基于GF遥感数据纹理分析识别制种玉米

仅利用多时相遥感数据识别作物,其精度难以满足制种玉米识别的实际需求。该文针对制种玉米种植特点,利用国产GF遥感数据,构建了制种玉米遥感识别方法。首先利用多时相国产卫星GF-1 WFV数据,依据研究区作物的物候历,构建各地类EVI时序曲线,提取玉米种植区域;进一步利用抽雄期的GF-2 PAN数据,以田块为对象,通过Sobel边缘检测算子,提取作物纹理信息,并利用Hough变换检测制种玉米田块中的条带状纹理信息,最终提取出制种玉米。该文以新疆维吾尔自治区奇台县坎尔孜乡为研究区,对该文构建的方法进行试验验证,试验结果显示,制种玉米识别精度为90.0%,Kappa系数为0.80。该文不但拓宽了中国国产...

Full description

Saved in:
Bibliographic Details
Published in农业工程学报 Vol. 32; no. 21; pp. 183 - 188
Main Author 张超 金虹杉 刘哲 李智晓 宁明宇 孙海艳
Format Journal Article
LanguageChinese
Published 国土资源部农用地质量与监控重点实验室，北京 100035%中国农业大学信息与电气工程学院,北京,100083%全国农业技术推广服务中心,北京,100125 2016
中国农业大学信息与电气工程学院，北京100083
Subjects
Hough变换
作物
制种玉米
纹理
遥感
高分一号WFV
高分二号PAN
纹理
GF-2 Pan
texture
作物
remote sensing
遥感
crops
Hough transform
Hough变换
高分二号PAN
制种玉米
seed maize
高分一号WFV
GF-1 WFV
Online AccessGet full text
ISSN1002-6819
DOI10.11975/j.issn.1002-6819.2016.21.024

Cover

More Information
Summary:仅利用多时相遥感数据识别作物,其精度难以满足制种玉米识别的实际需求。该文针对制种玉米种植特点,利用国产GF遥感数据,构建了制种玉米遥感识别方法。首先利用多时相国产卫星GF-1 WFV数据,依据研究区作物的物候历,构建各地类EVI时序曲线,提取玉米种植区域;进一步利用抽雄期的GF-2 PAN数据,以田块为对象,通过Sobel边缘检测算子,提取作物纹理信息,并利用Hough变换检测制种玉米田块中的条带状纹理信息,最终提取出制种玉米。该文以新疆维吾尔自治区奇台县坎尔孜乡为研究区,对该文构建的方法进行试验验证,试验结果显示,制种玉米识别精度为90.0%,Kappa系数为0.80。该文不但拓宽了中国国产遥感数据的应用领域,同时也为中国玉米制种监管提供了新的技术支撑。
Bibliography:11-2047/S
remote sensing; crops; texture; seed maize; GF-1 WFV; GF-2 Pan; Hough transform
In order to accurately gather the area and yield information of seed crops and guarantee the seed supplying safety, it＇s necessary to use the remote sensing technology to improve the efficiency and accuracy of the traditional manual statistic survey means. However, the crop identification based on multi-temporal remote sensing data cannot be used in actual seed maize identification because of the low accuracy. Now the domestic remote sensing satellite GF-1 and GF-2 have been launched, and the data characteristic of high spatial and high temporal resolution plays an important role in the field of remote sensing for agricultural condition. With the feature of high spatial and high temporal resolution of the data, this paper builds a method to identify the seed maize focusing on the planting feature of seed maize. At first, according to regional crop calendar, use the domestic GF-1＇s WFV sensor multi-temporal data ranging fro
ISSN:1002-6819
DOI:10.11975/j.issn.1002-6819.2016.21.024