森林植被遥感监测影像最佳分辨率选择

遥感影像分辨率的高低直接影响着森林植被监测的精度、成本和效率，故选择适合森林植被监测的影像最佳分辨率具有重要的应用价值。针对森林植被监测影像最佳分辨率选择方法及结果缺乏的问题，从林业实际应用出发，提出了基于1个步长的变异函数分析空间变异并综合考虑监测精度、成本和效率来确定森林植被监测影像最佳分辨率方法。基于最新的国产高分二号（GF-2）全色影像，利用1个步长的变异函数对湖南常宁洋泉镇林区3种典型分布类型森林植被进行拟合分析，初步确定适合森林植被监测的影像最低分辨率。然后对重采样形成的不同尺度多光谱影像分别进行监督分类，并对结果进行定量定性分析，结合影像成本和数据处理时间，找到适合不同类型森林植被监测的影像最佳分辨率。研究表明:不同分布类型的森林植被，适合遥感监测的影像最佳分辨率不同:①小冠幅森林植被3.2 m；②大冠幅森林植被16.0 m；③混合冠幅森林植被8.0 m。该森林植被遥感监测影像最佳分辨率确定方法和结果可为其他区域森林植被遥感监测影像最佳分辨率确定提供借鉴。

Optimal resolution selection for monitoring forest vegetation using remote sensing images

Image resolution directly affects precision, cost, and efficiency of forest vegetation extraction. To determine the best spatial resolution of remote sensing images for forest vegetation extraction, a method was proposed for optimal spatial resolution of remote sensing images when monitoring forest vegetation. Based on a panchromatic image of GF-2, a step length variation function was used to analyze forest vegetation of three distribution types considering monitoring precision, cost, and efficiency in Yangquan Town, Changning City, Hunan Province. A preliminary minimum resolution of the image suitable for forest vegetation extraction was determined. Then, through resampling of GF-2 multi-spectral images after fusion, middle and low resolution images of different scales were made to extract forest vegetation using the supervised classification method. Quantitative and qualitative analyses of results were carried out for accuracy of forest vegetation extraction, image cost, and data processing time. Results showed that the optimal spatial resolution for remote sensing monitoring images was different for different types of forest vegetation, with forest vegetation of small canopies being 3.2 m, for big canopies being 16.0 m, and for mixed canopies being 8.0 m. The optimal spatial resolution of remote sensing images to monitor forest vegetation could be used as a reference in other areas.