Back | 5. Accuracy Assessment |
5.1 Accuracy Assessment of LISS-3 classified map: The accuracy assessment was done with the collection of training sites data for the entire Chikballapur taluk The producer's, user's accuracy and overall accuracy corresponding to the various categories were computed, along with the error matrices for supervised and unsupervised classified MSS data of LISS-3, which is summarised in Table 2. The LISS-3 supervised classification accuracy assessment gave a kappa (k) value of 0.95 indicating that an observed classification is in agreement to the order of 95 percent.
Table 2: Producer's accuracy, user's accuracy and overall accuracy of land cover classification using LISS-3 MSS data for Chikballapur Taluk.
|
||||||||
Category |
Producer's Accuracy (%) |
User's Accuracy (%) |
Overall accuracy (%) |
Producer's Accuracy (%) |
User's Accuracy (%) |
Overall Accuracy (%) |
||
Agriculture |
94.21 |
84.54 |
95.63 |
94.47 |
83.39 |
90.22 |
||
Built up |
96.47 |
83.11 |
89.68 |
80.30 |
||||
Forest |
94.73 |
96.20 |
86.77 |
89.71 |
||||
Plantation |
92.27 |
91.73 |
84.44 |
90.10 |
||||
Waste land |
97.49 |
89.88 |
93.03 |
93.37 |
||||
Water bi\odies |
96.13 |
98.33 |
92.91 |
94.89 |
Back |
5.2 Accuracy Assessment of MODIS classified Maps
5.2.1 Accuracy Assessment using Error matrix - User's, Producer's and Overall accuracy assessment of the MODIS classified maps (using hard classifier) was done for Chikballapur taluk with the ground truth data and the results are listed in Table 3, 4 and 5.
Table 3: User's Accuracy of classified MODIS Data of Chikballapur taluk.
Algorithms |
Agriculture |
Built up |
Forest |
Plantation |
Waste land |
Water bodies |
NN (B1 to B7) |
94.00 |
80.80 |
94.65 |
59.40 |
93.87 |
45.55 |
NN (PCA) |
97.33 |
95.18 |
67.67 |
95.38 |
74.07 |
48.00 |
NN (MNF) |
93.89 |
94.46 |
89.13 |
85.60 |
74.22 |
59.40 |
Table 4: Producer's Accuracy of classified MODIS Data of Chikballapur taluk.
Algorithms |
Agriculture |
Built up |
Forest |
Plantation |
Waste land |
Water bodies |
NN (B1 to B7) |
56.73 |
99.00 |
73.07 |
96.60 |
89.53 |
68.56 |
NN (PCA) |
57.55 |
93.00 |
94.00 |
93.00 |
93.00 |
73.51 |
NN (MNF) |
69.99 |
91.89 |
87.24 |
99.00 |
95.00 |
56.93 |
Table 5: Overall Accuracy of classified MODIS Data of Chikballapur taluk.
Techniques |
Overall Accuracy |
NN on MODIS Surface reflectance bands (B1 to B7) |
68.88 |
NN on MODIS derived PCs (36 bands) |
71.02 |
NN on MODIS derived MNF Components (36 bands) |
86.11 |
Back |
5.2.2 Comparison based on land cover class percentage area - Land cover statistics were computed for all taluks pertaining to each classification algorithm at the taluk level.
5.2.3 Pixel to pixel analysis with LISS-3 MSS classified image - MODIS classified data were also compared with LISS-3 MSS classified data on a pixel by pixel basis for accuracy assessment of pure (homogenous) pixels. One pixel of MODIS spatially corresponds to 121 pixels (that is approximately equal to 258.5 m) of LISS-3. The error matrix was generated with user's accuracy, producer's accuracy and overall accuracy for the taluk and is listed in Table 6, 7 and 8.
Table 6: User's Accuracy obtained from pixel to pixel analysis with LISS-3 image comparison for Chikballapur taluk .
Algorithms |
Agriculture |
Built up |
Forest |
Plantation |
Wasteland |
Water bodies |
NN (B1 to B7) |
37 |
17 |
59 |
44 |
87 |
29 |
NN (PCA) |
20 |
45 |
61 |
69 |
81 |
56 |
NN (MNF) |
41 |
55 |
61 |
75 |
81 |
65 |
Table 7 : Producer's Accuracy obtained from pixel to pixel analysis with LISS-3 image comparison for Chikballapur taluk.
Algorithms |
Agriculture |
Built up |
Forest |
Plantation |
Wasteland |
Water bodies |
NN (B1 to B7) |
46 |
65 |
19 |
41 |
60 |
45 |
NN (PCA) |
61 |
29 |
38 |
69 |
78 |
45 |
NN (MNF) |
59 |
41 |
55 |
76 |
83 |
65 |
Table 8: Overall Accuracy of classified MODIS Data of Chikballapur taluk from pixel to pixel analysis with LISS-3 image comparison.
Technique |
Overall Accuracy |
NN on MODIS Surface reflectance bands (B1 to B7) |
51.34 |
NN on MODIS derived PCs (36 bands) |
63.69 |
NN on MODIS derived MNF Components (36 bands) |
69.87 |
The accuracy assessment showed that Neural Network classification on MNF components of MOFDIS bands 1 to 36 had highest overall accuracy followed by NN on PC's and NN on MODIS bands 1 to 7.