I am not sure how to describe my questions headline to point out my question in one sentence (english is not my native language so please do not flame me for it)
I have the following szenario.
An image is displayed on a computer display. a camera (webcam or similar) is pointed towards the disolay so the image observed will cover nearly the whole video frame, but might be a bit disorted.
How can i calibrate/detect the bounds of the image?
For calibration purposes i could display a test picture. The camera and displayed image will NOT change its position after calibration!
My only idea would be to click some reference points with my mouse. Automatic detection like in QR code would be preferred/perfect.
You can use a calibration points similarly to QR code. Place three black squires in the calibration image corners. As the result you'll obtain an area (inside a rectangle, set by calibration points) that can be used as a boundary for all further images.
Related
I want to detect the corners of a document in real time meaning that I want to draw a bounding quad around the document similar to how most doc-scanner app do. I also want to do this by scratch. I have already read a few corner detection algorithms such as the Harris corner detector, but I have a few concerns:
How do I implement this in the image analysis part of the camerax library and draw a bounding quad (not box since it can be slanted) around it in real-time meaning it is updated each frame of the camera preview?
Is there a more optimized approach since for the Harris corner detection, i will have to calculate the Ix, Iy, etc. each frame, and what if there isn't any significant movement in the image and the corners are relatively where they were before?
For laying detection result over preview, the question has been asked before. See: How to draw on PreviewView?
I don't know if there is a best practice for documents scanning. Dropbox has a series of well written blog posts on the topic: https://dropbox.tech/tag-results.doc-scanner
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.
I have a photo of a paper that I hold up to my webcam, and want to minimize the area of the photo to just the paper. This way, my OCR program will potentially be more accurate, as well as conceivably faster.
I have taken a couple steps thus far to isolate the paper from the background.
First, I use Canny Edge detection, set with high thresholds. This provides a two-color representation of the edges of my image. On it, I can see a rounded rectangle among some other artifacts that happen to have sharp edges in the background.
Next, I use a Hough transformation, to draw vectors with over 100 point hits in polar coordinates on a black background. The resulting image is as shown:
See that large (the largest), almost-rectangular figure right in the middle? That's the paper I'm holding. I need to isolate that trapezoid as a polygon, or somehow otherwise get the coordinates of its vertices.
I can use these coordinates on the original image to isolate a PNG of the paper and nothing else.
I would also highly appreciate if you could provide answers to any of these three sub-questions.
-How do you find the locations of the intersections of these lines on the image?
-How would I get rid of any lines that don't form the center trapezoidal polygon?
-With these points, is there anything better than convex hull that would allow me only to get the trapezoidal/rectangular shaped region of the image?
Here is another example, in which my program produced a better image:
I have a colored textual Image to be in black and white.
Specifically the text to be in black and background to be in white for whatever colored it might be
I am using JAVA to perform this operation.
Can anybody help me out to with a code snippet or point me to a discussion which is useful.
if you want to develop your own converter RGB to Black/White, please check Java - get pixel array from image . Here you import a image, access the pixel and with a specific threshold of the color you can decide to convert to black or white.
Or you use http://www.imagemagick.org/ as commandline tool (with api http://www.imagemagick.org/script/api.php).
Regards
Currently I am developing an application for decoding barcodes using mobile phones.
I have a problem with how to draw a line or a square on the camera screen to easily capture the barcode.
What is the easiest way of doing it?
Unfortunately this isn't as easy as it sounds. If you have a preview image from a phone's camera then it's often rendered as an overlay. This means that the camera preview image doesn't actually form any part of your application's canvas and you can't interact directly with the pixels. The phone simply draws the preview on top of your appliction, completely out of your control.
If you draw a line on your screen, then it will be drawn underneath the preview image.
The way around this isn't too pretty. You need to actually capture an image from the camera. Unfortunately this means capturing a JPEG or a PNG file into a byte buffer. You then load this image using Image.createImage and render that to the screen. You can then safely draw on top of that.
This also has the undesirable downside of giving you an appalling frame-rate. You might want to enumerate all the possible file formats you can capture in and try them all to see which one is quickest.
You can do this by using OverlayControl, assumming that your target devices support it.
I think i remember seeing a good example # Sony Ericsson developer forums.
Edit: found this (does not involve use of OverlayControl)