I got a webcam and I'm running some algorithm on the received images to find movements in it.
But, the automatic auto white balance is changing the excepted result of the pixels color variance.
That's why I'm trying to disable it.
Some one knows a way to get the web can image with out this automatic color balance or disable it ?
Tks
If you are on Linux, you can test disabling automatic white balance using the uvcdynctrl command line utility. This is not a permanent solution, since these settings are reset every time the webcam is disconnected from the computer. As of OpenCV 2.1, configuring white balance is not supported using cv::VideoCapture::set(). Assuming this is still the case in OpenCV 2.2, you will need to use another library to configure your webcam and capture frames.
If you on Linux, you can see an example of using direct Video4Linux (V4L) syscalls to do this in one of my Github projects.
In my case I found on driver two options that must be disabled, auto white exposure and auto white balance, I disable both of them and the image got as I needed.
Related
I am writing a script on the raspberry pi to detect the majority colour featured in a frame of a webcam and I seem to be having an issue. The following image is me holding up my phone with a blank red image on it. I seem to be getting an orange colour instead.
Now when I angle the phone I do in fact produce the red colour expected.
I am not sure why this is the case.
I am using a logitech c920 webcam that emits a blue light when activated and also have the monitor going. I am wondering whether the light from these two are causing this issue and when I angle it, these lights are not hitting it front on and thus not distributing the image.
I am still not heavily experienced in this area so I would enjoy hearing explanations and possible work arounds for my problem.
Thanks
There are a few things that can mess this up:
As you already mention, the light from the monitor and the camera.
The iPhone screen is a display, so flicker and sync might also be coming to play.
Reflection from the iPhone screen.
If your camera has automatic control for exposure and color balance etc., the picture quality can change as you move around.
I suggest using a colored piece of non-glossy paper so that you can remove the iPhone display's effects.
I am researching into the best way to detect test in a photo using open source libraries.
I think the standard way is as follows (note: steps 1 - 4 all use OpenCV):
1) detect outline of document
2) transform document so it's flat and cropped, using said outline
3) Make the background of document white, using a filter
4) Feed resulting image to Tesseract
Is this the optimum process, or is there a better way, or better tools?
Also, what happens for case if the photo doesn't have a document outline (It's possible that step 1 & 2 are redundant)?
Is there anyway to automatically detect document orientation (i.e. portrait / landscape)?
I think your process is fine. I've used a similar process for an Android project.
I think that the only way you can discover if a document is portrait/landscape is to reason with the length of the sides of the bounding box of your outline.
I don't think there's an automatic way to do this, maybe you can find the most external contour approximable with a 4 segment polyline (all doable in opencv). In order to get this you'll have to work with contour hierarchy and contous approximation (see cv2.approxPolyDP).
This is how I would go for automatic outline detection. As I said, the rest of your algorithm seems just fine to me.
PS. I'll leave my Android project GitHub link. I don't know if it can be useful to you, but here I specify the outline by dragging some handles, then transform the image and feed it to Tesseract, using Java and OpenCV. Yeah It's a very bad idea to do that in the main thread of an Android app and yeah, the app is not finished. I just wanted to experiment with OCR, so I didn't care much of performance and usability, since this was not intended to use, but just for studying.
Look up the uniform width transform.
What this does is detect edges which have more or less the same width with respect to their opposite edge. So things like drainpipes (which can be eliminated at a later pass) but also the majority of text. Whilst conceptually it's similar to a distance transform, the published method uses rather ad hoc normal projection methods and Canny edge detection.
I've looked all over Stack and other sources, but I haven't seen any code that seems to successfully emulate what a polarization filter does, reducing glare. The application I want for this code won't allow for a physical filter, so I was wondering if anyone had tried this.
I'm using OpenCV image processing (mat) in C++ on an Android platform, and glare is interfering with the results I'm trying to get. Imagine a lost object you're trying to find based on a finite set of Red/Green/Blue values; if the object is smooth, glare would render bad results. And that's my current problem.
OK, no, there's no virtual polarization that can be accomplished just with code. It's possible to find (via image color saturation) glare spots on shiny objects, and those can be overwritten with nearby pixels without glare, but that's not the same thing as real polarization. That requires a physical, metal mesh in front of the lens, or sensor, to eliminate those stray light waves that create glare.
Tell you what. The person who invents the virtual polarization filter, using just code, will be an instant billionaire since every cell phone and digital camera company will want to license the patent.
I'm using the two OpenCV functions mentioned above to retrieve frames from my webcam. No additional properties are set, just running with default parameters.
While reading frames in a loop I can see that the image changes, brightness and contrast seem to be adjusted automatically. It definitely seems to be a operation of OpenCV because the image captured by the camera is not changed and lit constantly.
So how can I disable this automated correction? I could not find a property that seems to be able to do that job.
You should try to play around with these three parameters:
CV_CAP_PROP_BRIGHTNESS Brightness of the image (only for cameras)
CV_CAP_PROP_CONTRAST Contrast of the image (only for cameras)
CV_CAP_PROP_SATURATION Saturation of the image (only for cameras)
Try to set them all to 50. Also (if it won't help) try to change another camera capture parameters from documentation.
To answer that for my own: OpenCV is buggy or outdated here.
it seems to be impossible to get images in native resolution of the camera, they're always 640x480; also forcing it to an other value by setting width and height properties does not change anything
it seems to be impossible to disable the automatic image correction, the properties mentioned above seem not to work
the brightness/contrast properties doesn't seem to work as well - or at least I could not find any good values for it or the automatic image correction always overrides them
To sum it up: I'd not recommend to use OpenCV for some more enhanced image capturing.
I am about to build a basic online image editor for my web application using rails and rmagick. I did a bit of googling but couldn't find any existing solution, however I'd like to be sure before I spend a lot of time rolling my own.
Is anyone aware of a plugin/gem that provides a pre-rolled image editor with undo/redo and minimal image degradation on multiple edits?
Thanks!
I know this exists: Rails Image Editor
http://github.com/heurionconsulting/rails_image_editor/tree/master
It might be too late but I'm building one too. I'm using the Pixastic and Raphael libraries to generate live previews. So people can actually preview the resulting image, without having to wait for the server to process it.
Features are:
Quick fit (make the image instantly fitting a given width, height or width and height)
Crop
Resize
Rotate left and right 90º (animated)
Flip horizontal and/or vertical
Colorize (color picker)
Make warmer (slider)
Make colder (slider)
Make greyscale (one click)
Make sepia (one click)
Glow (slider)
Blur (slider)
Sharpen (slider)
Brightness (slider)
Contrast (slider)
Details:
It is provided as a Rails plugin.
Currently it only works on Paperclip attachments.
Integration is piece of cake (js include and a link in the view, that's it!)
It is displayed as a lightbox-like overlay
All requests are Ajax (so no page reloading)
I skipped RMagick and built a dedicated ImageMagick module for serverside image processing
The whole frontend is based on Q, a javascript toolset providing a slider, color picker, floating windows, growl-like feature, I18n for javascript, decent Cookie management and much more.
configurable with a single YAML file
integrated languages are US English and Dutch
Possible downsides:
Live preview on color correction is not available on IE due to lack of HTML5 canvas support
The Q library is not free for commercial domains (only €49 per domain or €249 for a wildcard version)
This plugin will be available for free next month in Alpha release.