Breeding pond extraction method based on remote sensing time-space spectrum feature fusion

Abstract

The invention provides a breeding pond extraction method based on remote sensing time-space spectrum feature fusion, which is characterized by comprising the following steps: 1) performing water body feature and gradient calculation on a remote sensing image scene by scene, and extracting water body and edge features of a single scene; 2) carrying out probability synthesis on the water body and edge features extracted from the scene-by-scene image to obtain initial breeding thematic information; and 3) performing post-processing on the initial breeding thematic information by using the existing terrain, building area and coastline thematic data to obtain a high-precision breeding pond thematic product. The method is mainly characterized in that time sequence features and spectral features and spatial features of remote sensing images are comprehensively utilized, the breeding pond with specific water body features is distinguished from easily-mixed paddy fields, salt pans, rivers and other targets, and therefore high-precision extraction of the breeding pond is achieved.

Description

technical field [0001] The invention belongs to the technical field of digital image processing, and in particular relates to a method for extracting remote sensing breeding ponds. Background technique [0002] With the increase of global population and economic development, the human demand for food is also increasing, especially high-protein fish. Aquaculture is becoming an important player in the food production sector, especially in developing Asia. Statistics from the International Food and Agriculture Organization (FAO) show that in 2014, fish catches from aquaculture exceeded capture catches for the first time, and in 2016 they accounted for 53 percent of all fish catches. However, the vigorous development of aquaculture industry also faces many environmental problems. [0003] Breeding ponds account for a large proportion of the aquaculture industry, but there are problems such as encroaching on coastal mangroves, competing with farmland for water resources, pollut...

AI Technical Summary

Problems solved by technology

[0004] There are mainly three problems in the existing remote sensing extraction methods for aquaculture ponds: (1) methods of convolutional neural networks (Gusmawati N, Soulard B, Selmaoui-Folcher N, et al.Surveying shrimp aquaculture pond activity using multitemporal VHSR satellite images- case study from the Perancak estuary, Bali, Indonesia[J].Marine Pollution Bulletin, 2017, 131pb(jun.):49-60.) It is necessary to draw a large number of samples of different species of breeding areas, which is costly

But as far as the current research status is concerned (DuanY, Li X, Zhang L, et al. Mapping national-scale aquaculture ponds based on the Google Earth Engine in the Chinese coastal zone [J]. Aquaculture, 520.), this type of method does not Time-series image features are not effectively used, and it is difficult to distinguish them from rice fields and salt fields, and it is not conducive to the problem of missing regional data caused by cloud coverage of optical images during global exploration.

Images