# Tag Info

5

There are a lot of good books. Multiple View Geometry in Computer Vision, "Computer Vision: Algorithms and Applications", but anyway, replication of the theory is the necessary stage to me. You can start from the finishing course "3D Computer Vision" from the CTU. They published lectures and labs - in MatLAB. P.S. Here is a ready implementation http://...

4

[EDITED] Here's how it's done. Steps: 1. Isolate the Road Divider Part. Then, using Houghlines, find out the longest lines in Image. Find out the extrema points that cross image boundary. You got the Quadilateral points. I skipped this part by Manually Choosing them. In my case, the width of road at top of image is 10, and at bottom is 60. Now, for the ...

4

The computer vision system toolbox in Matlab includes Feature Detection, Extraction, and Matching. The line, edges, histogram of oriented gradients are all included other than the corner features. And there is also an example on point feature matching in the toolbox (the names are called corners but they are actually not).

3

This looks like a sort of ray tracing problem. If you know the position and orientation of your camera you should be able to calculate the 3x4 projection matrix and its inverse. This should allow you to convert from image points to 3d position (on the road). This discussion http://opencv-users.1802565.n2.nabble.com/2D-to-3D-projection-with-given-plane-...

2

As far as I can tell, Matlab's Computer Vision toolbox provides vision.CascadeObjectDetector for object detection, with support only for Frontal Face (CART),Frontal Face (LBP),Upper Body,Eye Pair,Single Eye,Single Eye (CART),Profile Face,Mouth,Nose. When you say: "the rectangle is drawn around the face instead.", that is happening because you are calling ...

2

Looking at the release notes for the computer vision toolbox, it looks like that functionality was not added until the R2013a release.

2

The seminal paper on background modeling is Adaptive background mixture models for real-time tracking by Stauffer and Grimson. If you have Matlab and the Computer Vision System Toolbox, then you can try using vision.ForegroundDetector, which implements a version of their algorithm.

2

You are not doing anything wrong. This particular example is there for a quick illustration, and it does not produce a very good detector. It trains a 5-stage detector from a very small training set. A decent detector may have 20 stages or more, and you would need thousands of positive samples and negative images to train it. Take a look at this tutorial ...

2

What I would try is to : Get the HSL (Hue Saturation Lightness) decomposition of the original picture Only keep the L component Multiply the image you have from the last steps with the image you get after the morphological operation (the black and white one, after deleting the noise) Do some thresholding on that Use what you obtained from that instead of ...

2

Thanks to helpful comments from @MBaz, I managed to come up with a solution: we can associate multiple audio samples with a single frame using the VideoFileWriter object. This fact and use-case is missing in the documentation. First, some stats about the audio and video files. The stereo audio samples are in a 2xN array signal. The video frames are in a ...

2

Here are a couple of examples from the computer vision system toolbox: Using kalman filter for object tracking For a more in-depth explanation the best book is Multiple Target Tracking with Radar Applications by Samuel Blackman.

1

It depends which types of the features are useful for you. If the corners are not useful/caracteristic, you should look elsewhere. For example, may be blob-like features will suit you need. Take a look at wikipedia article on feature detectors http://en.wikipedia.org/wiki/Feature_detection_(computer_vision). It is a good starting point and will give you an ...

1

You would have to connect both cameras to your computer, and either create two webcam objects or two videoinput objects to read images from them.

1

pcshow() was implemented in 2015b and in R2015a the equivalent function is called showPointCloud.

1

Since you have two webcams you can reconstruct the video scene using the frame sequence. If your webcams happends to be stereo pair (with known extrinsic parameters aka rotation and translation) your task becomes pretty easy. On the other hand if webcam locations are unknown the problem gets a little trickier, but it can still be solved using matlab's ...

1

You are correct, Haar features, as well as LBP and HOG are not rotation invariant. I have experimented with vision.CascadeObjectDetector in MATLAB's Computer Vision System Toolbox, and found that the face detector model it comes with can tolerate about 15 degrees of in-plane rotation. So if you want to handle in-plane rotation, the easiest thing is to ...

1

I know it is not matlab, but this post about training a 'banana' detector is really great and can give you some tips.

1

If you want to train the classifiers with your own database, you will only need 'trainCascadeObjectDetector' function and feed your images into the proper arguments (Positive, negative images). The output classifier will be in your 'outputXMLFilename' as in traincascadeobjectdetector trainingImageLabeler is helpful function for classifying positive/...

1

For pairwise stereo calibration try using the Stereo Camera Calibrator app in the Computer Vision System Toolbox. It is much easier to use than Caltech Camera Calibration toolbox. For starters it detects the checkerboard automatically.

1

Use an asymmetric calibration plate. Such as the new circle grid in OpenCV. Check here. You do not need to find the correspondences in the gui, but rather run the calibration directly. Also, do not forget to initialize the optimization problem using the median of the obtained poses. To calibrate multiple view setups, I would also recommend Multi Camera ...

1

In the case of video, the best approach to use is the ShiTomasi Corner Detector. (detectMinEigenFeatures). Now, if you want to detect corners that contains some invariant featrues, you should use SURF, BRISK, ORG, AKAZE, FREAK, etc... Example: im = rgb2gray(imread('image.jpg')); [r c] = corner_ST(im,20);

1

If you can calibrate your camera, and if you can detect some reference points on the road surface, then you can get 3D coordinates of the image points that are on the road. This assumes that the road is on a plane. In other words, if you can detect a car, then you can calculate the 3D coordinates of the bottom of the car. See this example in MATLAB using ...

1

There are a few of things you can try: Definitely move FaceDetect = vision.CascadeObjectDetector; outside of the loop. You only need to create the face detector object once. Re-creating it for every frame is definitely your performance bottleneck. vision.VideoFileReader returns a frame of class 'single' by default. If you change the output data type to '...

1

If you are after the simplest algorithm, you can basically try Genaralized Hough Transform using the image edges only (not corners). For really textureless objects, the following might guide you: 1) If you are willing to use an RGBD camera such as kinect, Hinterstoisser has come up with a quite good algorithm to do this: http://cvlabwww.epfl.ch/~lepetit/...

1

You should position the board at roughly the same distance from the camera as the distance between the camera and your objects of interest. The board should be big enough to cover a good portion of the field of view at that distance. Another point: ideally, your board should have an even number of squares along one side, and an odd number of squares along ...

1

Check out this tutorial on how to use trainCascadeObjectDetector. It describes common pitfalls, and includes examples. There is also a graphical tool for labeling training images on MATLAB Central, which makes preparing your training data much less painful. By the way, the link you posted is actually the documentation for vision.CascadeObjectDetector, ...

Only top voted, non community-wiki answers of a minimum length are eligible