landzs / tracking_multiple_objects_in_surveillance_cameras

Automatic detection and tracking of moving vehicles in a video from a surveillance camera

Tracking_Multiple_Objects_In_Surveillance_Cameras

Description

This is project for course EL2320 Applied Estimation. It is base on Motion-Based Multiple Object Tracking. This project performs automatic detection and tracking of moving vehicles in a video from a surveillance camera. Generally, the program contains 2 parts: 1)Detecting moving objects in each frame , 2) Associating the detections corresponding to the same object over time. The motion of each track is estimated by a particle filter.
You can find more details in Report

How to execute

Tracking_Cars('Data/SampleX.mp4') (X is number form 1 to 9)

Experimental results

Tracking Car By Particle Filter 1
Tracking Car By Particle Filter 2

tracking_multiple_objects_in_surveillance_cameras's Issues

Description of Weights

I have studied your paper and your code, but I could not understand why have you described about the dimension of particles as 100x2. I could understand this that 100 represents the number of particle, but what 2 represent in it?
Kindly elaborate this concept for me.
Thanks in advance

there is a problem when running

Improper assignment with rectangular empty matrix.

Error in detectionToTrackAssignment (line 19)
cost(i, :) = distance(mean(tracks(i).particles), centroids);

Error in Tracking_Cars (line 22)
[assignments, unassignedTracks, unassignedDetections] = ...

Error in Untitled (line 1)
Tracking_Cars('Data/Sample9.mp4') %(X is number form 1 to 9)