Video and Image Processing Blockset™ 2.5

Product Description

• Introduction and Key Features
• Modeling and Simulating Video and Imaging Systems
• Generating and Optimizing C Code
• Multimedia I/O, Video Viewer, and Display Blocks
• Video and Image Processing Algorithms
• Video and Image Analysis

Video and Image Analysis

Image analysis techniques enable you to extract information from video streams. For example, you can remove noise, correct for nonuniform illumination, or extract image features.

The Video and Image Processing Blockset lets you:

• Detect motion in a video sequence using techniques such as optical flow (Horn & Schunck and Lucas & Kanade), block matching for motion estimation, or 2-D sum of absolute differences (SAD)
• Match patterns to an existing template using techniques such as cross-correlation
• Judge the relative focus of a video scene
• Track and classify moving objects
• Separate foreground from background using segmentation techniques
• Identify object boundaries in an image frame using Canny, Sobel, Prewitt, or Roberts edge detection methods
• Use Blob Analysis to calculate statistics and return values that represent spatial coordinate locations, such as centroid or bounding boxes, for labeled regions in a binary image

The Statistics library enables you to compute 2-D statistical analyses, such as variance, correlation, max, min, mean, standard deviation, and histograms. Mean variance and standard deviation can be computed on an arbitrary Region of Interest (ROI).

Model illustrating a motion stabilization technique based on the SAD method. The stabilization subsystem (above) applies the SAD technique to remove unwanted translational camera motions and generate a stabilized video. Click on image to see enlarged view.