简介

提供了混淆矩阵类，可以实现对影像分类结果进行精度评价。

confusionMatrix

构造函数，返回一个由给定参数构建的混合矩阵对象。

函数	返回值
confusionMatrix(type=null, array=null)	ConfusionMatrix

参数	类型	说明
type	Array	混合矩阵中类型的数组对象，默认为null
array	Array	混合矩阵的数组对象，默认为null

示例

image = pie.Image('user/17090142114/PGDB001/World84').select(["B1","B2","B3"])
featureCollection = pie.FeatureCollection('user/17090142114/PGDB001/WorldROI')
training = image.sampleRegions(featureCollection,["type"],50000).getInfo()
classifer = pie.RT().train(training,"type",["B1","B2","B3"])
resultImage = image.classify(classifer,"classifyA")
result = classifer.confusionMatrix().getInfo()

matrix = pie.ConfusionMatrix(result.get('type'),result.get('array'))
kappa = matrix.kappa()
print(kappa.getInfo())

acc

返回混合矩阵的精确度。

函数	返回值
acc()	混合矩阵的精确度计算结果

示例

image = pie.Image('user/17090142114/PGDB001/World84').select(["B1","B2","B3"])
featureCollection = pie.FeatureCollection('user/17090142114/PGDB001/WorldROI')
training = image.sampleRegions(featureCollection,["type"],50000).getInfo()
classifer = pie.RT().train(training,"type",["B1","B2","B3"])
resultImage = image.classify(classifer,"classifyA")
result = classifer.confusionMatrix().getInfo()

matrix = pie.ConfusionMatrix(result.get('type'),result.get('array'))
acc = matrix.acc()
print('acc=',acc.getInfo())

kappa

返回混合矩阵的精确度。

函数	返回值
kappa()	混合矩阵的kappa系数计算结果

示例

image = pie.Image('user/17090142114/PGDB001/World84').select(["B1","B2","B3"])
featureCollection = pie.FeatureCollection('user/17090142114/PGDB001/WorldROI')
training = image.sampleRegions(featureCollection,["type"],50000).getInfo()
classifer = pie.RT().train(training,"type",["B1","B2","B3"])
resultImage = image.classify(classifer,"classifyA")
result = classifer.confusionMatrix().getInfo()

matrix = pie.ConfusionMatrix(result.get('type'),result.get('array'))
kappa = matrix.kappa()
print('kappa=', kappa.getInfo())