基于FbSP optimizer分割工具下的矿区土地利用分类研究—以济宁2号矿区为例

Classification of Land Use in Mining Areas Based on FbSP Optimizer Segmentation Tool—Setting No.1 Mining Area in Jining City as an Example

摘要:

影像分割是分类的基础，分割结果的好坏直接影响分类结果的精度。该文基于一种通过训练过程和模糊逻辑分析，确定最优分割参数的分割工具——Fuzzybased Segmentation Parameter optimizer (FbSP optimizer)来确定分割参数，并借助面向对象分类软件eCognition，以高分2号影像为基础，进行矿区土地利用的分类研究。结果表明，利用该工具不仅可以快速确定土地的最优分割尺度，同时结合eCognition，可较高精度地对土地利用进行分类。

Abstract:

Image segmentation is the basis of classification, and the result of segmentation has a direct impact on the accuracy of classification results. In this paper, a segmentation tool FbSP optimizer based on the training process and fuzzy logic analysis is proposed to determine the optimal segmentation parameters. The classification of land use in mining area based on the score 2 resolution image using the objectoriented classification software eCognition. It is showed that this tool can not only quickly determine the optimal scale of land use, but also can be used to classify land use according to the eCognition with high accuracy.