search

FirstSanny / Robi

Project for RoboCup
0 stars 0 forks source link

Interpreting Imagedata #3

Open FirstSanny opened 7 years ago

FirstSanny commented 7 years ago

We have to find our target objects at the images that we took.

Ju1yy commented 7 years ago

First of all, we have to register our targets. This is possible with void ALTrackerProxy::registerTarget(const std::string& TargetName, const AL::ALValue& Param) Luckily, the red ball is already a Target-Constant for the robot, so we just have to register it. Couldn't find out how we can register a new object, but it must be possible.

Hahaha, I also found ALFaceDetection which is specific for faces detection. So maybe we can use the professor as our target ;)

RebecaTU commented 7 years ago

I think the function that you said we can use for red ball is ALRedBallDetection, isn't it? it is good because we can fix our own size of Ball. We should try this function with Choregraphe Box on Tuesday!.

For face detection, it is said that is complicated to detect the face of a person in 3D because if the person turn the head, the image is another. or if the person is smiling in the photo and then sad in real it is also complicated. that is why we should use a photo of him in a papier and paste it in the class. It is said that alfacedetectiion returns the position, so after that we just walk to the target.

See here to learn a new object and two recognize it. We will need this to find the second object file:///home/rebeca/Documentos/programa/hands_on_nao/aldeb-doc-2.1.2.17/aldeb-doc-2.1.3.0/software/choregraphe/tutos/recognize_objects.html#choregraphe-tuto-learn-object