|
Loading images from a folder for object recognition
Anonymous |
13 year
|
I am trying to build an application that loads an image taken from another camera and send it as an input to the object recognition module in RR and gives the output as the object name that is the closest match.
Here is the approach I tried:
1. use the Load Image to load an image from a location.
2. Run the object recognition module.
3. Run the API to do getVariable (OBJECT_NAME) to get the result.
I tried used the load Image through the API (C#) as well . But the get variables method through the API doesnt return a result for OBJECT_NAME.
I think that the loadImage is not actually loading the image as an input to the object recognition module which is what I need to do to build the app.
Please let me know is this approach is incorrect and if yes, what is the best way to get the result using RR.
|
|
|
|
Anonymous |
13 year
|
Seems like you have the right approach. The trick is to find out in what stage things are going wrong. For example, if you startup RR in non-faceless module (as you would normally use it via your desktop) and then issue the load image via C# you should see the image suddenly appear in RR if that image load works. Whenever you issue an API command you should be able to see the change happen in RR (even button presses).
Keep in mind that using the load image module to load in an image will erase the pipeline when you issue that command. So if you are going to do it that way then be sure the load image pipeline also includes the object recognition pipeline.
An alternative way would be to have RR just load in a robofile that has the correct pipeline in it. You can use the execute or the LoadProgram API call to do so.
Finally, download the latest version of RR since there was an issue with the OR module not loading in previously trained templates.
Let me know if you get any further ...
STeven.
|
|
This Thread
