techfort / pycv Goto Github PK

in book (chapter 8, surveillance demo) says
"In this screenshot, the blue rectangle is the CAMShift detection and the green rectangle is the Kalman filter prediction with its center at the blue circle."

but when look the the code, green rectangles are contours, not tracking using camshift/kalman

this is some wrong with me,i run you code ,but this is not right for you example,have no line int the picture,
i do not know what is wrong
please tell me ,thank you very much!

what should I do to install opencv python on linux, I already compiled the opencv from source.

Hi,

I've tried to run the car_sliding_windows directly from the last repo here.
The issue I currently have is : after checking the path and the existance of the images files, I've anyway the following error:

396
397
398
399
OpenCV Error: Assertion failed (The data should normally be NULL!) in allocate, file /home/david/OpenCV/opencv-3.1.0/modules/python/src2/cv2.cpp, line 163
Traceback (most recent call last):
File "car_sliding_windows.py", line 17, in
svm, extractor = car_detector()
File "/home/david/OpenCV/pycv/chapter7/car_detector/detector.py", line 43, in car_detector
extract_bow.setVocabulary(vocabulary)
cv2.error: /home/david/OpenCV/opencv-3.1.0/modules/python/src2/cv2.cpp:163: error: (-215) The data should normally be NULL! in function allocate

This have been done under these conditions :

• changed all path strings: "/home/d3athmast3r/(....)" to reflect mine. This affects : detector.py, car_sliding_windows.py, bow.py, detect_hog_svm.py
• I'm running python 2.7.12 under virtualenwrapper
• under the same Virtualenv I've been compiled OpenCV 3.1.0 with the contribs of the same version

    k3 = ndimage.convolve(img, kernel_3x3)
  File "C:\Anaconda2\lib\site-packages\scipy\ndimage\filters.py", line 697, in convolve
    origin, True)
  File "C:\Anaconda2\lib\site-packages\scipy\ndimage\filters.py", line 530, in _correlate_or_convolve
    raise RuntimeError('filter weights array has incorrect shape.')
RuntimeError: filter weights array has incorrect shape.

Looks like scipy does not like the array, maybe I'm a newer version. Alternatively I wanted to use the VM to ensure I don't have any newer version however I could not find a the VM to download per the book info.

The code is reading twice the frames before applying it to background subtractor

Lines 122 and 127.
in line 122 frames are stored in variable called frane not frame