Prepare a  MATLAB code for Multi Object Tracking in aerial thermal video.

The requirements are as follows:

1. Reading video from the computer in RGB format.

2. Converting RGB video into Grayscale.

3. Setting of thresholds (adaptive thresholding) to transform Grayscale into Binary format.

4. Object detection using Connected Components Labelling in binarized video (frames).

5. Drawing bounding Box and assigning label (unique ID) to each bounding box (detected object).

6. Calculating confidence score for each detected object using assignment problem and auction algorithm (compare each object with other objects and ensure its not the other object).

7. Track all the detected objects using Kalman Filter.

8. The tracker shall be robust to:

a. Handle occlusion and does not switch ID assigned to different detected objects while tracking.

b. Various threshold levels of the thermal video.

9. Calculating accuracy, precision and IoU of the bounding box with ground-truth (annotated box) on a dataset.