Phase 6
Phase 6
Project Integration |week 11-12|
Project Integration |week 11-12|
Embedded Files
- Creating the final transfer learning network
- Robotic control station
- Send signal from Matlab to Arduino
OCS
OCS
The whole OCS System is composed of mechanisms that implement image acquisition, classification, sorting, and rejection. Their workflow is shown on figure 1.3. The image is captured using the web camera and pre-processed to go through our Convolutional Neural Network, titled OCS-CNN. Then, the classification process follows. It determines if the object placed on our platform is one of the three objects it was trained for. In case that the OCS network has 99% confidence or greater, the robotic arm will place the item in its allocated container. In the case that the item isn’t one of our three objects, a lever rejects the item off the platform to set space for another object to be classified.
OCS Systems Architecture is composed of image acquisition, deep learning, transfer learning, classification, rejection, and sorting
OCS Control station
OCS Control station
The OCS control station is created in Matlab by using the figure function and uicontrol tool to create the user interface buttons. The control station will be used to connect wireless communications and to start/stop the deep learning classification process.
Figure 1. OCS control station
OCS deep learning flow chart
OCS deep learning flow chart
In phase 1 of the deep learning flow chart, we have spent time to test the existing CNN's. However, the existing CNN's could not satisfy the requirement for object classification at an accuracy greater than 99%. Therefore, we have implemented a transfer learning to improve the classification accuracy. In phase 2, we have used the deep learning techniques to classify our target objects into their classes and send a dedicated signal to the robotic arm according to the classifications.
Control flow chart.pdfGoogle Sites
Report abuse