Crop fine classification method based on high-resolution remote sensing satellite image

Abstract

The invention discloses a crop fine classification method based on a high-resolution remote sensing satellite image. The method comprises the following steps: S1, collecting the high-resolution remote sensing satellite image and field sample data in a research area; s2, preprocessing the high-resolution remote sensing satellite image and the field sample data respectively; s3, processing the preprocessed high-resolution remote sensing satellite image by adopting a superpixel segmentation algorithm to obtain a segmented remote sensing image vector diagram, and superposing the remote sensing image vector diagram onto the preprocessed high-resolution remote sensing satellite image of the research area; s4, based on the result vector diagram of the superpixel segmentation algorithm, performing fine classification on crop species by using an object-oriented algorithm; the method mainly aims at a southern complex crop planting area, adopts a high-resolution remote sensing image, is combined with unmanned aerial vehicle remote sensing, is based on a superpixel segmentation algorithm and is combined with an object-oriented classification method, and crop species can be finely classified.

Description

technical field [0001] The invention relates to the technical field of remote sensing information classification, in particular to a method for fine classification of crops based on high-resolution remote sensing satellite images. Background technique [0002] my country is a big agricultural country, and agricultural issues have always been the focus of the common concern of the government and the people. Therefore, my country's modern agricultural production is gradually developing in the direction of intensification and precision. With the transformation of agriculture, the demand for spatial information, especially large-scale, dynamic, continuous and rapid crop information, becomes more and more important in the agricultural production process. Especially for southern agriculture, where the plots are small and complex, and there are many kinds, it is relatively difficult to obtain the information of species crops quickly and accurately. Therefore, timely acquisition o...

AI Technical Summary

Problems solved by technology

However, these methods are all based on the difference in spectral information of ground objects in remote sensing images. Their algorithms ignore the texture information and geometric information of ground objects, and it is difficult to distinguish between "same object with different spectrum" and "same spectrum with different object". Therefore, its classification accuracy is relatively low

With the development of deep learning, the effect of using deep learning algorithms to classify remote sensing images is relatively high. However, it is difficult to produce massive sample data, especially the refined classification sample data sets of species.

At the same time, the applicability of the deep learning algorithm is poor. The classification results of the algorithm model are quite different for the images of the same area at different times. The algorithm is not suitable for fast and accurate extraction of crop species information.

Images