An iterative h-minima improved watershed extraction method for an unmanned aerial vehicle high-resolution image single tree crown

An extraction method and watershed technology, applied in the field of single tree canopy extraction from UAV high-resolution images, can solve the problems of uneven brightness change, poor tree crown extraction effect, low detection accuracy of single tree crown position, etc. Reduce the effect of under-segmentation

Inactive Publication Date: 2018-12-11
NANJING FORESTRY UNIV
View PDF0 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, in the broad-leaved forest with high canopy density, due to the uneven brightness variation inside the canopy of the broad-leaved tree, the size of the canopy is different, the canopy is connected and overlapped, the shape of the canopy is irregular, and the boundary is not obvious. Canopy extraction is poor

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
  • An iterative h-minima improved watershed extraction method for an unmanned aerial vehicle high-resolution image single tree crown
  • An iterative h-minima improved watershed extraction method for an unmanned aerial vehicle high-resolution image single tree crown
  • An iterative h-minima improved watershed extraction method for an unmanned aerial vehicle high-resolution image single tree crown

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0041] The iterative h-minima of the present invention improves the watershed UAV high-resolution image single tree canopy extraction method. First, the image is preprocessed, and then the position of the single tree canopy is detected through mutual cooperation between different h values, and the false mark detection method is used. Filter invalid marks, and finally introduce the principle of circular arc symmetry to limit the watershed submersion process, avoiding the overgrowth of tree crown marks and the merger of unmarked tree crowns.

[0042] Such as figure 1 As shown, the image is preprocessed first, the grayscale image is smoothed based on the morphological opening operation, the gradient image is calculated using the Sobel operator, and the gradient image is removed by the mean filter. The specific process includes:

[0043] 1) Reduce the intensity difference inside the canopy in the original image based on the morphological theory, and use the morphological opening o...

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 an iterative h-minima improved watershed extraction method for an unmanned aerial vehicle high-resolution image single tree crown. Firstly, images are preprocessed, then the crown positions of individual trees are detected by cooperating with different h values, and false markers are detected to filter invalid markers. Finally, the principle of arc symmetry is introduced tolimit the watershed flooding process to avoid over-growth of tree crown markers and merging of unmarked tree crown. This method can effectively restrain over-segmentation and reduce under-segmentation, and extract the crown of different tree species, different shapes, different sizes and inconspicuous boundaries in the broadleaved forest with high canopy density. The method is an effective methodto extract single tree crown from high-score images of UAV, and exhibits good practicability.

Description

technical field [0001] The invention belongs to the technical field of forestry remote sensing, and in particular relates to a single-tree canopy extraction method of a high-resolution image of an unmanned aerial vehicle that iteratively h-minima improves a watershed. Background technique [0002] Individual tree canopy information can not only be used to quantify the properties of individual trees such as canopy surface area, volume, diameter, tree height, and canopy canopy density, but also to estimate forest characteristics such as aboveground biomass and forest carbon storage. These variables are important input parameters for stand-scale modeling analysis (such as stock volume and bioenergy supply), and individual tree canopy extraction is also the first step in tree species classification, which has an important impact on tree species classification accuracy. [0003] At present, the mainstream methods for crown extraction of single trees can be divided into two catego...

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 Applications(China)
IPC IPC(8): G06T7/155G06T7/12G06T7/11G06T7/136G06T7/187
CPCG06T2207/10032G06T2207/20152G06T2207/30188G06T7/11G06T7/12G06T7/136G06T7/155G06T7/187
Inventor 温小荣腾文秀林国忠佘光辉
Owner NANJING FORESTRY UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap