The Impact of Multiple Pond Conditions on the Performance of Dike-Pond Extraction

Dike-ponds in fisheries often present multiple pond conditions such as pure, suspended sediment, water bloom, semidry conditions, etc. However, the impact of these on performance extracting dike-pond from remote sensing images has not been studied. To solve this problem, we explore existence impacts by comparing four rule-based methods two groups test regions. The first group few while second more. results show that various measure values deteriorate proportion regions increases. All performed worse than first, where overall accuracy decreased 8.80%, misclassification error increased 3.69%, omission raised 10.53%, and correct quantity rate dropped 8.23%, respectively. extraction method ingested indistinguishably other group. it outperformed group, with a 4.22% improvement accuracy, 10.25% decrease error, 19.03% increase rate. These findings suggest can negatively should be considered applications require precise size, number, shape.