
This SDK has been patched by Embedded Artists for the iMXRT1062 Developer's Kit.
The SDK was released on 2022-11-18 and is based on NXP's 2.12.1 SDK (SDK_2_12_1_MIMXRT1062xxxxA.zip).

This is what has been patched:
* Set CPU speed according to Commercial/Industrial CPU
* Correction of the VDD_SOC_IN voltage.
* Flash settings (speed, algorithm, size, driver) to work with the 4MB OctalSPI ATXP032
* LWIP projects - added reading of the MAC address from the onboard I2C EEPROM
* Added an I2C driver for the gpio expander (PCA6416) and code to use it
* Modified pin muxing
* SEMC projects - changed algorithm for memory test and now test entire 32MB instead of only 4KB
* Adjusted the USB interface number for USB Host examples (it is different for host and device examples)
* Added a software_reset() function in board.c/.h to issue a JEDEC reset before NVIC_SystemReset()
* Changed the Wi-Fi examples to use the Embedded Artists 1XK M.2 Module (EAR00385) as default
* 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).
* Embedded Wizard project 'ew_gui_smart_thermostat' was incorrectly setup for EVKB
* Changed the default display to RK043FN02H as it is the one mounted on the Developer's Kits

This has been added:
* New WDOG examples that work
* I2C probe example
* Example to show the use of software_reset()

This has been removed:
* All projects for the EVK - only keeping EVKB which is then patched
* The original WDOG and RTWDOG examples as those were not working

Important things to note:
* Read section "8 - Known Issues" in docs/MCUXpresso SDK Release Notes for EVK-MIMXRT1060.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:
* J22 (micro USB) is the default UART and unless specified otherwise it is setup for 115200 8/N/1


!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
THIS PROJECT IS NOT DIRECTLY COMPATIBLE WITH THE HARDWARE AND WILL NOT WORK.

This example for the EVKB requires requires a camera but that is not
supported on the iMX OEM Carrier Board.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!



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



Overview
========
Convolutional neural network (CNN) example with the use of
convolution, ReLU activation, pooling and fully-connected functions.

The CNN model in the example was trained using the scripts available at [1]
with the CifarNet model. 
The configuration of the model was modified to match the neural 
network structure in the CMSIS-NN CIFAR-10 example.
The example source code is a modified version of the Label Image
example from the TensorFlow Lite examples [2], adjusted to run on MCUs.
The neural network consists of 3 convolution layers interspersed by
ReLU activation and max pooling layers, followed by a fully-connected layer
at the end. The input to the network is a 32x32 pixel color image, which is 
classified into one of the 10 output classes. The model size is 91 KB.

Firstly a static ship image is used as input regardless camera is connected or not.
Secondly runtime image processing from camera in the case camera and display
is connected. Camera data are displayed on LCD. 

HOW TO USE THE APPLICATION:
To classify an image, place an image in front of the camera so that it fits in the
white rectangle in the middle of the LCD display. 
Note that semihosting implementation causes slower or discontinuous video experience. 
Select UART in 'Project Options' during project import for using external debug console 
via UART (virtual COM port).

[1] https://github.com/tensorflow/models/tree/master/research/slim
[2] https://github.com/tensorflow/tensorflow/tree/r2.3/tensorflow/lite/examples/label_image

Files:
  main.cpp - example main function
  ship.bmp - shrinked picture of the object to recognize
    (source: https://en.wikipedia.org/wiki/File:Christian_Radich_aft_foto_Ulrich_Grun.jpg)
  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('ship.bmp')
    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    with open('image_data.h', 'w') as fout:
      print('#define STATIC_IMAGE_NAME "ship"', file=fout)
      print('static const uint8_t image_data[] = {', file=fout)
      img.tofile(fout, ', ', '0x%02X')
      print('};\n', file=fout)
  timer.c - timer source code
  image/* - image capture and pre-processing code
  model/get_top_n.cpp - top results retrieval
  model/model_data.h - model data converted from a .tflite file to a C language
    array using the xxd tool (distributed with the Vim editor at www.vim.org)
  model/model.cpp - model initialization and inference code
  model/model_cifarnet_ops.cpp - model operations registration
  model/output_postproc.cpp - model output processing
  video/* - camera and display handling


Toolchain supported
===================
- IAR embedded Workbench  9.30.1
- Keil MDK  5.37
- GCC ARM Embedded  10.3.1
- MCUXpresso  11.6.0

Hardware requirements
=====================
- Mini/micro USB cable
- EVK-MIMXRT1060 board
- Personal computer
- MT9M114 camera (optional)
- RK043FN02H-CT display (optional)

Board settings
==============
Connect the camera to J35 (optional)
Connect the display to A1-A40 and B1-B6 (optional)
Connect external 5V power supply to J2, set J1 to 1-2

Prepare the Demo
================
1. Connect a USB cable between the host PC and the OpenSDA USB port on the target board. 
2. Open a serial terminal with the following settings:
   - 115200 baud rate
   - 8 data bits
   - No parity
   - One stop bit
   - No flow control
3. Download the program to the target board.
4. Either press the reset button on your board or launch the debugger in your IDE to begin running the demo.

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

CIFAR-10 object recognition example using a TensorFlow Lite model.
Detection threshold: 60%
Expected category: ship
Model: cifarnet_quant

Static data processing:
----------------------------------------
     Inference time: 65 ms
     Detected:       ship (100%)
----------------------------------------


Camera data processing:
Data for inference are ready
----------------------------------------
     Inference time: 64 ms
     Detected:        dog (95%)
----------------------------------------

Data for inference are ready
----------------------------------------
     Inference time: 65 ms
     Detected: No label detected (0%)
----------------------------------------

Data for inference are ready
----------------------------------------
     Inference time: 64 ms
     Detected:      horse (60%)
----------------------------------------
