State-and-Evolution Detection Models: A Framework for Continuously Monitoring Landscape Pattern Change

Lingwen Tian,Meiling Liu,Ling Wu,Xiangnan Liu

doi:10.1109/tgrs.2021.3088537

Abstract

Detecting the evolution of large-area landscape patterns using long-term remote-sensing images is helpful in supporting research on the relationship between landscape patterns and ecological processes, as well as the development of ecological process simulations and spatiotemporal interaction models. However, detection methods have generally been developed as separate applications, each with a separate type of landscape pattern change; remote-sensing images are acquired at epochal timesteps. Consequently, in practical applications, many omission changes for some types of pattern changes and inaccurate evolution time are presented in the detected map. In this article, state-and-evolution detection models (SEDMs) are promoted to obtain complete information about the evolution of landscape patterns based on yearly land cover data. In the proposed framework, we first define the major categories of landscape pattern changes to comprehensively reveal the characteristics of landscape pattern changes associated with real change cases. Next, a morphological rule-based pattern recognition approach is proposed for quantitative discrimination among these categories. This approach is then applied in annual land cover data to continuously detect landscape pattern evolution processes and evolution time. Finally, the detected evolution time in different evolution processes is applied to measure the timestep between two disparate types. The performances of the SEDMs are presented by Landsat-derived land cover evolution in Shanxi, China. The detected results are indirectly verified by the land cover conversion matrix and connect index, indicating strong robustness and generalization ability of the SEDMs.

Highlights

A LANDSCAPE is a heterogeneous land area containing multiple ecosystems or a mosaic of different land use/cover [1]
An stateand-evolution detection models (SEDMs) represents changes over time in state of each cell as a stochastic continuous-time process {Xt : t ≥ 0}, where the state denotes the type of landscape pattern change (CT) at a certain phase; the state space (SS) is a set consisting of k discrete state types (Xt ∈ SS); t is discrete timesteps [Fig. 1(a)]
The validation based on land cover conversion matrix and connect index indicates that SEDMs can accurately detect CTs and their evolutions, with average overall accuracy (OA) of 78.27%, 74.90%, and 69.65% for dissection, aggregation, and creation, respectively

Summary

Introduction

A LANDSCAPE is a heterogeneous land area containing multiple ecosystems or a mosaic of different land use/cover [1]. The pattern change analysis on a small spatial scale [10]–[14] and within a short time period lacks repeatability, which cannot explain the spatial patterns and processes on larger temporal (decades or longer) and spatial scales (such as regional landscape levels and higher levels). These large-scale phenomena are important because most environmental and resource management issues occur on large/medium-scale and large-scale patterns, and processes must be linked with small-scale ones to understand nature [15]. Numerous studies on landscape pattern change analysis have emerged based on landscape cells (a group of land cover pixels within a given window) using multiple remote-sensing images that support timely and accurate land cover change information [16]–[19]

Methods

Findings

Discussion

Conclusion