Method and device for synthesizing feature image time series based on multi-source middle division

Abstract

The invention discloses a method for synthesizing time series of characteristic images based on multi-source central division, comprising the steps of: using GF1-WFV, HJ1-A / B and Landsat to respectively collect remote sensing images to form multi-source central division remote sensing images; Preprocess the remote sensing images and output radiation-normalized time series feature images; extract the cloud coverage area in the time series feature images and divide the cloud-free time series feature images, and output the feature image time series empty tile set; collect MODIS The time-series feature images of several types of crops were extracted from the single-crop growth curves to generate a multi-crop growth curve library; the time-series empty tile set and the single-crop growth curve were added to the radiation-normalized time-series feature images for time phase Normalize to construct feature image time series tile sets. The present invention completes the construction of time series of characteristic images with large scale, long time series and large amount of data.

Description

technical field [0001] The present invention relates to the technical field of remote sensing image processing, and more specifically, the present invention relates to a method and device for synthesizing time series of characteristic images based on multi-source middle division. Background technique [0002] Crop planting area is an important basis for formulating grain production policies and determining the quantity of grain trade, and is the basis for the adjustment of agricultural planting structure. The growth of crops has its own laws, and the same crop has a relatively stable growth and development law in the same area. According to the differences in time and biomass in the development process of different crops, combined with some phenological information, time can be used. Sequence remote sensing images classify and identify different crops, so as to achieve the purpose of extracting crop planting information. In this context, the time series data of remote sensi...

AI Technical Summary

Problems solved by technology

Such a time interval will cause the characteristic index to fail to reflect the crop growth status in the same period, and generally cause significant differences in crop growth in different regions of the study area.

In addition, since the resolution of feature images has been significantly improved compared with traditional time series data, the increase in data volume is also huge, especially large-scale (provincial) time series data will inevitably cause a sharp increase in data volume. Taking Jiangxi Province as an example, using HJ satellite data to cover the entire area in a single phase requires 6.5G of data for 5 scenes, and the crop growth period is set as three months. At least 9 periods of data are required to construct a ten-day time series during the growth period. When synthesizing according to the traditional band combination method, at least 100G or more characteristic image time-phase sequence data will be generated. Ordinary stand-alone machines cannot perform follow-up analysis and calculation on such a huge amount of data.

Images