YaddyVirus's profile - activity

Call each value by a separate name and hence call the respective trackbar.. sweet! But how do I see the script's output on my screen?

I have made a traffic light detection algorithm with OpenCV and Python. My script has some trackbars which I need to adjust and see the output to set a desired stage.

My script runs on my Raspberry Pi, is there any way I can send the trackbar values from another device on the same network on the RPi? For example, maybe make a phone app (I use Windows Phone BTW) which sends the trackbar data to the script and shows the output on my phone screen.

@jmbapps Didn't get a word of that... can you please elaborate? I think that C

Hi there!

I want to extract centre coordinates of detected circles using cv2. HoughCircles, here's the code-

import cv2
import numpy as np
def f(x): return

img = cv2.imread('C:/Users/Yardy/img1.png')
cimg = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
edges = cv2.Canny(cimg,50,150,apertureSize=3)
circles = cv2.HoughCircles(edges, cv2.HOUGH_GRADIENT, 1.2, 200,param2=30)

if circles is not None:
    circles = np.round(circles[0, :]).astype("int")
    X = circles

print X

When I run this code, in X I get the centre coordinates as well as the radius of the circle. How do I store the coordinates and radius in separate variables?

Also if my code detects more than one circle, I get an error -

Need more than one value to repack.

Can anybody help me?

Actually Its quite simple, just threshold the image, and use cv2.HoughCircles to find a circle, now put a condtional statement that as soon as the pixel intensity at the detected circle goes below a minimun threshold, activate the robot.

@berak ohh the minimum and max ranges of the y and x axis respectively right?

@Tetragramm I actually came up with another method to solve the problem. I used np.count_nonzero, and restricted the ROI to one half of the image, and thanks to @berak , it worked!
Now I have another question at hand, I have to track two lines, one white and one yellow, is it possible to extract both lines in one threshold?

@berak it works! Thanks a ton! But I don't understand in what order you've written the coordinates

@Tetragramm, I've been through the documentary for whole nights, and it helped me a lot, but it fails to answer several specific questions...

Hi there!

I am working on a script that reads the non-zero pixels in a binary image. I am using np.count_nonzero() to count non zero pixels, which works fine until, I specify a range of coordinates.

ele = np.count_nonzero(img)

Works fine
But, when I specify a range of pixels, for example I am working with a 640*480 res image, Now I want that only the pixels falling in the range of (320,0) and (640,480), that is, the right half, be checked for non zero values, so I modify my code to -

ele = np.count_nonzero(((320,0)<gray) & (gray <(640,480)))

But this line gives an error -

ValueError: operands could not be broadcast together with shapes (480,640) (2,)

To me it sounds like the image layers (RGB or similar) might be causing this problem. Can anyone tell me how to fix it?

@Tetragramm Actually, I don't know how do I split my binary image into 2 parts so that I can use np.count_nonzero to compare non zero pixels in each part

@berak I was in the process of doing that :P My code is just the regular one used for thresholding, create trackbars, convert image to HSV and then threshold.

Hi there!

I am wokring on an autonomous robot which has to run in a lane 2m wide with a yellow line marking the outer boundary and white line marking the inner boundary. What I have thought of is, that I threshold the image to obtain a binary image containing the lines. Then divide that binary image into 2 parts vertically and then find out the non zero pixels in each part, the part with less pixels indicate that that the lane is turning, and hence the robot turns to the side of less pixels.

I'm trying to do something like this-

Problem is with all my opencv+python knowledge I am stuck at the stage where I threshold the image. I don't know how to proceed after that. I tried searching the documentary but I didn't find anything.

So how can I achieve this?

Thanks.

It was nothing more than an indentation error in the for loop. :P

I tried that. It didn't work. But I have solved my problem. I made an empty function which returns nothing, and the I called that. @Tetragramm

Hi there!
I'm trying to threshold an image. I have used the cv2.createTrackbar function as-
cv2.createTrackbar('High H','image',0,179, None).
Now the last part is what I'm having trouble with. Further in my code, I use highH = cv2.getTrackbarPos('High H','image') to get my trackbar value and use it in the cv2.inRange function. So it becomes pretty obvious that I do not need to call a function as the last argument of the function. Now the problem is I can't seem to type in the function. I tried removing the last part, I got an error-

cv2.createTrackbar only works with 5 arguements. Only 4 given.
Hmm, okay seems like I can't skip a part.

Next I tried callback and nothing. I got this error:-
When used nothing:- NameError: name 'nothing' is not defined
When used callback:- NameError: name 'callback' is not defined

Okay after a while I tried using None. Got this error:- TypeError: on_change must be callable
So how do I use the cv2.createTrackbar function without calling a function?

Thanks!

Hi there! I'm getting this unexpected error in this code. I found some solutions after googling a whil, I've tried all of them but none work. Here is the code-

import cv2
import numpy as np
img = cv2.imread('circleTest.jpg',0)
img = cv2.medianBlur(img,5)
cimg = cv2.cvtColor(img,cv2.COLOR_GRAY2BGR)

circles = cv2.HoughCircles(img,cv2.HOUGH_GRADIENT,1,20, param1=50,param2=30,minRadius=0,maxRadius=0)

circles = np.uint16(np.around(circles))
for i in circles[0,:]: cv2.circle(cimg,(i[0],i[1]),i[2],(0,255,0),2)
cv2.circle(cimg,(i[0],i[1]),2,(0,0,255),3)
cv2.imshow('detected circles',cimg)
cv2.waitKey(0)
cv2.destroyAllWindows()

The code is borrowed from a tutorial about HoughCircles from here- http://opencv-python-tutroals.readthe...

Whenever I try to run the code I get an error saying Unsupported or Unrecognized array type, ( you must have seen the entire error). I tried copying DLLs, putting the image in the working directory, giving the path to the image but no good. I can run other programs like of object detection fine, on Python and C/C++.

Any help would be greatly appreciated.
Thanks!

@berak , I did, that was the first thing I did. But that's only for RGB and I couldn't possibly understand how to implement it when I have 6 parameters (High H,S,V and Low H,S,V) I maybe can type in the cv2.inRange function correctly but I can't really understand how to use my trackbars to change the required values

Hi there!
I'm kinda new to Python, have been struggling with this language along with the OpenCV libraries for a while now. I've come across various code examples for first converting an image (in my case camera feed from my laptop's webcam) from BGR to HSV and then using the cv2.inRange function to threshold the image to obtain a binary image with the desired colors filtered out. I can't understand the implementation of the cv2.createTrackbar function.

So my obvious question is-
How do I make trackbars to control the Min and Max H,S,V values to threshold my image. I cant understand how the trackbars affect the individual variables.

Hi there! I am working on a project which requires the detection of road signs. I was wondering that if I could take sample images of the road signs, save them onto my PC(final version to run on Raspberry Pi), and then as the camera runs, it keeps on checking the presence of the sign stored in the memory and when it is found takes appropriate action(For example if I implement it in a robot, whenever the robot sees a stop sign, it should stop for, say 10 seconds and then start moving again.)

Can anyone help me out on this?

Hi there. I am a complete beginner to OpenCV. I downloaded it today and tried installing it on my pc using the documentation on this site. After a while I figured out that the documentation was for version 2.14 and was very, very complex. They were using VS 2010 and I have 2012. So I headed to YouTube and found these tutorials-

1. Installing OpenCV - https://www.youtube.com/watch?v=EcFte...

2. Configuring it with Visual Studio- https://www.youtube.com/watchv=tHX3ML...

However, I had already completed the first part of the documentation(which does something with a command prompt and makes environment variables. But I ignored that and followed both the tutorials and when I tried to run the sample code. I got errors on two lines. Here's the code-

include iostream

include opencv2/opencv.hpp // VS says it cannot open source file.

void main()

{ std::cout << "OpenCV Version: " << CV_VERSION << std::endl;// VS says indetifier CV_VERSION is unidentified. }

This program is supposed to show the OpenCV version I have on my pc, but I cannot get it to run. Is my OpenCV installation valid? If not then how can I repair it? I have already checked the paths a lot of times. I just can' seem to figure out what to do. Any and all help will be greatly appreciated.

Thank you -YaddyVirus

Thanks a lot @berak

Hey there! I am new to OpenCV and while I originally intend to use it on a Raspberry Pi, I couldn't wait for my board to ship and thought that I'll start coding around on my PC itself. I just wanted to know that to run OpenCV will the Visual Studio Express 2012 Express edition is enough or will I have to download the full edition? Moreover the set up tutorial is very complex(I'm a geek and code in Java and C# but still it went over my head). Can anyone tell me a simplified procedure?