The correct way to define a ROI - python

I'm currently working on a image processing project opencv python based, and essentially it's quite simple, it uses cv2.matchTemplate and a database to make all the matches, but I think something is done wrong.
The images are get from a few cameras, then the processing starts, based on predefined masks, the ROI of each object is returned the ROI of to then be matched with a database of images.
The problem is, the database has crops are the same size of the ROI predefined, which means the cv2.matchTemplate is always trying to match an image of the same size as the one in the database.
Here's an example with a printed circuit board (which is what my project does, it looks for bad positioned components/missing components).
Simple circuit board
This would be the full raw image, taken from one of the cameras.
Next a predefined filter would be applied to the raw image, I usually convert it to grayscale.
Now, my predefined ROI would crop the image.
Image ROI
Image crop
Notice though it crops way more than the component itself, and this crop will be saved in the database for eventual matches.
But shouldn't I be manually feeding my database with properly images? I mean, is it ideal? wouldn't it take way more matches (images) to get my desired threshold?
I need arguments to convince someone of it, I want to feed my database with well cropped images, but this person just wants to save it automatically.
Thanks!

Related

Find Coordinates of cropped image (JPG) from it's original

I have a database of original images and for each original images there are various cropped versions.
This is an example of how the image look like:
Original
Horizontal Crop
Square Crop
This is a very simple example, but most images are like this, some might taken a smaller section of the original image than others.
I was looking at OpenCV in python but I'm very new to this kind of image processing.
The idea is to be able to save the cropping information separate from the image to save space and then generate all the cropping and different aspect ratio on the fly with a cache system instead.
The method you are looking for is called "template matching". You find examples here
https://docs.opencv.org/trunk/d4/dc6/tutorial_py_template_matching.html
For your problem, given the large images, it might be a good idea to constrain the search space by resizing both images by the same factor. So that searching a position that isn't as precise, but allows then to constrain the actual full pixel sized search to a smaller region around that point.

"Normalizing" (de-skewing, re-scaling) images as preprocessing for OCR in Python

I have a bunch of scanned images of documents of the same layout (strict forms filled out with variable data) that I need to process with OCR. I can more or less cope with the OCR process itself (convert text images to text) but still have to cope with the annoying fact that the scanned images are distorted either by different degree of rotation, different scaling or both.
Because my method focuses on reading pieces of information from respective cells that are defined as bounding boxes by pixels, I must convert all pictures to a "standard" version where every corresponding cells are in the same pixel position, otherwise my reader "misreads". My question is, how could I "normalize" the distorted images?
I use Python.
Today in high-volume form-scanning jobs we use commercial software with adaptive template matching, which does deskew and selective binarization to prepare the images, but then it adapts field boxes per image, not placing boxes on XY-location.
Deskeing process overall increases the image size. It is visible in this random image from online search:
https://github.com/tesseract-ocr/tesseract/wiki/skew-linedetection.png
Notice how the title of the document was near the top border, and in the deskewed image it is shifted down. In this oversimplified example an XY-based box would not catch it.
I use commercial software for deskewing and image pre-processing. It is quite inexpensive but good. Unfortunately, I believe it will take you only part-way if the data capture method relies on xy-coordinate field matching. I sense your frustration with dealing with it, thus appropriate tools were already created for handling that.
I run a service bureau for such form processing. If you are interested I can further share privately methods how we process.then

Homography of soccer field

Okay so i am trying to find homography of a soccer match. What i have till now is
Read images from a folder which is basically many cropped images of a template soccer field. Basically this has images for center circle and penalty lines etc.
Read video stream from a file and crop it into many smaller segments.
Loop inside the images in video stream and inside that another loop for images that i read from folder.
Now in the two images that i get through iteration , i applied a green filter because of my assumption that field is green
Use orb to find points and then find matches.
Now the Problem is that because of players and some noise from croud, i am unable to find proper matches for homography. Also removing them is a problem because that also tends to hide the soccer field lines that i need to calculate the homography on.
Any suggestions on this is greatly appreciated. Also below are some sample code and images that i am using.
"Code being used"
Sample images
Output that i am getting
The image on right of output is a frame from video and that on left is the same sample image that i uploaded after filterGreen function as can be seen from the code.
Finally what i want is for the image to properly map to center circle so i can draw a cube in center, Somewhat similar to "This example" . Thanks in advance for helping me out.
An interesting technique to throw at this problem is RASL. It computes homographies that align stacks of related images. It does not require that you specify corresponding points on the images, but operates directly on the image pixels. It is robust against image occlusions (eg, players moving in the foreground).
I've just released a Python implementation here: https://github.com/welch/rasl
(there are also links there to the original RASL paper, MATLAB implementation, and data).
I am unsure if you'd want to crop the input images to that center circle, or if the entire frames can be aligned. Try both and see.

Automatic extract subimage with a frame

I am trying to extract a subimage from a scanned paper like this:
https://cloud.kopa.ch/index.php/s/gGZm5xeMYlPfU81
The extracted images should be georeferenced and added to a webmap service, but thats not the question here.
How can I get the frame / its pixel coordinates to crop the image?
I am also free in creating the "layout" (similar to the example), which means I could add markers to get the frame better after scanning it again.
The workflow is:
generate layout - print map - draw on the map - scan it - crop "map-frame" - georeferencing this frame - show it on a webmap
The "map-frames" are preprocessed and I know their location/extent
Has anybody an idea how to crop the (scanned) images automatically to this "map-frame"?
I have to work with python and have the packages PIL, pillow and imagemagick for the image processing
Thanks for you help!
If you need more information, don't hesitate to ask
Here's an example I adapted form the Pillow docs, check them out for any further processing that you might need to perform:
from Pillow import Image
Image.open("/path/to/image.jpg")
box = (100, 100, 400, 400)
region = im.crop(box)
Also, it might prove valuable to search Stack Overflow for this kind of operation, I'm sure it has been discussed earlier.
As for finding the actual rectangle to crop you'll have to do some form of image analysis. In it's simplest form, conceptually that could be something along these lines:
Applying an S-curve filter to a black-and-white representation of your image
Iterate over all of the pixels in the image
Keep track of horizontal and vertical lines that has sufficiently black pixel values.
Use this data to determine the bounding box of the portion of the image your interested in.
Depending on your needs you might want to look into some computer vision library instead, which are well optimized for this and similar tasks. The one that springs to mind is OpenCV which is I would guess is well optimized and documented, and there's a python module available as well.

how to test for a specific watermark in an image with python2.7

there is this project am currently working on, which requires me to watermark every uploaded image. i have tried series of examples online, but they are not giving me what i really want as result
for example
i have an image A with image B watermarked on it, the two images are of the same dimensions. i applied opacity of 0.5 on image B before placing it on image A
now, i would really appreciate if anyone could help with a boolean function to check if image A has already been watermarked with image B before watermarking it.
thanks.
This depends on several factors that you'll need to provide more information for.
For instance, how complex are these images? Is there a lot of noise? Are the images that are uploaded similar in any way, or are they heterogenous? Are the watermarks always the same, or are they different?
As a general principle for extracting objects from images, you should look into processes such as color deconvolution, thresholding, and blob extraction.
In short--some sample images would go a long way...
yeah, finally found a dubious way to solve the problem, by hiding a specific text on the alpha layer of the image after watermarking it using steganography.
so on every upload, i get the image, iterate through the lowest pixels of the image's alpha layer, then compare the result to the text. if the result matches the text, definitely, the image has been watermarked.

Categories

Resources