Skip to main content
. Author manuscript; available in PMC: 2019 Dec 4.
Published in final edited form as: Irrig Sci. 2018 Dec 3;1:1–23. doi: 10.1007/s00271-018-0613-9

Table 3.

Assessment accuracy between different methods and manually extracted method for a small part of the study of area

Date of flight Item Method Unsupervised
Supervised
Index based
Physical based
Method
Method
Method
Method
Classes Shadow Non-shadow Shadow Non-shadow Shadow Non-shadow Shadow Non-shadow

2014, August Manually extracted Shadow 27,039 6742 31,292 2489 32,683 1098 29,455 4326
Non-shadow 20,485 170,695 8433 182,747 5608 185,572 10,598 180,582
Total 47,524 177,437 39,725 185,236 38,291 186,670 40,053 184,908
Assessment accuracy (correctness metric) 80.6% 93.4% 96.7% 87.2%
2015, June Manually extracted Shadow 19,038 3917 21,038 1917 21,393 1562 20,084 2871
Non-shadow 2566 199,440 2109 199,897 1192 200,814 2456 199,550
Total 21,604 203,357 23,147 201,814 22,585 202,376 22,540 202,421
Assessment accuracy (correctness metric) 82.9% 91.6% 93.2% 87.5%
2015, July Manually extracted Shadow 11,845 2416 13,030 1231 13,320 941 12,497 1764
Non-shadow 3454 207,246 2561 208,139 1459 209,241 2964 207,736
Total 15,299 209,662 15,591 209,370 14,779 210,182 15,461 209,500
Assessment accuracy (correctness metric) 83.1% 91.3% 93.4% 87.6%
2016, May Manually extracted Shadow 18,301 3459 19,668 2092 20,268 1492 18,796 2964
Non-shadow 5697 197,504 3294 199,907 2314 200,887 4198 199,003
Total 23,998 200,963 22,962 201,999 22,582 202,379 22,994 201,967
Assessment accuracy (correctness metric) 84.1% 90.4% 93.1% 86.4%