This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
The better the resolution of the camera the better are the results.See the camera resolution of 640480 would also do good if you have haar-trained or train-cascaded your micro-controller(I mean the brain of UAV) well.For that you would require atleast 2000 positive images of the object(2020 to 100*100 sized images,if the resolution of images is more,the time of training will be more),and a set of atleast 4000 negative images.Once trained properly your UAV would be able to detect the object even with poor resolution.The whole process of haartraining can be viewed from the following blog:- http://abhishek4273.wordpress.com/2014/02/10/opencv-haar-training/
If you have any further questions ping me up on the same blog or here. I hope this helps you :)
| | 2 | No.2 Revision |
The better the resolution of the camera the better are the results.See results.
See the camera resolution of 640480 640*480 would also do good if you have haar-trained or train-cascaded your micro-controller(I mean the brain of UAV) well.For well.
For that you would require atleast 2000 positive images of the object(2020 to 100*100 100100 sized images,if the resolution of images is more,the time of training will be more),and a set of atleast 4000 negative images.Once trained properly your UAV would be able to detect the object even with poor resolution.The resolution.
The whole process of haartraining can be viewed from the following blog:- http://abhishek4273.wordpress.com/2014/02/10/opencv-haar-training/
If you have any further questions ping me up on the same blog or here. I hope this helps you :)
