Inundated_Area_Identification

Version 3.1

Algorithm:

• Step 1: Find minimum pixel value (Vmin) in image/matrix.
• Step 2: Set Global threshold (t) to SD of the whole image.
• Step 3: Find water pixel p and draw a window around pixel p.
• Step 4: Calculate mean (m1) of the pixels in the window in range [Vmin, Vmin + t] using global threshold t.
• Step 5: Increase the threshold, t by a small value Δt. Let, this be the local threshold, t1’.
• Step 6: Now calculate the mean (m2) of the window again.
• Step 7: If abs(m1-m2) > x then increase t1’ by Δt
• Step 8: Terminate when mean difference is constant or negligible (<= x ).
• Step 9: Else repeat from step 3 to step 7.
• Step 10: With the final value of Local Threshold ∂ grow the region.

P.s.: mean difference acceptable range x are choosen by trial and error

Limitation

• Cannot Identify small regions or cluster of small regions

Conclusion

• Final Solution of the Project.