Click here to Skip to main content
15,848,905 members
Please Sign up or sign in to vote.
0.00/5 (No votes)
Hi.
I want detect the silver objects in different places like this Canister in this picture :

http://8pic.ir/images/87258923242677915413.png[^]

whats the best way to do this with open cv ?
Posted
Updated 25-Dec-13 9:35am
v4
Comments
[no name] 25-Dec-13 8:40am    
Please define "silver shapes"...
mohammadali1375 25-Dec-13 8:58am    
for example a ball with this material : http://8pic.ir/images/87258923242677915413.png
[no name] 25-Dec-13 9:38am    
Can you imagine now the problem? Your picture is "just" an example, but you need to define "silver shape" in a mathematical way.
[no name] 25-Dec-13 12:47pm    
To be serious: The only way is through reverse ray tracing. And that is not an easy job.
mohammadali1375 25-Dec-13 14:44pm    
with Laser ?

1 solution

If I understand you question correctly, you are working on a machine learning project and part of the objective is to detect objects of silver materials, say silver ball, silver spoon etc.
However, before you can do that, you have to train your machine to differentiate objects of different materials, say silver, glass, wood etc. This is called supervised training or supervised learning in AI.

Generally, the approach towards building a supervised training AI project is as follows:
1. Collect a lot of data with known categories, in your case, images of objects of known but different materials the quantity of each should be about the same so as to maintain data balance;
2. Split the data into a training set and test set, say 60% training set and 40% test set;
3. Train the machine learning model on the training set;
4. Test the trained model on the test set to measure the accuracy.

There are many supervised training AI techniques that you can use to achieve this, ANN and SVM are just some of them.

Since you are interested in OpenCV, there are no better place than to visit its official website for more information.
 
Share this answer
 
v2

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)



CodeProject, 20 Bay Street, 11th Floor Toronto, Ontario, Canada M5J 2N8 +1 (416) 849-8900