Algorithm 1: Aircraft-tracking algorithm

Input: Image sequence.

Output: Aircraft location with bounding box.

1: Generate region proposals by clustering.

2: for n =1 to m (m is the number of candidate regions) do

3: Calculate gray distribution field.

4: Compute structural distribution.

5: Calculate similarity between target’s model and candidate region.

6: end for

7: Detect occlusion via measuring the variation of the model distance.

8: Select a region with the minimal distance as the tracking region.

9: Update the target’s model.