研究目的
To distinguish target points from background points in SAR images using a salient seed extraction based method.
研究成果
The proposed salient seed extraction based method achieves competitive performance in terms of detection rate and improves false alarm rate as well as computational efficiency compared to state-of-the-art methods.
研究不足
The method's performance is sensitive to the parameters step and searching size controlling parameter, although it shows stability within certain ranges.
1:Experimental Design and Method Selection:
The proposed method employs a salient point to region scheme, starting with salient seed extraction via mean-shift and region feature based approach, followed by pixel assignment to the most similar seed, and finally CFAR operation for target detection.
2:Sample Selection and Data Sources:
Single look HH polarization TerraSAR-X images from Boneyard in Davis, with 0.5 m resolution in both range and azimuth directions, and the image size is 728 and 784 pixels in the range and azimuth directions respectively.
3:5 m resolution in both range and azimuth directions, and the image size is 728 and 784 pixels in the range and azimuth directions respectively.
List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: Not explicitly mentioned.
4:Experimental Procedures and Operational Workflow:
Initialization of seeds, mean-shift based seed selection, region feature based salient seed extraction, pixel assignment, and CFAR detection.
5:Data Analysis Methods:
Detection rate (DR) and false alarm rate (FAR) are used for quantitative comparison with state-of-the-art methods.
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容