
This SDK has been patched by Embedded Artists for the iMXRT1176 Developer's Kit.
The SDK was released on 2022-03-02 and is based on NXP's 2.11.0 SDK (SDK_2_11_0_MIMXRT1176xxxxx.zip).

This is what has been patched:
* LWIP projects - added reading of the MAC address from I2C EEPROM either on the 100Mbit adapter
  or on the uCOM board
* Added an I2C driver for the gpio expander (PCA6416) and code to use it
* Added an I2C driver for the PWM gpio expander (PCA9530) and code to use it
* SDRAM size has been corrected to 32Mb (including linker files, MPU and DCD)
* 1G Ethernet PHY has been changed from RTL8211F to AR8031DS
* SEMC projects - changed to correct settings for the SDRAM
* SEMC projects - changed algorithm for memory test and now test entire 32MB instead of only 4KB
* Examples using eLCDIF/LCDIFv2 have been updated to use PCA6416/PCA9530 for
  RST/PWR/BL signals
* BOARD_USER_BUTTON has been redirected to SW5/WAKEUP button on the uCOM Carrier Board
* USER_LED has been changed to the blue RGB LED using PCA6416
* Touch: I2C bus and GPIOs have been changed for RST/INT
* Camera pins
* Adjusted the USB interface number (it is different for host and device examples)
* Changed the Wi-Fi examples to use the Embedded Artists 1XK M.2 Module (EAR00385) as default
* Corrected the ethernet PHY addresses
* Changed CORE clock depending on speed grading of MCU (798MHz for Industrial, 996MHz for Commercial)
* Changed SEMC clock to be within maximum speed for SDRAM (now 148.5MHz, was 198MHz)
* Many of the projects have been updated to use a more complete pin_mux.c file where all
  necessary pins have been initialized. The SDK examples used to only configure the pins
  that they use (and often not every pin) and most of the time the configuration was only
  for MUX:ing and not the PAD settings (pull up/down/none, drive strength and slew).
* Converted the AzureRTOS examples to use the BOARD_NETWORK_USE_100M_ENET_PORT (same as all
  other networking examples) instead of using the old EXAMPLE_USE_1G_ENET_PORT

This has been added:
* HDMI support to most GUI examples. HDMI at 1024x768@60 is the default resolution but
  that can be changed per project in display_support.h/elcdif_support.h/lcdifv2_support.h.
* Added ADT example for TensorFlow Lite
* I2C probe example
* EDID reader example
* Wi-Fi (serial) examples for the CMWC1ZZABR-107-EVB (a.k.a ABR Module)

This has been removed:
* All projects for the expansion board AGM01

Important things to note:
* Read section "8 - Known Issues" in docs/MCUXpresso SDK Release Notes for MIMXRT1170-EVK.pdf
  to see known issues with the current version of the SDK.
* For Iperf examples, set compiler optimization to -O3 or similar to improve performance.
* If the hardware seems unresponsive and the debugger cannot connect/flash/erase the current program
  then the most likely cause is the running program preventing the access. To stop the currently
  running program and regain control:
  1) Press and hold down the ISP_ENABLE button (SW1)
  2) Press and hold down the RESET button (SW3)
  3) Let go of the RESET button
  4) Wait an extra second or two
  5) Release the ISP_ENABLE button
  6) The hardware is now in a mode where programming/erasing it should work


Connectors:
* J29 (micro USB) is the default UART for the CM7 core and unless specified otherwise it is setup for 115200 8/N/1
* J30 (micro USB) is the default UART for the CM4 core and unless specified otherwise it is setup for 115200 8/N/1
* For 1Gbit Ethernet examples, use connector J25 on uCOM Carrier Board
* For 100Mbit Ethernet examples, use ethernet adapter connected between J12 on uCOM Carrier Board
  and J37 on the adapter. These four connections are also needed:
    1) uCOM Carrier Board, JP38:1 -> adapter JP37:1
    2) uCOM Carrier Board, JP38:2 -> adapter JP37:2
    3) uCOM Carrier Board, JP27:1 -> adapter JP39:2
    4) uCOM Carrier Board, JP27:2 -> adapter JP39:1
* The two CSI examples can use either an OV5640 camera in connector J23 or a camera in connector J24
* The EIQ examples that use a camera expects the camera in connector J24 (J23 might work for some
  of the examples but runs much slower)
* Some GUI examples are configured for the RK055AHD091 display which should be in connector "C" on
  the uCOM Carrier Board. Ignore the readme text about connecting extra 5V power.
* The default for GUI examples is to use an HDMI adapter in connector "C" on
  the uCOM Carrier Board. Ignore the readme text about connecting extra 5V power.


Everything below this line is the original content of the readme file.
=======================================================================



Overview
================
DeepViewRT Camera label iamge Demo for IMXRT. This project shows the demonstration of real-time object recognition using DeepViewRT on an IMXRT platform together with a camera and LCD module.  The firmware uses a user defined model and real-time camera to classify object.

This program is a deepviewrt demonstration program which uses the SDRAM/TCM to allocate the memory pool. The cache can be allocated on SDRAM or TCM, the performance will be different then.The firmware will use PXP API's to crop the region of interest (ROI) to appropriate size for the neural network (resize by calls to PXP API's). The results of labels can be shown in the LCD, and some other processed results will be redirected to UART console and/or toolchain GUI console. 

Cache can be placed on DTC or SDRAM.

Experiments
===========
The deepview-rt-cortex-m7f library is located in 'libs' folder. By default, a pre-trained mobilenetv1 width multiplier of 0.25, pixel of 160x160 model for classification is used and located in 'models' folder.

Files:
  main.cpp - example main function
  image_data.h - image file converted to a C language array of RGB values
    using Python with the OpenCV and Numpy packages:
    import cv2
    import numpy as np
    img = cv2.imread('giant_panda.jpeg')
    img = cv2.resize(img, (160, 160))
    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    with open('image_data.h', 'w') as fout:
      print('#define STATIC_IMAGE_NAME "giant panda"\n', file=fout)
      print('static const uint8_t image_data[] = {', file=fout)
      img.tofile(fout, ', ', '0x%02X')
      print('};\n', file=fout)
  mobilenet_v1_0.25_160.rtm - pre-trained DeepviewRT model
  video/* - camera and display handling

Running the demo
================
The log below shows the output of the demo in the terminal window (compiled with MDK):

==========================================================================
                    DeepviewRT camera label image Demo
===========================================================================

Static data processing:
Label index = 389, score = 0.986, runtime: 195
        MATCH - giant panda - 389

Camera data processing:

Data for inference are ready
Label index = 795, score = 0.050, runtime: 202
        NO MATCH - shower curtain - 795

Data for inference are ready
Label index = 795, score = 0.061, runtime: 201
        NO MATCH - shower curtain - 795

Data for inference are ready
Label index = 509, score = 0.403, runtime: 201
        MATCH - computer keyboard - 509

