Computer Vision Toolbox provides algorithms and tools for the design and simulation of computer vision and video processing systems. You can detect and track objects in video frames, recognize objects, calibrate cameras, perform stereo vision, and process 3D point clouds. You can detect objects such as faces, facial features, and pedestrians, and also create your own detectors.

The toolbox provides algorithms and functions to create image recognition and image retrieval systems. One available approach to do this is using the bag-of-words method. You can create your own object detection or recognition system by selecting and assigning objects of interest and training a classifier. Detected objects can be tracked over time using KLT and Kalman filter algorithms. In this example, a face is tracked with feature point tracking using the KLT algorithm.

For 3D computer vision, you can calibrate single and stereo cameras using the Camera Calibrator and Stereo Camera Calibrator apps. With stereo vision, you can calculate the depth of points in a scene and perform 3D reconstruction. 3D point cloud processing techniques are used to process data from 3D sensors such as LiDARs, stereo, and RGB-D cameras.

You can register and stitch together 3D point clouds and fit geometric shapes to 3D point clouds. Feature detection, extraction, and matching can be used to solve many computer vision problems, including image registration and object detection. The toolbox also includes over 50 Simulink^® blocks. As shown in this example, the lane markings on the road are detected to determine when a vehicle departs from its lane.

The toolbox also supports C-code generation using MATLAB Coder™. For more information on Computer Vision Toolbox, return to the product page or choose a link below.