Small-scale forest land range extraction method based on remote sensing image

Through multi-scale image segmentation and vector guidance technology combined with feature extraction and classification post-processing, the problems of low accuracy and excessive manual intervention in complex land processing are solved, and the classification of remote sensing image with high precision and low manual intervention are achieved.

CN120182820APending Publication Date: 2025-06-20国家林业和草原局中南调查规划院
0 Cites 0 Cited by

Patent Information

Application Number
CN202510258682.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-03-06
Publication Date
2025-06-20

AI Technical Summary

Technical Problem

Traditional remote sensing image classification methods have low accuracy when dealing with complex land boundaries and morphology, and require a lot of manual intervention, making it difficult to adapt to the application needs of large-scale and complex scenarios.

Method used

Multi-scale image segmentation and vector-guided image segmentation are used to realize image pixel spatial clustering through bottom-up methods, combining feature definition, calculation, statistics and object screening to realize object-oriented image classification, and perform classification post-processing to improve accuracy.

Benefits of technology

It improves classification accuracy, reduces manual intervention, improves processing efficiency, can meet the needs of many fields such as forest land resource management and environmental monitoring, and provides new solutions for the further development of remote sensing image processing technology.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention belongs to the field of remote sensing image processing, and particularly discloses a small-scale forest land range extraction method based on a remote sensing image, which comprises the following steps: S1, adopting multi-scale image segmentation and vector guide image segmentation to obtain a segmented pattern spot object data set; s2, obtaining feature parameters capable of describing the forest land through feature definition, feature calculation, feature statistics and object screening based on feature statistical values of the sample data; s3, realizing object-oriented image classification according to characteristic parameters capable of describing the forest land; s4, performing classification post-processing on the classification result; s5, providing batch processing according to the processed classification result to obtain a remote sensing image forest land target extraction result; according to the method, image segmentation, feature extraction, classification and post-processing are integrated, an advanced remote sensing image processing technology and a deep learning model are applied, the classification precision is improved, meanwhile, manual intervention is reduced, and the requirements of multiple fields such as forest land resource management and environment monitoring can be met.
Need to check novelty before this filing date? Find Prior Art