Skip to main content
. 2015 Nov 4;15(11):27969–27989. doi: 10.3390/s151127969

Table 3.

Confusion matrix of visual classification (VC) versus Artificial Neural Network (ANN) classification. Feature codes have been abbreviated as follows: side bars (SB), erosion (ER), riffle (RI), deep water (DW), shallow water (SW), tree (TR), shadow (SH), vegetation (VG), vegetated bar (VB), vegetated bank (VK), submerged vegetation (SV), emergent vegetation (EV) and grass (GR). GE stands for georeferencing error.

ANN Classification
Feature VC SB ER RI DW SW TE SH VG GE Total
SB 1334 1097 - 8 - 2 - 10 214 3 1334
ER 287 - 22 13 1 3 - 10 238 - 287
RI 3339 - 1 2717 - 318 - 219 76 8 3339
DW 2082 - - 60 1927 54 - 8 29 4 2082
SW 2573 - - 262 80 1514 - 493 217 7 2573
TR 1755 - - 76 1 29 496 135 1013 5 1755
VB 299 - - - - - - - 299 - 299
VK 313 - 10 - 6 - - 15 281 1 313
SV 468 - - 160 - 125 - 46 135 2 468
EV 71 - 1 9 - 2 - 1 58 - 71
GR 344 - - - - - - - 343 1 344
SH 220 - 4 - - - - 180 31 - 220
Total 13,085 1097 38 3305 2015 2052 496 1117 2934 31 13,085