The present repository containts a simple code to detect and track an object. Image processing is based on thresholding after deleting background and particle tracking on Euclidean distance calculation.
We want to detect and track objects in a image, few examples marked with blue circles in figure 1.
The code runs using the main_particle_tracking.m script. Set the parameters in Define Parameters section:
- file name and location
- First, Last and Frame rate
- time convertion, from frame to time [s, min,..]
The code starts, one frame of the video pops up and you must select the area of interest, blue rectangle in next figure. Once you select the place, right click, and select Crop Image.
Next, on the Command Window, you receive the following requests:
The number is needed to convert pixels to space units, then insert a known dimension of the blue box you selected before (figure 2).
The results are listed as a set of cells in the variable Particles, where each one of them represents one particle with the following information
| time | x position | y position |
|---|---|---|
| t1 | x1 | y1 |
| . | . | . |
| . | . | . |
| . | . | . |
| . | . | . |
| tn | xn | yn |
A second variable is listed as Particle_Velocity with the following information
| time | x velocity | y velocity | net velocity |
|---|---|---|---|
| t1 | vx1 | vy1 | v1 |
| . | . | . | . |
| . | . | . | . |
| . | . | . | . |
| . | . | . | . |
| tn-1 | vxn-1 | vyn-1 | vn-1 |
It is calculated as first-orden difference.
Mac Intyre J., Simple Object Tracking (2023), GitHub repository, https://github.com/macinj1/Object_Tracking