论文部分内容阅读
提出一种软硬变化检测的作物识别方法 SHLUCD(Soft and Hard Land Use/Cover Change Detection Method)。该方法利用多期遥感影像能够有效表达作物的生长物候特征,以达到在离散变化区(即纯净像元区,包括完全转换成作物的突变区域和非作物区域)和连续变化区(即渐变区,混合像元区,是部分转化为作物的区域)准确进行作物的识别。在北京市选择一个研究区,以冬小麦为研究对象,选用2011年10月6日(播种期)和2012年4月16日(拔节期)两期环境减灾1号卫星影像,分别采用硬变化检测方法 HLUCD(Hard Land Use/Cover Change Detection Method)、软变化检测方法 SLUCD(Soft Land Use/Cover Change Detection Method)和SHLUCD进行冬小麦的识别。实验结果表明:在不同尺度窗口下,SHLUCD较传统方法表现出较明显的优势,具有更低的均方根误差RMSE(SHLUCD为[0.14,0.07],HLUCD为[0.15,0.07],SLUCD为[0.16,0.08])和偏差bias(SHLUCD为-0.0008,HLUCD为-0.007,SLUCD为0.014)和更高的决定系数R2(SHLUCD为[0.68,0.86],HLUCD为[0.62,0.86],SLUCD为[0.60,0.86])。针对冬小麦突变区域、冬小麦渐变区域和非冬小麦区域分别进行评价,表明SHLUCD识别精度接近各区最佳的识别方法,进一步验证了SHLUCD的灵活性和适用性。SHLUCD方法在离散变化区能够通过土地覆盖类型状态变化来有效地识别出冬小麦,在连续变化区可识别出土地覆盖的状态变化程度定量表达冬小麦的丰度,是其他作物多时相遥感变化检测的前期实验基础。
A soft and hard Land Use / Cover Change Detection Method (SHLUCD) is proposed. In this method, the multi-temporal remote sensing image can effectively express the phenological characteristics of crops so as to achieve the characteristics of crop phenology in discrete change regions (ie pure pixel regions, including mutant and non-crop regions completely converted into crops) and continuous change regions , Mixed pixel area, is part of the area into the crop) accurate identification of crops. A study area was selected in Beijing. Taking winter wheat as the research object, we selected two satellite images of Environmental Disaster Reduction No. 1 on October 6, 2011 (sowing date) and April 16, 2012 (jointing period), and adopted hard change detection Methods HLUCD (Hard Land Use / Cover Change Detection Method), Soft Land Use / Cover Change Detection Method (SLUCD) and SHLUCD were used to identify winter wheat. The experimental results show that SHLUCD has a lower RMSE than traditional methods (SHLUCD [0.14,0.07], HLUCD [0.15,0.07], SLUCD [ 0.16,0.08]) and bias bias (SHLUCD -0.0008, HLUCD -0.007 and SLUCD 0.014) and higher decision coefficients R2 (SHLUCD [0.68,0.86], HLUCD [0.62,0.86] and SLUCD [ 0.60, 0.86]). The mutagenesis of winter wheat, the gradual change of winter wheat and the area of non-winter wheat were evaluated respectively, which showed that the recognition accuracy of SHLUCD was close to the best identification method in each area, further verifying the flexibility and applicability of SHLUCD. The SHLUCD method can effectively identify winter wheat in the discrete variation area through the change of land cover type state, and the abundance of winter wheat can be quantitatively expressed by recognizing the degree of state change of land cover in the continuous variation area, which is the pre-detection of multi-temporal remote sensing change of other crops Experimental basis.