IdentifEYEr
- NVIDIA Jetson Nano 2GB
- Arducam Camera
- Arducam Upgraded Camera Pan Tilt Platform
- GH-S37D Servo Motor (x2)
- HC-05 Bluetooth Module
- ATmega328P
- Push Button
- Joystick
- NVIDIA Jetson --- Python
-Pytorch
-OpenCV
-sklearn
-SSD MobileNetV1
- ATmega328P --- ANSI C
- <avr/io.h>
-<util/delay.h>
-<avr/sfr_defs.h>
-<string.h>
User Interface Software Block Diagram
Jetson Searching Software Block Diagram
- Identify 8 different items in various locations in the room with a confidence level of 90%
- Items can be of the same class of items used in the development phase, but need to be significantly unique.
- Lighting conditions must vary to represent realistic situations.
- Able to generalize to multiple backgrounds, meaning that the confidence does not drop by more than 5% with a different background
- Able to locate the item within 1 minute of the user selecting the item
-
UI system is smaller than 6 inches by 12 inches
- Detection and camera system must be smaller than 6 inches by 12 inches
- Transmit signal over bluetooth 20/20 times