Unsupervised Segmentation Evaluation Using Area-Weighted Variance and Jeffries-Matusita Distance for Remote Sensing Images

Yongji Wang,Ying Liu,Qingwen Qi

doi:10.3390/rs10081193

Abstract

Image segmentation is an important process and a prerequisite for object-based image analysis. Thus, evaluating the performance of segmentation algorithms is essential to identify effective segmentation methods and to optimize the scale. In this paper, we propose an unsupervised evaluation (UE) method using the area-weighted variance (WV) and Jeffries-Matusita (JM) distance to compare two image partitions to evaluate segmentation quality. The two measures were calculated based on the local measure criteria, and the JM distance was improved by considering the contribution of the common border between adjacent segments and the area of each segment in the JM distance formula, which makes the heterogeneity measure more effective and objective. Then the two measures were presented as a curve when changing the scale from 8 to 20, which can reflect the segmentation quality in both over- and under-segmentation. Furthermore, the WV and JM distance measures were combined by using three different strategies. The effectiveness of the combined indicators was illustrated through supervised evaluation (SE) methods to clearly reveal the segmentation quality and capture the trade-off between the two measures. In these experiments, the multiresolution segmentation (MRS) method was adopted for evaluation. The proposed UE method was compared with two existing UE methods to further confirm their capabilities. The visual and quantitative SE results demonstrated that the proposed UE method can improve the segmentation quality.

Highlights

With the rapid development of high-resolution satellite sensor technology, the phenomenon often occurs that different geo-objects have the same spectral reflectance, or the same geo-objects have different spectral reflectance in a high spatial resolution (HSR) remote sensing image, resulting in the poor performance of traditional pixel-based image analysis in HSR images
A unsupervised evaluation (UE) method of the weighted variance (WV) and JM distance measures based on local measure criteria has been proposed for evaluating segmentation quality, and the JM distance is improved by considering the contribution of the common border between adjacent segments and the area of each segment to make the heterogeneity measure more effective and objective
The proposed UE method is compared with the Zhang and Espindola methods to further demonstrate the effectiveness of the proposed UE method

Summary

Introduction

With the rapid development of high-resolution satellite sensor technology, the phenomenon often occurs that different geo-objects have the same spectral reflectance, or the same geo-objects have different spectral reflectance in a high spatial resolution (HSR) remote sensing image, resulting in the poor performance of traditional pixel-based image analysis in HSR images. The image segmentation process, which partitions a remote sensing image into spatially contiguous and spectrally homogeneous regions [9], is commonly considered to be a prerequisite for GEOBIA because GEOBIA performance is directly affected by the segmentation quality. The evaluation of segmentation quality is considered to be important for GEOBIA in determining optimal scales and obtaining effective segmentation results for subsequent analysis

Methods

Discussion

Conclusion