东北林业大学学报
東北林業大學學報
동북임업대학학보
JOURNAL OF NORTHEAST FORESTRY UNIVERSITY
2015年
7期
110-115
,共6页
猫群优化算法%高光谱图像%森林类型
貓群優化算法%高光譜圖像%森林類型
묘군우화산법%고광보도상%삼림류형
Cat swarm algorithm%Hyperspectral image%Forest type
以吉林省汪清林业局为研究区,通过猫群位置寻优的过程对阔叶林、针叶林和混交林进行聚类分析。结果表明:森林类型区分精度达到83.5%,Kappa系数0.793,与传统高光谱聚类方法相比,能较好的识别森林类型。
以吉林省汪清林業跼為研究區,通過貓群位置尋優的過程對闊葉林、針葉林和混交林進行聚類分析。結果錶明:森林類型區分精度達到83.5%,Kappa繫數0.793,與傳統高光譜聚類方法相比,能較好的識彆森林類型。
이길림성왕청임업국위연구구,통과묘군위치심우적과정대활협림、침협림화혼교림진행취류분석。결과표명:삼림류형구분정도체도83.5%,Kappa계수0.793,여전통고광보취류방법상비,능교호적식별삼림류형。
In Wangqing Forestry Bureau of Jilin Province, by location optimization process of cat swarm algorithm, we clustered the forest types between broad-leaved forest, coniferous forest and mixed forest with better effect compared with traditional hyperspectral clustering method, and the classification accuracy of forest type clustering was 83.5%, and Kappa coefficient was 0.793.