Asked: 2019-09-07 01:09:50 -0600
Seen: 618 times
Last updated: Sep 07 '19
Area of a single pixel object in OpenCV
how to understand which functions available in python bindings?
Problems installing opencv on mac with python
build problems for android_binary_package - Eclipse Indigo, Ubuntu 12.04
OpenCV DescriptorMatcher matches
Can't compile .cu file when including opencv.hpp
Weird result while finding angle
cv2.perspectiveTransform() with Python
Using OpenCV's stitching module, strange error when compositing images
In what circumstances? Using what hardware?
real time video is sent to the laptop .... i'm using intel core i7 7700HQ and Nvidia GTX 1050 but i don't know the rightapproach
real time video is sent to the laptop .... i'm using intel core i7 7700HQ and Nvidia GTX 1050 but i don't know the rightapproach
using a webcam? at what resolution? where are these logos? On dirty barrels/containers in a warehouse with poor lighting or in a clean bright lab? You need to provide more information. Possible approaches: template matching, keypoint extractoin and matching deep neural networks (e.g. Yolo) It all depends, however, on where and how you are going to recognize these logos.
camera resolutions is not so high and lighting condition varies alot :( also logos might have been turned around or be in gray scale... and they are printed on paper i tried template matching and feature matching BUT not good result :(
if so, go for yolo (or other deep network), with sufficient number of training samples of your logos, it should be able to do a decent job.
OK , TNX