentation of small remote sensing images. For example, when splitting, the coordinates of the upper left corner of the first horizontal image are (0,0), the second is (112,0), the third is (224,0) and so on, and the vertical coordinates of the upper corner are ( 0,112), (0,224) and so on. When the sample s...

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 detecting image changes with a deep convolutional neural network combined with morphology, which can segment registered remote sensing images of different time phases; rotate and mirror the segmented images, and then correspond to different time phases The remote sensing image of the location is merged into an 8-channel image; the obtained 8-channel image data is input into the SegNet network model for training, and a 2-channel image is output; the image is filled with holes by using an AND operation, and then the corrosion operation is used to remove Noise information to obtain an image processing model; the number of remote sensing images to be detected is segmented and input to the previous model for processing, and the image is output; the output image is merged into the size of the original remote sensing image to be detected, and the image change detection is completed. The invention adopts a deep convolutional neural network combined with a morphological method, which has high detection accuracy, effective denoising, simple method, and high accuracy and robustness for building change detection.

Description

technical field [0001] The invention relates to an image change detection method, in particular to a method for detecting image changes using a deep convolutional neural network combined with morphology. Background technique [0002] In recent years, with the rapid development of computer technology and artificial intelligence, land supervision has also become increasingly intelligent. The supervision of land resources is beneficial to the country's rational distribution and utilization of land resources. A major problem in land supervision is that the land resources are extremely large, and it takes a lot of manpower to conduct on-the-spot inspections and investigations in real life. Using remote sensing images to compare images in different phases can effectively find out the differences in changes in buildings in different phases, thereby realizing effective supervision of land resources. However, for remote sensing images of large areas, it takes a lot of human resourc...

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): G06T7/00G06T7/10G06T7/30G06N3/04
CPCG06T7/0002G06T7/10G06T7/30G06T2207/30181G06T2207/20084G06T2207/20081G06T2207/10032G06N3/045
Inventor 徐梦溪吴晓彬朱斌王鑫石爱业陈哲韩磊
Owner NANJING INST OF TECH