FooBar's profile - activity

"P != eye(3)" The projection matrix is 3x4, so why do you compare it to a 3x3 matrix? "and the number is larger than 4."

Just on small comment on patterns: Please do not print out paper and somehow fix it to a cardboard. Invest 20$ and let i

"But how can I know that camera calibtation was optimal? " That's a tricky question. A good rmse (well below a pixel fo

(8,8,0): I've never seen an 8m calibration target :)

"But how can I know that camera calibtation was optimal? " That's a tricky question. A good rmse (well below a pixel fo

TriangulatePoints with known object? Hey! I have a calibrated stereo setup (using the IR and RGB Camera from a older Pr

This representation is numerically unstable and not able to handle vertical lines.

I am trying to detect a textured artwork from their photocopied or printed versions using image processing. The texture as such is very thin (nearly 100 microns) so shape and shadow technique did not give result that well. Any other method through which I can achieve the same?

I have a mako camera and want to capture real time video from it. I do not know if I capture directly from vimba or I use opencv? If I use opencv so how I connect vimba with opencv? Thanks

This is your free choice. "projectPoints" IS a tool to simulate the scene.

You can generate a set of 3d points and a known transformation. Then use projectPoints with these values and some intrinsic camera matrix. This will create a set of 3d points, a transformation and 2d points that you can use as a test case.

so roughly

projectPoints(3d_points, R, t, camMatrix) -> 2d_points

and

solvePnP(2d_points, 3d_points, camMatrix) -> (R,t)

Traceback (most recent call last): File "C:\dfsdf.py", line 13, in <module> cv2.imshow("input", img) error: ......\opencv-2.4.13.2\modules\highgui\src\window.cpp:269: error: (-215) size.width>0 && size.height>0 in function cv::imshow

This is not possible. Visual odometry is only accurate up to scale so you will need additional information like an object with known size. This fact gets more intuitive if you think about adding a camera to a toy car that is driving through a toy city. The images will look the same as in the real world.

This is not a "do my homework"-forum. What did you try before asking here?

i have install open cv version 2.4.9 and visual studio 2010 X86,i have windows 8 64 bit system , now i want to configure open cv with python i have tried this lib and software

matplotlib-1.3.0.win32-py2.7

numpy-1.7.1-win32-superpack-python2.7

python-2.7.12 shel (for python x64) results are : Python 2.7.12 (v2.7.12:d33e0cf91556, Jun 27 2016, 15:19:22) [MSC v.1500 32 bit (Intel)] on win32 Type "copyright", "credits" or "license()" for more information.

import numpy import matplotlib import cv2

Traceback (most recent call last): File "<pyshell#2>", line 1, in <module> import cv2 ImportError: DLL load failed: %1 is not a valid Win32 application.

python-2.7.12 shel (for python x86) results are : Python 2.7.12 (v2.7.12:d33e0cf91556, Jun 27 2016, 15:19:22) [MSC v.1500 32 bit (Intel)] on win32 Type "copyright", "credits" or "license()" for more information.

import numpy import matplotlib import cv2

Traceback (most recent call last): File "<pyshell#2>", line 1, in <module> import cv2 ImportError: DLL load failed: %1 is not a valid Win32 application.

I am using following code to differentiate between a sky blue region (required region) in my image

and the background. I have applied PyrMeanShiftFiltering function to do segmentation of the image. After applying code the result

image have successfully segmented sky blue section but now i want to know the location of this sky blue region (mid point). Can any one help me how i can know the coordinates of this blue region in my image after doing this image segmentation.

#include "cv.h"
#include "highgui.h"
#include "math.h"
#include <iostream.h>


int main(int argc, char** argv)
{
IplImage* output_image;

    IplImage* image = cvLoadImage("31.jpg",CV_LOAD_IMAGE_COLOR);
  CvMemStorage* storage = cvCreateMemStorage(0);
        cvNamedWindow( "origional image", 1 );
        cvShowImage( "origional image", image);
        cvWaitKey(0);

    IplImage *filtered = cvCreateImage(cvGetSize(image),image->depth,image->nChannels);

        cvCopy(image,filtered,NULL);

  int level = 3;
  int spatial_radius = 40;
  int color_radius  = 40;

  filtered->width &= -(1<<level);
  filtered->height &= -(1<<level);

  cvPyrMeanShiftFiltering(filtered, filtered,spatial_radius,color_radius,level);
  cvNamedWindow( "fourth", 1 );
        cvShowImage( "fourth", filtered);
        cvWaitKey(0);

   return 0;
}

If you have additional requirements (e.g. cards without barcode) it would be helpful to integrate this information into the question. How do you expect someone to guess this?

Hey! These are possible approaches, have you also thought about template matching without features: cv::matchTemplate I also thought about detecting (or reading the barcode. e.g. using zbar). OpenCV has also some OCR capabilities: OCRTesseract

It took me 5 seconds to come up with around four ideas and you really have none? Where is your own work?

Hi forum,

How can I detect an object and define its position in an image using opencv c++?

Do you think using Cascade of Haar-Like Classifier for image are correct way? If not, what can i do?

Thanks for your help!! Minh.

You most probably will use another feature like orb or brisk which have similar performance.

The patent is 50 years old... Only the algorithms in the nonfree-modules are patented, the rest in OpenCV can be used commercially.

Have a look at cv::remap

import numpy as np
import cv2
from matplotlib import pyplot as plt
while(1):
    img=cv2.imread('D:\IMG_0590_1.jpg')    
    ball = img[278:104, 330:158]
    img[181:355, 100:274] = ball

    cv2.imshow('img',img)

    if cv2.waitKey(10) & 0xFF == ord('q'):
            break
cv2.destroyAllWindows()

This is my code but it is giving the error mentioned in the question. What is the problem?

Could you post example images of what you want to achieve?

"whereas saving the image to disk involves multiple RAM and disk accesses." You could write the file to a RamDisk which would already speed up the current implementation by just changing the directory.

I prefer to create my calibration targets from aluminum composites: https://us.whitewall.com/photo-lab/al... They are planar enough, can be rather large and are extremely cheap.

Please add the code here so that the question can still be useful after the link expires. "falling on image "13380-sony.jpg"." How do you expect someone to know what is happing?

Please don't post text as an image. (Why not print it, scan it and then post it?) And what is the meaning of the matrices?

Do you first get the intrinsic calibration for the two cameras?

'cv::initUndistortRectifyMap' That should be the inverse function. It creates map so that it applies the distortion model to a pixel. (The pixel at (x,y) in my dst image is seen at map(x, y)=distort(x, y) in the src image). Thanks for also running an evaluation. Have you also measured (or estimated) the variance in the runs? I had some iterations where the remap was around 2 or 3 times slower than in the mean case.

Hello!

I'm currently working on distorting an image with remap to simulate a camera. This is how I create the maps:

Mat float_distortion_map_x(height_, width_, CV_32FC1);
Mat float_distortion_map_y(height_, width_, CV_32FC1);

Point2f undistorted;
for (int x=0; x<width_; ++x)
    for (int y=0; y < height_; ++y)
    {
        {
            undistort_point(Point2f(x, y), undistorted); // wrapper for cv::undistortPoints
            float_distortion_map_x.at<float>(y, x) = undistorted.x;
            float_distortion_map_y.at<float>(y, x) = undistorted.y;
        }
    }

#if 1
    ROS_INFO("NO Conversion");
    float_distortion_map_x.copyTo(distortion_map_1_);
    float_distortion_map_y.copyTo(distortion_map_2_);
#else
    ROS_INFO("CONVERTING");
    distortion_map_1_ = Mat(height_, width_, CV_16SC2);
    distortion_map_1_ = Mat(height_, width_, CV_16UC1);
    convertMaps(float_distortion_map_x, float_distortion_map_y,  distortion_map_1_, distortion_map_2_, CV_16SC2, false);
#endif

The results look great, If I run a cv::undistort on the result, I again get my original image, so that this part should be ok. However, the remap documentation claims a speed up of around 2 for the converted version which I cannot reproduce. I called remap with different interpolation types and the run time for a VGA image have been (mean for 100 conversions)

> LANCZOS4: 33ms,  
> Linear: 0.5ms, 
> Cubic: 1.3ms

But the same for the original maps and the converted ones. I could not see any difference. Has anyone also tested this claim and found something similar?

imshow is in the highgui-module: highgui so you have to include that.

# USAGE
# python motion_detector.py
# python motion_detector.py --video videos/example_01.mp4

# import the necessary packages
import argparse
import datetime
import imutils
import time
import cv2

# construct the argument parser and parse the arguments
ap = argparse.ArgumentParser()
ap.add_argument("-v", "--video", help="path to the video file")
ap.add_argument("-a", "--min-area", type=int, default=500, help="minimum area size")
args = vars(ap.parse_args())

# if the video argument is None, then we are reading from webcam
if args.get("video", None) is None:
    camera = cv2.VideoCapture(0)
    time.sleep(0.25)

# otherwise, we are reading from a video file
else:
    camera = cv2.VideoCapture(args["video"])

# initialize the first frame in the video stream
firstFrame = None

# loop over the frames of the video
while True:
    # grab the current frame and initialize the occupied/unoccupied
    # text
    (grabbed, frame) = camera.read()
    text = "Unoccupied"

    # if the frame could not be grabbed, then we have reached the end
    # of the video
    if not grabbed:
        break

    # resize the frame, convert it to grayscale, and blur it
    frame = imutils.resize(frame, width=500)
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    gray = cv2.GaussianBlur(gray, (21, 21), 0)

    # if the first frame is None, initialize it
    if firstFrame is None:
        firstFrame = gray
        continue

    # compute the absolute difference between the current frame and
    # first frame
    frameDelta = cv2.absdiff(firstFrame, gray)
    thresh = cv2.threshold(frameDelta, 25, 255, cv2.THRESH_BINARY)[1]

    # dilate the thresholded image to fill in holes, then find contours
    # on thresholded image
    thresh = cv2.dilate(thresh, None, iterations=2)
    _, contours, _ = cv2.findContours(thresh.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

    # loop over the contours
    for c in cnts:
        # if the contour is too small, ignore it
        if cv2.contourArea(c) < args["min_area"]:
            continue

        # compute the bounding box for the contour, draw it on the frame,
        # and update the text
        (x, y, w, h) = cv2.boundingRect(c)
        cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 0), 2)
        text = "Occupied"

    # draw the text and timestamp on the frame
    cv2.putText(frame, "Room Status: {}".format(text), (10, 20),
                cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 2)
    cv2.putText(frame, datetime.datetime.now().strftime("%A %d %B %Y %I:%M:%S%p"),
                (10, frame.shape[0] - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.35, (0, 0, 255), 1)

    # show the frame and record if the user presses a key
    cv2.imshow("Security Feed", frame)
    cv2.imshow("Thresh", thresh)
    cv2.imshow("Frame Delta", frameDelta)
    key = cv2.waitKey(1) & 0xFF

    # if the `q` key is pressed, break from the lop
    if key == ord("q"):
        break

# cleanup the camera and close any open windows
camera.release()
cv2.destroyAllWindows()

You could try to use split and merge: Split/Merge

split your image, call setTo(0) for your green channel and merge them again.

Here is a good tutorial