Correction Method of Land Use Classification Results of Remote Sensing Images

Abstract

The invention relates to the field of image processing, and provides a method for correcting land use classification results of remote sensing images. The method includes defining a solution space, organizing the solution space by a method suitable for searching, and traversing the solution space by using depth priority. The technical solution proposed by the invention is more conducive to the correction of categories with low classification accuracy, and the accuracy correction operates in a direction that is beneficial to automatic identification with low classification accuracy, so it is more conducive to balancing the classification accuracy of various object categories.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a correction method for land use classification results of remote sensing images. Background technique [0002] Land use usually refers to the different attributes of different lands, for economic and social purposes, using various scientific and technological means to carry out periodic or long-term various activities on the land to meet human needs. Land Use and Land Cover (LULC) change research is the primary consideration for scientific management and decision-making in the sustainable use of resources and environmental protection. Using remote sensing data to monitor land use and land cover changes, the object-oriented classification result data based on fuzzy mathematics has the following shortcomings: [0003] ⑴Influenced by the phenomenon of "same object with different spectrum, different objects with same spectrum", there are many mixed pixels at the boundary of the obje...

AI Technical Summary

Problems solved by technology

[0003] ⑴Influenced by the phenomenon of "same object with different spectrum, different objects with same spectrum", there are many mixed pixels at the boundary of the object category, which makes the segmentation boundary of some object types unsatisfactory, and the results extracted by the algorithm are often inaccurate;

[0004] ⑵The phenomenon of "salt and pepper noise" exists, and the existence of broken and small objects is out of date, resulting in sporadic objects with small distribution areas on the same ground feature in the same area;

There are often shadows on the shady side of the mountains, and the classification is not consistent with the normal situation in terms of brightness value and average value, which affects the effect of automatic classification;

[0006] ⑷For large-scale automatic classification, it is inevitable that there will be some misclassification or omission

[0007] At present, in order to improve the classification progress, only the classification method can be changed

Images