Shulou(Shulou.com)05/31 Report--

This article mainly introduces the relevant knowledge of "how python uses OpenCV to achieve multi-target tracking". Xiaobian shows you the operation process through practical cases. The operation method is simple, fast and practical. I hope this article "how python uses OpenCV to achieve multi-target tracking" can help you solve the problem.

1 Background

Most beginners in computer vision and machine learning learn object detection. If you're a beginner, you might wonder why we need object tracking. Can't we just detect objects in each frame?

Let's explore a few reasons tracking is useful:

First, when multiple objects, such as people, are detected in a video frame, tracking helps determine the identity of the object across frames.

Second, in some cases, object detection may fail, but it is still possible to track an object because tracking takes into account the position and appearance of the object in the previous frame.

Third, some tracking algorithms are very fast because they search locally rather than globally. Therefore, we can obtain very high performance for our system by performing target detection every nth frame and tracking objects in the intermediate frames.

So why not track objects indefinitely after the first detection? Tracking algorithms can sometimes lose objects they are tracking. For example, when the motion of an object is too large, the tracking algorithm may not be able to keep up. Target detection is usually repeated after a period of target tracking.

In this tutorial, we will focus only on the tracing section. The objects we want to track will be captured by specifying bounding boxes around them.

2 Multi-target tracking based on MultiTracker

The MultiTracker class in OpenCV provides an implementation of multi-target tracking. But this is only a preliminary implementation, as it only deals with tracked objects and does not perform any optimization on tracked objects.

2.1 Create a single object tracker

A multi-object tracker is just a collection of individual object trackers. We first define a function that takes a tracker type as input and creates a tracker object.

OpenCV has 8 different tracker types: BOOSTING, MIL, KCF, TLD, MEDIANFLOW, GOTURN, MOSSE, CSRT. This article does not use the GON TURN tracker. Typically we give the tracker class name, return a single tracker object, and build multiple tracker classes.

C++ Code:

vector trackerTypes = {"BOOSTING", "MIL", "KCF", "TLD", "MEDIANFLOW", "GOTURN", "MOSSE", "CSRT"};/** * @brief Create a Tracker By Name object Initialize tracker according to set type * * @param trackerType * @return Ptr */Ptr createTrackerByName(string trackerType){ Ptr tracker; if (trackerType == trackerTypes[0]) tracker = TrackerBoosting::create(); else if (trackerType == trackerTypes[1]) tracker = TrackerMIL::create(); else if (trackerType == trackerTypes[2]) tracker = TrackerKCF::create(); else if (trackerType == trackerTypes[3]) tracker = TrackerTLD::create(); else if (trackerType == trackerTypes[4]) tracker = TrackerMedianFlow::create(); else if (trackerType == trackerTypes[5]) tracker = TrackerGOTURN::create(); else if (trackerType == trackerTypes[6]) tracker = TrackerMOSSE::create(); else if (trackerType == trackerTypes[7]) tracker = TrackerCSRT::create(); else { cout

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.