使用nnet包进行光栅图像分类

Question

我在“nnet＆＃39;”中遇到了栅格分类问题。 R中的包。我无法创建具有预测的单元格类别的栅格。有我的代码：

library(sp)
library(nnet)
library(caret)
library(rgdal)
library(raster)

images_list=list.files(path="maps", full.names=TRUE, ignore.case = TRUE) 

rasters_stack <- stack(images_list)   

table = read.csv("points.csv", sep = ",")

indeks <- createDataPartition(table$class, p=2/3, list=FALSE)
training <- table[indeks, ] 
test    <- table[-indeks, ]

model = nnet(factor(class) ~ blue + green + red, data = training, size=2)

predict(rasters_stack, model, type="class", filename="output.img",
         progress="text", overwrite=TRUE)

我收到了这些错误：

Error in v * x@data@gain : non-numeric argument to binary operator
In addition: Warning messages:
1: In rgdal::putRasterData(x@file@transient, v, band = 1, offset = off) :
  NAs introduced by coercion
2: In rgdal::putRasterData(x@file@transient, v, band = 1, offset = off) :
  NAs introduced by coercion
3: In rgdal::putRasterData(x@file@transient, v, band = 1, offset = off) :
  NAs introduced by coercion
4: In rgdal::putRasterData(x@file@transient, v, band = 1, offset = off) :
  NAs introduced by coercion
5: In .gd_SetStatistics(object, ...) : NAs introduced by coercion