adamsss's profile - activity

I'm working with large tiff images (~~16k x 16k). I used slicing image for 9 tiles in my algorithm (library: image_slicer) before image processing.

I'm looking for some faster alternative method of slicing image for tiles. This one is so slow - when i open 4-5 images in once time, algorith slicing images for few minutes.

I tried to use img=img[x:y , x:y], but this method crashed my python application.

I'm working in my programme with "huge" tiffs (8411x8361,60Mb). When i tried use template matching function on the unprocessed image, console showed me this kind of error:

OpenCV Error: Insufficient memory (Failed to allocate 140682288 bytes) in cv::OutOfMemoryError, file ......\sources\modules\core\src\alloc.cpp, line 52 OpenCV Error: Assertion failed (u != 0) in cv::Mat::create, file ......\sources\modules\core\src\matrix.cpp, line 424

I understand meaning of this error (I hope so), so i sliced my image for 4 tiles and use cutted part of my TIFF in template matching algorithm. Console showed me the same error message - of course with lower number of "failed to allocate X bytes".

Moreover, template matching is only one part of my programme. In previous steps it wasn't any problems with open,use and edit unprocessed huge TIFFs.

Any ideas?

It was loaded correctly. I can show it in grayscale and colour.

Additionally, this code works fine with "test.jpg", but doesn't work with TIF file

My code, now:

image_dir='7_7026_polowa.tif'   
image = cv2.imread(image_dir)
..........
img_gray = cv2.imread(image_dir, 0)
template = cv2.imread('template.tif', 0)
w, h = template.shape[::-1]

res = cv2.matchTemplate(img_gray,template,cv2.TM_CCOEFF_NORMED)
threshold = 0.8
loc = np.where( res >= threshold)
for pt in zip(*loc[::-1]):
      cv2.rectangle(image, pt, (pt[0] + w, pt[1] + h), (0,0,255), 2)

 cv2.imwrite('res.png',image)

Ok, i corrected it, but now another error:

res = cv2.matchTemplate(img_gray,template,cv2.TM_CCOEFF_NORMED)
cv2.error: ..\..\..\sources\modules\core\src\matrix.cpp:424: error: (-215) u != 0 in function     cv::Mat::create

Hello, I would like to use this code in my programme:

img_rgb = image
img_gray = cv2.cvtColor(img_rgb, cv2.COLOR_BGR2GRAY)
template = roi
w, h = template.shape[::-1]

res = cv2.matchTemplate(img_gray,template,cv2.TM_CCOEFF_NORMED)
threshold = 0.8
loc = np.where( res >= threshold)
for pt in zip(*loc[::-1]):
     cv2.rectangle(img_rgb, pt, (pt[0] + w, pt[1] + h), (0,0,255), 2)

cv2.imwrite('res.png',img_rgb)

image and roi are the images (tiff format) readed in earlier part of my code

Now, i got error:

w, h = template.shape[::-1] ValueError: too many values to unpack (expected 2)

Where is the problem?!

Hello,

I found angle and scale invariant template matching code, but not for python language... Have anybody got it?

I will be very grateful !

Ok, i try again and it works now correctly. Thanks for all advices!

I see, that cv2.pyd file is not created in both places: in opencv folder and in python's libraries folder ...

I tried all process again and now it was built successfull ... but when i try import cv2 in IDLE still error msg:

 import cv2

Traceback (most recent call last): File "<pyshell#2>", line 1, in <module> import cv2 ImportError: No module named 'cv2'

BTW... results from RELEASE mode: ========== Build: 2 succeeded, 1 failed, 53 up-to-date, 0 skipped ==========

and still doesn't work

both in RELEASE mode?

Now, i'm trying build in DEBUG mode (BUILD_SHARED_LIBS set off) ...

Result: ========== Build: 55 succeeded, 1 failed, 0 up-to-date, 0 skipped ==========

and result of build the INSTALL:

========== Build: 55 succeeded, 2 failed, 0 up-to-date, 0 skipped ==========

but still doesn't work...

Link to raport from cmake: link text

========== Build: 3 succeeded, 1 failed, 53 up-to-date, 0 skipped ========== and python 3.5 folder is still empty :(

I built it from src... like this: link text

but it still doens't import cv2 (the same error) !

I have the same problem with "ImportError: DLL load failed: The specified module could not be found" with OpenCV 3.1 and Python 3.5 ... but installed Microsoft Visual C++ 2015 Redistributable still didn't solve my problem. Why? Any suggestions?