Object regognition and object handeling by robot arm and grabber "hand"
In many situations people are used to idetify objects and handle them.
An example is picking fruit.
We need a ( camera and ) sofware that can accuratly recognise “un-standarised” objects like kiwifruit and other fruits.
For this purpose TensorFlow can be used, and an additional ( Phyton ) script must be written to trigger events, so the robot conducts a sequence of movements to actually pick the fruit recognized.
Kiwifruit have a set of common visible similarities that the sofware (TensorFlow) must identify.
When the object is identified, a serie of events are trigged.
These events sould be able to be control from a sofware on a PC or laptop.
The procedure will be as follow.
1. Camera and softaware recognise the fruit on the branch between the leaves
2. Estimate the location in distance and directions.
3. Organize the robot arm and grabber “hand” to approach object / apple
4. Feel the object with sensors in the grabber “hand”
5. Grab the object with the grabber
6. Rotate the grabber to allow the apple stalk to lose from the tree twig
7. Retract the the robot arm
8. Grabber drop the apple on designated place
9. Sarching for new object near by...