I am new to computer vision but I am trying to code an android app which does the following:
Get the live camera preview and try to detect one logo in that (i have the logo in my resources). In real-time. Draw a rect around the logo if found. If there is no match, dont draw the rectangle.
I already tried a couple of things including template-matching and feature detection using ORB.
Why that didnt work:
Template-matching:
Issues with scaling and rotation. I tried a multi scale variant of it but a) the performance was really bad and b) the rectangle was of course always shown trying to search for the image. There was no way to actually confirm in the code if the logo was found or not.
ORB feature detection:
Also pretty slow (5-6 fps) but it worked ok-ish. The other problem was that also i never could be sure if the logo was in the picture or not. ORB found random matches even if the logo was not in the picture.
Like I said, I am very new to this. I would appreciate the help on what would be the best way to achieve:
Confirm if a picture A (around 200x200 pixels) is in ROI of camera picture (around 600x600 pixels).
This shouldnt take longer than 50ms per frame. I dont know if thats even possible though. So if a correct way to do this would take a bit longer than that, I would just do the work in a seperate thread and only analyze like every fifth camera frame or so.
Would appreciate any hints or code examples on how to achieve that. Thank you!
With logo detection, I would highly recommend using OpenCV HaarClassifier. It is easy to generate training samples from a collection of images of the logo, or one logo image with many distortions.
If you can use a few rules like the minimum and maximum size of the logo to be detected, and possible regions on the image where it can appear, you can run the detector at a speed better than you mention with ORB.
Related
i am actually writing a "simple" version of the game "Diablo" with JavaFX.
At the moment the screen size is absolute 800x600, everything is working fine. Now our team wants to use fullscreen on every desktop.
My Problem is: We are using images for walls, floor etc with an absolute size of 25x25. If we want to work in fullscreen mode, we would neet to scale them.
But as far as i know, you only can scale ImageView or BufferedImage. Of course we are reusing the images. e.g. a wall texture appears more than 20 times on the Pane. That's easy with an image, since you can use it again and again, but you would need 20 ImageViews for the same thing.
The screen is update about every 0.2 seconds, so we can't always scale all the ImageViews.
My ideas were:
Very many scaled ImageViews in a HashMap (bad performance)
Resizing all original images into a new resource directory at the start of the game (bad storage solution - duplicated images)
I am sure, there is a better solution and a intelligent standard workaround, but i didn't find it yet.
Thank you for every help,
TailorDurden
Okay, i found the answer myself. Simple as usual:
Image myImage = new Image("img.png", 15, 20, false false);
scales the image in the loading process to 15x20..
I have been trying to "cut" an image for some time now, I ll explain why and what I tried.
So I wanted to create an hp "bar" except it's not a bar but a heart and so I though it would be easy all I had to do is have two pictures draw them on top of each other and then just cut one to make it appear as in hp was being lost, but I was not able to find a way to cut the image.
Setting the height just resizes the image as you might have guessed
I tried using textureRegion to kind of hack it but it didn't go so well
I found a method called clip begin which also uses scissors but for some reason that just doesn't seem to be working.
I might be using the clip begin wrong but I can't really find any real documentation on it, all I'm doing is:
image.clipBegin(x,y,height,weight);
image.clipEnd();
I almost forgot, I'm using a scene2d Image, might be a better way to go around it but not sure what that would be.
I would appreciate any ideas on how to do this, thank you.
You want to use the OpenGL Scissor support that Libgdx exposes. See the Libgdx Clipping wiki
and the Libgdx ScissorStack documentation.
The API isn't particularly friendly (its designed to support dynamically pushing multiple constraining rectangles, which as far as I've seen, isn't used very often).
The important point to remember with the scissor stack is that it only applies to actual draw commands that get issued. Since most APIs try to batch up draw commands, this means actual drawing might not happen when it looks like it should happen. To ensure clipping is happening you must flush any buffered draws before pushing the scissor (otherwise the wrong thing might get clipped) and you must flush any draw calls before popping the scissor (otherwise things you want clipped might avoid the scissors).
See libgdx ScissorStack not working as expected or libGDX - How to clip or How to draw on just a portion of the screen with SpriteBatch in libgdx? or Making a Group hide Actors outside of its bounds.
I'm making a game in Java and I have it so that when the player presses the left arrow button, three images are drawn one after the other. But when the third image is drawn, you can still see the image drawn before it (I rotated the images to make it look more realistic). Is there anyway to "undraw" the images? I've done some research and haven't found anything specific.
No, you can't undraw an image. It sounds like the images aren't fully opaque.
Not to be an over-complication junkie, but how about using OpenGL to draw them on quads. That way you can hide, show, rotate, animate, accelerate etc so much easier. And it will all be tons faster than with Swing/AWT. JOGL library + NeHe Tutorials will quickly show you how to achieve this (really, you only need this and this). Or use Unity if you wanna do it even simpler and more portable.
Best reader,
I'm stuck on one of my concepts.
I'm making a program which classroom children can measure themselves with.
This is what the program includes;
- 1 webcam (only used for a simple webcam view.)
- 2 phidgets (don't mind these.)
So, this was my plan. I'll draw a rectangle on the webcamview and make it repaint itself constantly.
When the repainting is stopped by one of the phidgets, the rectangle's value will be returned in centimeters or meters.
I've already written the code of the rectangle that's repainting itself and this was my result:
(It's a roundRectangle, the lines are kind of hard to see in this image, sorry about that.)
As you can see, the background is now simply black.
I want to set the background of this JFrame as a webcam view (if possible) and then draw the
rectangle over the webcam view instead of the black background.
I've already looked into jmf, fmj and such but am getting errors even after checking my webcam path and adding the needed jar libraries. So I want to try other options.
So;
- I simply want to open my webcam, use it as background (yes live stream, if possible in some way).
And then draw this rectangle over it.
I'm thus wondering if this is possible, or if there's other options for me to achieve this.
Hope you understand my situation, and please ask if anything's unclear.
EDIT:
I got my camera to open now trough java. The running camera is of type "Process".
This is where I got the code for my camera to open: http://www.linglom.com/2007/06/06/how-to-run-command-line-or-execute-external-application-from-java/
I adjusted mine a little so it'll open my camera instead.
But now I'm wondering; is it possible to set a process as background of a JFrame?
Or can I somehow add the process to a JPanel and then add it to a JFrame?
I've tried several things without any succes.
My program as it is now, when I run it, opens the measuring frame and the camera view seperatly.
But the goal is to fuse them and make the repainting-rectangle paint over the camera view.
Help much appreciated!
I don't think it's a matter of setting a webcam stream as the background for your interface. More likely, you need to create a media player component, add it to your GUI, and then overlay your rectangles on top of that component.
As you probably know from searching for Java webcam solutions in Stack Overflow already, it's not easy, but hopefully the JMF Specs and API Guide will help you through it. The API guide is a PDF and has sections on receiving media streams, as well as sample code.
I'm making an Android live wallpaper using OpenGL.
I would like to let the user pick an image to use for the background. Can someone give me any advice on an easy way to do this?
I know that I need to down sample bitmaps before loading them to avoid using huge amounts of memory for large bitmaps.
I also know OpenGL only supports texture sizes that are powers of two. It looks like messy work having to calculate the down scale factor then calculate the closest good texture size and then scale a quad with that texture appropriately to the screen.
It would also be nice to have the wallpaper panning feature when homescreens change.
Is there any easier way than doing all of the above myself? If not, is there sample code available for what appears to be a very common task?