Scale characteristics of remote sensing information: a case of shoreline
25 Jul 2005-Vol. 8, pp 5383-5385
TL;DR: In this paper, a series of satellite remotely sensed images with different resolutions are adapted to study the scale characteristics of remote sensing, and the results show that the natural shoreline is a representative fractal object, and its length is dependent on the scale or the remote sensing resolution.
Abstract: The scale issue is the most important item in the remote sensing application. The scale affects the information contents of the remote sensing imagery. Usually, the scale is determined by the sensor resolution. In this paper, a series satellite remotely sensed images with different resolutions are adapted to study the scale characteristics of remote sensing. The results show that the natural shoreline is a representative fractal object, and its length is dependent on the scale or the remote sensing resolution. The scale fractal characteristics are the important to monitoring of shoreline. The length at a special scale can be calculated from property remote sensing images based on the study, which is meaningful for coastal remote sensing surveying.. I. INTRODUCTION The scale issue is the most important item in the remote sensing application. The scale affects the information contents of the remote sensing imagery. Usually, the scale of remote sensing information is determined by the sensor resolution. The results of remote sensing monitoring for the same object with different resolution images are discrepant. For example, the length of shoreline is related to the measure scale, here is the resolution of sensor. The shoreline derived from high resolution image is longer than that derived from lower resolution image.
Citations
More filters
01 Jul 2006
TL;DR: The purpose of the study is to find the quantitative relationship between remote sensing information and resolution, and then to develop a scaling model for remote sensing shoreline in order to get more accurate measurement.
Abstract: Spatial scale is one of the fundamental problems in geosciences. The spatial scale of remote sensing information and scaling have become the important study area of many scientific studies with the applications of remote sensing data. Usually, the scale of remote sensing image is determined by the sensor's resolution, and the resolution affects the information contents of the remote sensing imagery. The purpose of the study is to find the quantitative relationship between remote sensing information and resolution, and then to develop a scaling model for remote sensing shoreline in order to get more accurate measurement. In this paper, the spatial scale of shoreline information extracted from satellite images with different resolutions has been studied based on fractal geometry. A spatial scaling model has been developed. The model has been used to scale the shorelines spatial scale extracted from IKONOS, IRS, SPOT and ETM+ images to 1:10000 and 1:50000 scales. The results have been compared with standard topographic map scales. It is shown that the model works well.
References
More filters
TL;DR: A brief introduction to fractals is provided and how they can be used by ecologists to answer a variety of basic questions about scale, measurement and hierarchy in, ecological systems is reported on.
Abstract: Fractal models describe the geometry of a wide variety of natural objects such as coastlines, island chains, coral reefs, satellite ocean-color images and patches of vegetation. Cast in the form of modified diffusion models, they can mimic natural and artificial landscapes having different types of complexity of shape. This article provides a brief introduction to fractals and reports on how they can be used by ecologists to answer a variety of basic questions, about scale, measurement and hierarchy in, ecological systems.
438Â citations
TL;DR: It is found that a number of important structural features of the intact old-growth landscape do not occur in the disturbed landscape, and significant landscape heterogeneity in this glaciated region is produced by landforms alone, without natural or human disturbances.
Abstract: We used geographic information systems (GIS) to analyze the structure of a second-growth forest landscape (9600 ha) that contains scattered old-growth patches. We compared this landscape to a nearby, unaltered old-growth landscape on comparable land- forms and soils to assess the effects of human activity on forest spatial pattern. Our objective is to determine if characteristic landscape structural patterns distinguish the primary old- growth forest landscape from the disturbed landscape. Characteristic patterns of old-growth landscape structure would be useful in enhancing and restoring old-growth ecosystem functioning in managed landscapes. Our natural old-growth landscape is still dominated by the original forest cover of eastern hemlock (Tsuga canadensis), sugar maple (Acer saccharum), and yellow birch (Betula allegheniensis). The disturbed landscape has only scattered, remnant patches of old-growth ecosystems among a greater number of early successional hardwood and conifer forest types. Human disturbances can either increase or decrease landscape heterogeneity depending on the parameter and spatial scale examined. In this study, we found that a number of important structural features of the intact old-growth landscape do not occur in the dis- turbed landscape. The disturbed landscape has significantly more small forest patches and fewer large, matrix patches than the intact landscape. Forest patches in the fragmented landscape are significantly simpler in shape (lower fractal dimension, D) than in the intact old-growth landscape. Change in fractal dimension with patch size, a relationship that may be characteristic of differing processes of patch formation at different scales, is present within the intact landscape but has been obscured by human activity in the disturbed landscape. Important ecosystem juxtapositions of the old-growth landscape, such as hem- lock with lowland conifers, have been lost in the disturbed landscape. In addition, significant landscape heterogeneity in this glaciated region is produced by landforms alone, without natural or human disturbances. The features that distinguish disturbed and old-growth forest landscape structure that we have described need to be examined elsewhere to determine if such features are char- acteristic of other landscapes and regions. Such forest landscape structural differences that exist more broadly could form the basis of landscape principles to be applied both to the restoration of old-growth forest landscapes and the modification of general forest man- agement for enhancing biodiversity. These principles may be particularly useful for con- structing integrated landscapes managed for both commodity production and biodiversity protection.
369Â citations
TL;DR: This paper summarizes the state of the art and introduces several updated developments in analysis and description of patch patterns and patch dynamics by means of Mandelbrot’s fractal analysis, with an emphasis on current research results and a personal view.
194Â citations
01 Jan 2000
TL;DR: In this article, the pattern of the vegetated landscape in Beijing was analyzed using fractals and the GIS software ARC/ INFO, and significant positive linear correlations were found between fractal dimensions calculated using two different methods, and between each of the two fractal dimension and average patch size.
Abstract: The pattern of the vegetated landscape in Beijing was analyzed using fractals and the GIS software ARC/ INFO. Patch complexity was examined for a range of defined vegetation types. Significant positive linear correlations were found between fractal dimensions calculated using two different methods, and between each of the two fractal dimensions and average patch size. Fractal dimensions for the types within a type in the higher level are more similar than those not within a type in the higher level. Two domains of scale were identified: those larger and smaller than about 2. 7km2. Patches in the large domain are more complex than those in the smaller domain. The main reason for which is that large patches are of ten inlaid with smaller ones while this phenomenon rarely occurs among smaller patches.
6Â citations