Very impressive programming. Did you use a hough detection to do this or did you find a faster way? I am in need of a faster way of detecting elipses. I would need an image analyzed almost in real time (about .25 sec and image at most). If I can do this I can take this project I am working on to the next level. Wondering If you can help?
I didn't use the hough method. Rather, i wrote an algorithm that goes through a number of steps to find the ellipse by using other means of identification, such as its calculated area vs. actual area (with a small amount of tolerance of course). Also, for the robotics competition, the camera will never be at a location where the height of the ellipse will be significantly less than the width. I used simple trigonometry in a complex sort of way to ensure that the object was circular as well.
Oh, i forgot to mention.... the algorithm works on the image pretty quickly, that was a 640x480 image and it was able to have about 20 fps. That in combination with the actual network restrictions in retrieving the image consistently (i did not use a video stream or that sort), the algorithm is pretty time-efficient.
Very impressive programming. Did you use a hough detection to do this or did you find a faster way? I am in need of a faster way of detecting elipses. I would need an image analyzed almost in real time (about .25 sec and image at most). If I can do this I can take this project I am working on to the next level. Wondering If you can help?
WinderKid2000 2 years ago
I didn't use the hough method. Rather, i wrote an algorithm that goes through a number of steps to find the ellipse by using other means of identification, such as its calculated area vs. actual area (with a small amount of tolerance of course). Also, for the robotics competition, the camera will never be at a location where the height of the ellipse will be significantly less than the width. I used simple trigonometry in a complex sort of way to ensure that the object was circular as well.
wolf1oo 2 years ago
Oh, i forgot to mention.... the algorithm works on the image pretty quickly, that was a 640x480 image and it was able to have about 20 fps. That in combination with the actual network restrictions in retrieving the image consistently (i did not use a video stream or that sort), the algorithm is pretty time-efficient.
wolf1oo 2 years ago