【论文学习二】Optimum Features Selection for oil Spill Detection in SAR Image

目录

## Summary

## Research Objective

## Problem Statement

## Method(s)

## Evaluation

## Conclusion

---

## Summary

## Research Objective

In this paper, an algorithm is presented for selecting an optimum set of features from SAR images; which maximizes the discrimination between oil spills and their lookalikes in such images.

## Problem Statement

Generally, the oil spill detection methods use randomly selected statistical qualitative and quantitative features to classify the dark objects in SAR images into oil spills and lookalikes; and there is no agreement among the researchers as to the set of features that should be used to differentiate between oil spills and lookalikes.

（一般来说，溢油检测方法采用随机选择的统计定性和定量特征，将SAR图像中的黑色物体归类为溢油和相似物;研究人员对用来区分溢油和相似物的特征也没有达成一致。）

## Method(s)

The proposed algorithm for optimum feature selection in SAR images is shown as a block diagram in Fig. 1. As shown in Fig. 1, this algorithm consists of several steps; first, windows containing dark spots are extracted from the original SAR image. The separated images are then sent to the dark spot detection block, and their dark spots are detected. Afterwards, in the feature extraction block, the desired features are extracted based on the previously detected  spots and their surrounding regions. When the desired features of the objects in the SAR image are extracted, they are sent to the feature selection step for further processing. The classification rate is determined based on this selected set,and the result is evaluated and discussed to validate the selected set of features.

本文提出的SAR图像特征优化选择算法如图1所示。如图1所示，该算法由多个步骤组成;首先，从原始SAR图像中提取含有黑点的窗口。然后将分离的图像发送到黑点检测块，并对其黑点进行检测。然后，在特征提取块中，根据之前检测到的斑点及其周围区域提取去分类特征。提取出SAR图像中目标的期望特征后，将其发送到特征选择步骤进行进一步处理。根据所选择的特征集确定分类率，并对结果进行评估和讨论，以验证所选择的特征集。

 

 

 

## Evaluation

## Conclusion

strong conclusion:

weak conclusion：

 ## Notes