Forest illegal reclamation land parcel detection method based on deep learning target detection

A technology of deep learning and target detection, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as small scope, poor timeliness, and poor model versatility

Active Publication Date: 2021-09-10
CHANGGUANG SATELLITE TECH CO LTD
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Quantitative remote sensing model identification methods rely on the spectral characteristics of plants, and the identification effect is relatively good. However, on the one hand, the model is not universal due to the high requirements for satellite image radiation correction accuracy; / The resolution of hyperspectral imagery is low and it is difficult to accurately locate the boundaries of land parcels
The above problems have caused the problems of low accuracy and poor timeliness of traditional remote sensing interpretation methods

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
  • Forest illegal reclamation land parcel detection method based on deep learning target detection
  • Forest illegal reclamation land parcel detection method based on deep learning target detection
  • Forest illegal reclamation land parcel detection method based on deep learning target detection

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment 1

[0046] according to Figure 1 to Figure 5 As shown, the present invention provides a method for detecting illegally reclaimed land in forest areas based on deep learning target detection, comprising the following steps:

[0047] The invention develops a deep convolutional neural network that can be migrated and has strong generalization performance, and combines the feature pyramid extraction mode in the network to achieve the expression of multiple scale features of the target to be detected, and realizes the high-resolution remote sensing image in the complex environment. Illegal land reclamation identification.

[0048] First, expert knowledge is used to construct a sample dataset suitable for deep learning training. The construction principles include: (1) The illegally reclaimed forest land plots are located in dense forests and hilly forest areas, with traces of artificial reclamation, and the land area is generally small; (2) The illegally reclaimed forest land plots a...

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 relates to a forest illegal reclamation land parcel detection method based on deep learning target detection. The invention relates to the technical field of forest illegal reclamation land parcel detection, and the method comprises the steps of constructing a sample data set of deep learning training based on expert knowledge; inputting a sample image, and establishing a multi-scale feature pyramid recognition network; based on a continuous learning mode, training model fine tuning is carried out; and according to the fine-adjusted training model, applying the fine-adjusted training model to a test image, and detecting an illegal reclamation land parcel in a forest region. In order to improve the recall rate of detection model identification and greatly improve the precision of machine interpretation, the method of the invention adopts a continuous learning mode to carry out two-stage model training, and can prevent bare land parcels with similar spectrums and textures from being mistakenly classified.

Description

technical field [0001] The invention relates to the technical field of detecting illegally reclaimed land in forest areas, and relates to a method for detecting illegally reclaimed land in forest areas based on deep learning target detection. Background technique [0002] Illegal land reclamation in forest areas is mostly distributed in inaccessible mountainous areas, dense forests, and hilly areas. The plots are scattered and small in size, with strong concealment, making it very difficult to discover and report. The remote sensing monitoring of illegally reclaimed land in forest areas can quickly obtain the situation of deforestation and reclamation, and provide practical and effective help for forestry management departments to carry out forest harvesting management. [0003] At present, based on high-resolution satellite remote sensing images, the most common identification methods for illegal land reclamation in forest areas are man-machine interactive interpretation an...

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): G06K9/62G06K9/46G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/214
Inventor 李竺强朱瑞飞马经宇张国亮刘思言田德宇
Owner CHANGGUANG SATELLITE TECH CO LTD
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