Hardware/Neuro Vision TFT

Neuro Vision TFT Getting Started

The Neuro Vision TFT carrier with the OV5640 camera, an 800x480 LCD with capacitive touch, a speaker, the strobe LED and time of flight sensor, and the TFT-W variant with the wireless module.

beginner3 min read

Neuro Vision TFT

The Neuro Vision TFT is a carrier for the Neuro N6 that mounts the OV5640 camera module together with an 800x480 colour LCD, a Goodix GT911 capacitive touch controller and a speaker driven by a MAX98357A amplifier. Like the Neuro Vision OV5640 it carries the strobe LED and a VL53L1CB time of flight sensor. It is the only carrier family with a display, and it allows a sketch to show the camera image and model results with no computer attached. The Neuro Vision TFT-W is the same carrier with a u-blox MAYA-W476 WiFi and Bluetooth LE module added.

Specifications

TFTTFT-W
CameraOmniVision OV5640, 5 MP, autofocus. Carrier wires the CSI lanes invertedSame
Display800x480 LCD on the 16-bit RGB565 LTDC interface, two layers: camera and ARGB4444 overlaySame
TouchGoodix GT911 capacitive, up to five points, on the module I2C bus at 0x5D or 0x14Same
Refresh30 to 60 Hz, following the camera frame rate under TRANSPORT_LCDSame
SpeakerMAX98357A I2S class D amplifier on SAI1 with a speakerSame
Strobe LEDOn the sensor strobe outputSame
Time of flightST VL53L1CB, 4x4 zonesSame
WirelessNoneu-blox MAYA-W476, WiFi 2.4 GHz and Bluetooth LE
Part numberOL-DEV-N6-006, hardware revision AOL-DEV-N6-007, hardware revision B

The inertial unit, magnetometer and microphone are on the Neuro N6 board itself.

Fitting

The carrier mates with the two board-to-board connectors on the Neuro N6 with the board unpowered. The display faces away from the Neuro N6.

In a sketch

Both variants are declared with the carrier argument:

cpp
#include <OV5640_Arduino.h>

OV5640 camera(CARRIER_TFT);

The argument records the inverted CSI lane wiring and enables the display. Declaring a plain OV5640 camera; on this carrier produces no frames.

The ObjectSegmentationDisplay example, under File > Examples > NeuroN6Examples > 03.NeuralNetwork, runs segmentation and draws the result on the panel. cfg.transport = TRANSPORT_LCD selects the panel as the output. See The On-board Display.

Display

The camera image is written to the panel continuously by hardware and is double buffered, so it does not tear. Post-processors draw boxes, masks and skeletons on the overlay layer above it, and a sketch can draw its own graphics there with the ohmlab_lcd_draw primitives. The panel shows WVGA (800x480) at full size and VGA centred with bars; larger camera modes are for USB streaming and switch the panel off.

The panel can also mirror the image while video streams to a host, with cfg.lcd = true.

Touch

The GT911 library reads up to five touch points and provides pressed() and clicked() rectangle tests, so an on-screen button is a drawn rectangle and a test against the same rectangle. The default coordinate mapping is correct on this carrier. See Touch Input.

Speaker

The Speaker library plays tones and streams 16-bit mono PCM through the amplifier. See Speaker.

Wireless (TFT-W)

The MAYA-W476 provides WiFi video streaming and Bluetooth LE. The WiFiAccessPoint example runs an access point, a BLE peripheral, the camera, the panel and the speaker together. See WiFi.

Examples for this carrier

ExampleLibraryRequires
CameraOptions, ObjectDetectionDisplay, ObjectSegmentationDisplayNeuroN6ExamplesTFT or TFT-W
QrReaderDisplayQuircTFT or TFT-W
BlobTrackerDisplayN6ImgprocTFT or TFT-W
FaceRecognitionFaceIDTFT or TFT-W
CalibrateTouchGT911TFT or TFT-W
SpeakerTone, SpeakerClipSpeakerTFT or TFT-W
AudioSpectrum, AudioWaterfallN6AudioTFT or TFT-W
AudioEventDetectionDisplayN6AudioAITFT or TFT-W
WiFiAccessPointMayaW476TFT-W

Downloads and Regulatory Compliance.