Video movement data extraction and analysis

I want to track an object in a recorded video and extract the movement data e.g the direction it turns and the angle at which it turns at. Can anyone give me some guidance?

[Walter Roberson]

https://www.mathworks.com/help/vision/tracking-and-motion-estimation.html

Answer 1

[Simar]

Hi Abigail,

I understand that you are looking for guidance related to tracking an object in a recorded video and extract its movement data. This involves several steps that can be accomplished using MATLAB and Computer Vision Toolbox. Here is a general guide to get started:

Step 1: Read the Video

Firstly, import video into MATLAB. Use the “VideoReader” function to create a video reader object for reading video files.

v = VideoReader('yourVideoFile.mp4'); 

Step 2: Object Detection

Before tracking, detect the object in the first frame. There are several approaches to object detection, including:

• Thresholding and Morphological Operations: For objects with distinct colour.
• Feature Detection and Matching: Useful for objects with distinct features.
• Machine Learning or Deep Learning Models: Pre-trained models like YOLO, SSD, or custom-trained models for complex object detection.

Choose the method based on object characteristics.

Step 3: Object Tracking

Once object is detected, track it across frames. MATLAB offers several tracking algorithms:

• vision.PointTracker: Tracks a set of points.
• vision.KalmanFilter: Predicts the location of a moving object based on its previous state.
• vision.HistogramBasedTracker: Tracks objects based on colour histograms.

For example, to use a simple point tracker:

tracker = vision.PointTracker('MaxBidirectionalError', 2); 
initialize(tracker, initialLocation, initialFrame); 

Step 4: Analyze Movement

While tracking the object frame by frame, analyze its movement by comparing its positions between frames. To determine the direction and angle of turn:

• Direction: Calculate the vector between positions in consecutive frames.
• Angle of Turn: Use the dot product between consecutive direction vectors to find the angle.

angleInRadians = acos(dot(vector1, vector2) / (norm(vector1) * norm(vector2))); 
angleInDegrees = rad2deg(angleInRadians); 

Step 5: Visualize and Extract Data

Visualize the tracked path directly on the video frames using “insertShape” or “insertMarker” functions. To extract and analyse the movement data, store the positions, directions, and angles in arrays for further processing or visualization.

Utilize MATLAB's Image Processing Toolbox and Computer Vision Toolbox for advanced functions and algorithms that simplify these tasks. Experiment with different methods and parameters to find the best fit for specific video and object of interest. As already shared by Walter this is a great resource to start with - https://www.mathworks.com/help/vision/tracking-and-motion-estimation.html

Please refer to the following documentation links for more information-

• Image Processing Toolbox- https://www.mathworks.com/help/images/getting-started-with-image-processing-toolbox.html
• Computer Vision Toolbox- https://www.mathworks.com/help/vision/?s_tid=srchbrcm
• Video Reader- https://www.mathworks.com/help/matlab/ref/videoreader.html?searchHighlight=videoreader&s_tid=srchtitle_support_results_1_videoreader
• Insert Shape- https://www.mathworks.com/help/vision/ref/insertshape.html?searchHighlight=insertshape&s_tid=srchtitle_support_results_1_insertshape
• Insert Marker- https://www.mathworks.com/help/vision/ref/insertmarker.html?searchHighlight=insert%20marker&s_tid=srchtitle_support_results_1_insert%20marker

Hope it helps!

Best Regards,

Simar

[Abigail]

Thank you that is really helpful!