Tracking Objects Based on Multiple Particle Filters for Multipart Combined Moving Directions Information

Algorithm 2

Particle filter for random processes .

	Input: pf particles sample set (based on Algorithm 4), k th image (where k starts from the second image)
	Output: The new set of particles represents , estimate the state of the object in the k th image.
	Step 1:
	for i = 1 to do
	Beginfor
	/∗(with is the i th particle)∗/
	Get
	Get
	Get
	Calculate
	Calculate by Algorithm 8
	Update weight for i th particle:
	Endfor
	Step 2:
	Calculate sum
	for i = 1 to do
	Beginfor
	Standardize weight:
	Endfor
	Step 3: Calculate
	Step 4:
	if then
	use boostrap
	Endif
	Step 5: Estimate the state of the object in the k th image by calculating the average of the new set of particles