Forest land change detection method based on early-stage forest land vector data

Abstract

The invention discloses a forest land change detection method based on early-stage forest land vector data, and the method comprises the steps: 1, carrying out the multi-scale segmentation of a front-stage remote sensing image and a rear-stage remote sensing image based on the early-stage forest land vector data, and obtaining image spots corresponding to the front-stage remote sensing image and the rear-stage remote sensing image; Step 2, extracting multidimensional features of the two periods of image spots, and calculating the difference degree of the image spots by adopting a feature fusion method based on a sparse representation theory; Step 3, adaptively selecting a training sample, and determining a change threshold of the difference degree between the two stages of image spots by using a Bayesian threshold determination method based on a maximum expectation algorithm; And step 4, performing binary segmentation on the weighted difference image obtained in the step 2 by using thechange threshold obtained in the step 3 to obtain a forest land change detection result. According to the method, multi-scale segmentation can be carried out on the front and rear stages of remote sensing images by using early-stage forest land vector data, so that a forest land change detection result is obtained.

Description

technical field [0001] The invention belongs to the field of remote sensing image data processing, and relates to a detection method for performing multi-scale segmentation based on previous woodland vector data, and then extracting forestland change information. Background technique [0002] With the implementation of the national sustainable development strategy, marked by the full start of the six key forestry projects, my country's forestry has entered a new stage of comprehensively promoting leapfrog development guided by the theory of sustainable development. However, the current forest resource survey and monitoring is mainly based on manual surveys on the ground. The technical means are backward, and the informatization and automation level of monitoring data collection, transmission and storage are low. It is greatly affected by human factors, and it is necessary to promote the development of forest land change detection to automation. The multi-scale segmentation ...

AI Technical Summary

Problems solved by technology

At present, multi-scale segmentation methods are mostly focused on the research of quantitative and qualitative identification of optimal parameters. Many multi-scale segmentation algorithms have been developed, but the actual application effect is not ideal, and it is difficult to apply to actual production on a large scale.

Images