遥感技术在毒草识别中的研究进展

毒草的滋生蔓延严重破坏草地生境，制约草地畜牧业的发展。遥感技术作为牧场管理的一种重要的技术手段，其传感器自身的空间分辨率和光谱分辨率的高低是决定毒草识别成功与否的关键。于毒草独特的物候特征出现时获取影像数据能帮助提高分类识别的精度。回顾了3种遥感技术在毒草识别中的研究进展。航空摄影成本高、数据处理复杂，难于得到推广 ；多光谱卫星遥感大多空间分辨率低，仅在识别大面积滋生、密度较大的毒草方面展现出了一定的潜力 ；高光谱遥感的出现改善了对植被分类识别的精度，是未来毒草识别的主要依据。由于高光谱数据的冗余性和复杂性，数据处理技术和分类方法的选择也是影响毒草识别精度的重要因素。

A Review of Poisonous Weeds Detection Using Remote Sensing Technology

The rapid spread of poisonous weeds often causes serious damage to grassland habitats,and limits the development of animal husbandry. Remote sensing technology offers the advantage of efficient natural resource investigation than ground survey,and it has become one of important techniques in pasture management. The spatial resolution and spectral resolution of the sensors are the key factors which will determine the ability of detecting poisonous weeds. Acquiring imagery data at proper phenological stage will help to improve the accuracy of discrimination. This paper reviewed three types of remote sensing technology in detecting poisonous weeds. Aerial photography has not been widely used because of its high cost and complex data processing. Most current multispectral satellite sensors have relatively coarse spatial resolution and only show some potential in detecting dense and large areas of poisonous weeds. The emergence of hy-perspectral remote sensing has improved the accuracy of vegetation classification and identification,and it will be the main basis for high accuracy poisonous weed detecting and mapping in the future. By reason of the redundancy and complexity of the hyperspectral remote sensing data,the methods of data handling and classification will be challenges which will influence the accuracy of detecting poisonous weeds.