Graph – Based High Resolution Satellite Image Segmentation for Object Recognition

Object based image processing and analysis is challenging research in very high resolution satellite utilisation. Commonly ei ther pixel based classification or visual interpretation is used to recognize and delineate land cover categories. The pixel based classification techniques use rich spectral content of satellite images and fail to utilise spatial relations. To overcome th is drawback, traditional time consuming visual interpretation methods are being used operational ly for preparation of thematic maps. This paper addresses computational vision principles to object level image segmentation. In this study, computer vision algorithms are developed to define the boundary between two object regions and segmentation by representing image as graph. Image is represented as a graph G (V, E), where nodes belong to pixels and, edges (E) connect nodes belonging to neighbouring pixels. The transformed Mahalanobis distance has been used to define a weight function for partition of graph into components such that each component represents the region of land category. This implies that edges between two vertices in the same component have relatively low weights and edges between vertices in different components should have higher weights. The derived segments are categorised to different land cover using supervised classification. <br><br> The paper presents the experimental results on real world multi-spectral remote sensing images of different landscapes such as Urban, agriculture and mixed land cover. Graph construction done in C program and list the run time for both graph construction and segmentation calculation on dual core Intel i7 system with 16 GB RAM, running 64bit window 7.

Download Full-text

Development of segmentation algorithm based region growing for software GIS GRASS

Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis ◽

10.11118/actaun201058010207 ◽

2010 ◽

Vol 58 (1) ◽

pp. 207-216

Author(s):

Aleš Tippner

Keyword(s):

Image Segmentation ◽

High Resolution ◽

Land Cover ◽

Mathematical Morphology ◽

Satellite Images ◽

Satellite Image ◽

Region Growing ◽

Segmentation Algorithm ◽

Aerial Photographs ◽

High Resolution Satellite Image

Image segmentation is fundamental prerequisite for new satellite images interpretation methods. GIS GRASS provides segmentation tools enabling global image segmentation only. We designed procedure enabling local segmentation using existing GRASS tools and segmentation algorithm based on region growing that we developed with C++. This algorithm applies mathematical morphology operators to output segments, too. Principial aim of the project is creation of useful input for differentiation of base land cover classes in panchromatic high-resolution satellite image (or historical aerial photographs for example).

Download Full-text

Land Cover Classification of Nine Perennial Crops Using Sentinel-1 and -2 Data

Remote Sensing ◽

10.3390/rs12010096 ◽

2019 ◽

Vol 12 (1) ◽

pp. 96 ◽

Cited By ~ 6

Author(s):

James Brinkhoff ◽

Justin Vardanega ◽

Andrew J. Robson

Keyword(s):

Land Cover ◽

Large Scale ◽

Satellite Image ◽

Resource Planning ◽

Support Vector ◽

Perennial Crop ◽

Perennial Crops ◽

Object Based ◽

Rbf Kernel

Land cover mapping of intensive cropping areas facilitates an enhanced regional response to biosecurity threats and to natural disasters such as drought and flooding. Such maps also provide information for natural resource planning and analysis of the temporal and spatial trends in crop distribution and gross production. In this work, 10 meter resolution land cover maps were generated over a 6200 km2 area of the Riverina region in New South Wales (NSW), Australia, with a focus on locating the most important perennial crops in the region. The maps discriminated between 12 classes, including nine perennial crop classes. A satellite image time series (SITS) of freely available Sentinel-1 synthetic aperture radar (SAR) and Sentinel-2 multispectral imagery was used. A segmentation technique grouped spectrally similar adjacent pixels together, to enable object-based image analysis (OBIA). K-means unsupervised clustering was used to filter training points and classify some map areas, which improved supervised classification of the remaining areas. The support vector machine (SVM) supervised classifier with radial basis function (RBF) kernel gave the best results among several algorithms trialled. The accuracies of maps generated using several combinations of the multispectral and radar bands were compared to assess the relative value of each combination. An object-based post classification refinement step was developed, enabling optimization of the tradeoff between producers’ accuracy and users’ accuracy. Accuracy was assessed against randomly sampled segments, and the final map achieved an overall count-based accuracy of 84.8% and area-weighted accuracy of 90.9%. Producers’ accuracies for the perennial crop classes ranged from 78 to 100%, and users’ accuracies ranged from 63 to 100%. This work develops methods to generate detailed and large-scale maps that accurately discriminate between many perennial crops and can be updated frequently.

Download Full-text