基于GPS照片数据处理系统的地面样方调查

Ground Sample Survey Based on GPS Photos Processing System

集成GPS模块的相机拍摄的农作物照片可同时记录地物类型和空间位置。为了处理大量的地面样方照片数据并及时转换为遥感监督分类的训练样本,开发了一个能够高效处理样方GPS照片数据的软件系统。该系统通过人机交互判识照片中的地物类型,同时提取照片的地理位置信息,并对样本的地理位置进行修正或补充,在地图与遥感影像的辅助下生成有效的样本信息。以宁夏银川灌区农作物采样样本为例,对系统进行了测试。结果表明,1 000个典型的训练样本可在1 h内产生,减少了样本的处理时间,提高了样本的质量。

Crop photos taken by camera with integrated GPS module can mark topographic feature type & spatial position. In order to facilitate processing and converting a number of ground sample GPS photos into the training samples in support of the supervised classification with remote sensing images,a highly efficient software is developed. The human-mazhine interactive identification is used to help visually identify and interpret the type of the objects in the photo, extract and offset the longitude and latitude of the location information from the photo EXIF attribute and finally output the resulting valid sample data with the basic maps and remote sensing images. The system is tested that the samples were collected in the Ningxia Northern irrigation area are processed. The result shows that one thousand samples can be processed within one hour, thus the processing time of the sample is reduced and the sample quality is improved.