Merge branch 'main' of git.beagleboard.org:docs/docs.beagleboard.io

Dockerfile

@@ -2,6 +2,7 @@ FROM python:3.9-alpine as sphinx-build-env
 RUN apk update
 RUN python -m pip install --upgrade pip
 RUN pip install -U sphinx
+RUN pip install -U sphinx_design
 RUN pip install sphinxcontrib-svg2pdfconverter
 RUN apk add librsvg
 RUN pip install sphinx_rtd_theme

VERSION

 VERSION_MAJOR = 0
 VERSION_MINOR = 0
-PATCHLEVEL = 6
-VERSION_TWEAK = 1
+PATCHLEVEL = 7
+VERSION_TWEAK = 2
 EXTRAVERSION = beta

beaglebone-ai-64/edge_ai_apps/configuration_file.rst

+.. _ai_64_edgeai_configuration:
+
+Demo Configuration file
+#########################
+
+The demo config file uses YAML format to define input sources, models, outputs
+and finally the flows which defines how everything is connected. Config files
+for out-of-box demos are kept in ``edge_ai_apps/configs`` folder. The
+folder contains config files for all the use cases and also multi-input and
+multi-inference case. The folder also has a template YAML file
+``app_config_template.yaml`` which has detailed explanation of all the
+parameters supported in the config file.
+
+Config file is divided in 4 sections:
+
+#. Inputs
+#. Models
+#. Outputs
+#. Flows
+
+Inputs
+======
+
+The input section defines a list of supported inputs like camera, filesrc etc.
+Their properties like shown below.
+
+.. code-block:: yaml
+
+    inputs:
+        input0:                                         #Camera Input
+            source: /dev/video2                         #Device file entry of the camera
+            format: jpeg                                #Input data format suported by camera
+            width: 1280                                 #Width and Height of the input
+            height: 720
+            framerate: 30                               #Framerate of the source
+
+        input1:                                         #Video Input
+            source: ../data/videos/video_0000_h264.mp4  #Video file
+            format: h264                                #File encoding format
+            width: 1280
+            height: 720
+            framerate: 25
+
+        input2:                                         #Image Input
+            source: ../data/images/%04d.jpg             #Sequence of Image files, printf style formatting is used
+            width: 1280
+            height: 720
+            index: 0                                    #Starting Index (optional)
+            framerate: 1
+
+All supported inputs are listed in template config file.
+Below are the details of most commonly used inputs.
+
+.. _ai_64_edgeai_camera_sources:
+
+Camera sources (v4l2)
+---------------------
+
+**v4l2src** GStreamer element is used to capture frames from camera sources
+which are exposed as v4l2 devices. In Linux, there are many devices which are
+implemented as v4l2 devices. Not all of them will be camera devices. You need
+to make sure the correct device is configured for running the demo successfully.
+
+``init_script.sh`` is ran as part of systemd, which detects all cameras connected
+and prints the detail like below in the UART console:
+
+.. code-block:: bash
+
+   debian@beaglebone:/opt/edge_ai_apps# ./init_script.sh
+   USB Camera detected
+       device = /dev/video18
+       format = jpeg
+   CSI Camera 0 detected
+       device = /dev/video2
+       name = imx219 8-0010
+       format = [fmt:SRGGB8_1X8/1920x1080]
+       subdev_id = 2
+       isp_required = yes
+   IMX390 Camera 0 detected
+       device = /dev/video18
+       name = imx390 10-001a
+       format = [fmt:SRGGB12_1X12/1936x1100 field: none]
+       subdev_id = /dev/v4l-subdev7
+       isp_required = yes
+       ldc_required = yes
+
+script can also be run manually later to get the camera details.
+
+From the above log we can determine that 1 USB camera is connected
+(/dev/video18), and 1 CSI camera is connected (/dev/video2) which is imx219 raw
+sensor and needs ISP. IMX390 camera needs both ISP and LDC.
+
+Using this method, you can configure correct device for camera capture in the
+input section of config file.
+
+.. code-block:: bash
+
+    input0:
+        source: /dev/video18  #USB Camera
+        format: jpeg          #if connected USB camera supports jpeg
+        width: 1280
+        height: 720
+        framerate: 30
+
+    input1:
+        source: /dev/video2  #CSI Camera
+        format: auto         #let the gstreamer negotiate the format
+        width: 1280
+        height: 720
+        framerate: 30
+
+    input2:
+        source: /dev/video2  #IMX219 raw sensor that nees ISP
+        format: rggb         #ISP will be added in the pipeline
+        width: 1920
+        height: 1080
+        framerate: 30
+        subdev-id: 2         #needed by ISP to control sensor params via ioctls
+
+    input3:
+        source: /dev/video2  #IMX390 raw sensor that nees ISP
+        width: 1936
+        height: 1100
+        format: rggb12       #ISP will be added in the pipeline
+        subdev-id: 2         #needed by ISP to control sensor params via ioctls
+        framerate: 30
+        sen-id: imx390
+        ldc: True            #LDC will be added in the pipeline
+
+Make sure to configure correct ``format`` for camera input. ``jpeg`` for USB
+camera that supports MJPEG (Ex. C270 logitech USB camera). ``auto`` for CSI
+camera to allow gstreamer to negotiate the format. ``rggb`` for sensor
+that needs ISP.
+
+Video sources
+-------------
+
+H.264 and H.265 encoded videos can be provided as input sources to the demos.
+Sample video files are provided under ``/opt/edge_ai_apps/data/videos/video_0000_h264.mp4``
+and ``/opt/edge_ai_apps/data/videos/video_000_h265.mp4``
+
+.. code-block:: yaml
+
+    input1:
+        source: ../data/videos/video_0000_h264.mp4
+        format: h264
+        width: 1280
+        height: 720
+        framerate: 25
+
+    input2:
+        source: ../data/videos/video_0000_h265.mp4
+        format: h265
+        width: 1280
+        height: 720
+        framerate: 25
+
+Make sure to configure correct ``format`` for video input as shown above.
+By default the format is set to ``auto`` which will then use the GStreamer
+bin ``decodebin`` instead.
+
+Image sources
+-------------
+
+JPEG compressed images can be provided as inputs to the demos. A sample set of
+images are provided under ``/opt/edge_ai_apps/data/images``. The names of the
+files are numbered sequentially and incrementally and the demo plays the files
+at the fps specified by the user.
+
+.. code-block:: yaml
+
+    input2:
+        source: ../data/images/%04d.jpg
+        width: 1280
+        height: 720
+        index: 0
+        framerate: 1
+
+RTSP sources
+------------
+
+H.264 encoded video streams either coming from a RTSP compliant IP camera or
+via RTSP server running on a remote PC can be provided as inputs to the demo.
+
+.. code-block:: yaml
+
+    input0:
+        source: rtsp://172.24.145.220:8554/test # rtsp stream url, replace this with correct url
+        width: 1280
+        height: 720
+        framerate: 30
+
+.. note::
+
+   Usually video streams from any IP camera will be encrypted and cannot be
+   played back directly without a decryption key. We tested RTSP source by
+   setting up an RTSP server on a Ubuntu 18.04 PC by refering to this writeup,
+   `Setting up RTSP server on PC
+   <https://gist.github.com/Santiago-vdk/80c378a315722a1b813ae5da1661f890>`_
+
+Models
+======
+
+The model section defines a list of models that are used in the demo. Path to
+the model directory is a required argument for each model and rest are optional
+properties specific to given use cases like shown below.
+
+.. code-block:: yaml
+
+    models:
+        model0:
+            model_path: ../models/segmentation/ONR-SS-871-deeplabv3lite-mobv2-cocoseg21-512x512   #Model Directory
+            alpha: 0.4                                                                            #alpha for blending segmentation mask (optional)
+        model1:
+            model_path: ../models/detection/TFL-OD-202-ssdLite-mobDet-DSP-coco-320x320
+            viz_threshold: 0.3                                                                    #Visualization threshold for adding bounding boxes (optional)
+        model2:
+            model_path: ../models/classification/TVM-CL-338-mobileNetV2-qat
+            topN: 5                                                                               #Number of top N classes (optional)
+
+Below are some of the use case specific properties:
+
+#. **alpha**: This determines the weight of the mask for blending the semantic
+   segmentation output with the input image ``alpha * mask + (1 - alpha) * image``
+#. **viz_threshold**: Score threshold to draw the bounding boxes for detected
+   objects in object detection. This can be used to control the number of boxes
+   in the output, increase if there are too many and decrease if there are very
+   few
+#. **topN**: Number of most probable classes to overlay on image classification
+   output
+
+The content of the model directory and its structure is discussed in detail in
+:ref:`pub_edgeai_import_custom_models`
+
+
+Outputs
+=======
+
+The output section defines a list of supported outputs.
+
+.. code-block:: yaml
+
+   outputs:
+       output0:                                                     #Display Output
+           sink: kmssink
+           width: 1920                                              #Width and Height of the output
+           height: 1080
+           connector: 39                                            #Connector ID for kmssink (optional)
+
+       output1:                                                     #Video Output
+           sink: ../data/output/videos/output_video.mkv             #Output video file
+           width: 1920
+           height: 1080
+
+       output2:                                                     #Image Output
+           sink: ../data/output/images/output_image_%04d.jpg        #Image file name, printf style formatting is used
+           width: 1920
+           height: 1080
+
+All supported outputs are listed in template config file.
+Below are the details of most commonly used outputs
+
+Display Sink (kmssink)
+----------------------
+
+When you have only one display connected to the SK, kmssink will try to use
+it for displaying the output buffers. In case you have connected multiple
+display monitors (e.g. Display Port and HDMI), you can select a specific display
+for kmssink by passing a specific connector ID number.
+Following command finds out the connected displays available to use.
+
+**Note**: Run this command outside docker container. The first number in each
+line is the connector-id which we will use in next step.
+
+.. code-block:: bash
+
+    debian@beaglebone:/opt/edge_ai_apps# modetest -M tidss -c | grep connected
+    39      38      connected       DP-1            530x300         12      38
+    48      0       disconnected    HDMI-A-1        0x0             0       47
+
+From above output, we can see that connector ID 39 is connected. Configure the
+connector ID in the output section of the config file.
+
+Video sinks
+-----------
+The post-processed outputs can be encoded in H.264 format and stored on disk.
+Please specify the location of the video file in the configuration file.
+
+.. code-block:: yaml
+
+    output1:
+        sink: ../data/output/videos/output_video.mkv
+        width: 1920
+        height: 1080
+
+Image sinks
+-----------
+The post-processed outputs can be stored as JPEG compressed images.
+Please specify the location of the image files in the configuration file.
+The images will be named sequentially and incrementally as shown.
+
+.. code-block:: yaml
+
+    output2:
+        sink: ../data/output/images/output_image_%04d.jpg
+        width: 1920
+        height: 1080
+
+Flows
+=====
+
+The flows section defines how inputs, models and outputs are connected.
+Multiple flows can be defined to achieve multi input, multi inference like
+below.
+
+.. code-block:: yaml
+
+    flows:
+        flow0:                              #First Flow
+            input: input0                   #Input for the Flow
+            models: [model1, model2]        #List of models to be used
+            outputs: [output0, output0]     #Outputs to be used for each model inference output
+            mosaic:                         #Positions to place the inference outputs in the output frame
+                mosaic0:
+                    width:  800
+                    height: 450
+                    pos_x:  160
+                    pos_y:  90
+                mosaic1:
+                    width:  800
+                    height: 450
+                    pos_x:  960
+                    pos_y:  90
+        flow1:                              #Second Flow
+            input: input1
+            models: [model0, model3]
+            outputs: [output0, output0]
+            mosaic:
+                mosaic0:
+                    width:  800
+                    height: 450
+                    pos_x:  160
+                    pos_y:  540
+                mosaic1:
+                    width:  800
+                    height: 450
+                    pos_x:  960
+                    pos_y:  540
+
+Each flow should have exactly **1 input**, **n models** to infer the given input
+and **n outputs** to render the output of each inference. Along with input, models
+and outputs it is required to define **n mosaics** which are the position of the
+inference output in the final output plane. This is needed because multiple
+inference outputs can be rendered to same output (Ex: Display).
+
+Command line arguments
+----------------------
+
+Limited set of command line arguments can be provided, run with '-h' or '--help'
+option to list the supported parameters.
+
+.. code-block:: bash
+
+    usage: Run : ./app_edgeai.py -h for help
+
+    positional arguments:
+      config           Path to demo config file
+                           ex: ./app_edgeai.py ../configs/app_config.yaml
+
+    optional arguments:
+      -h, --help       show this help message and exit
+      -n, --no-curses  Disable curses report
+                       default: Disabled
+      -v, --verbose    Verbose option to print profile info on stdout
+                       default: Disabled

beaglebone-ai-64/edge_ai_apps/datasheet.rst

+.. _ai_64_edgeai_datasheet:
+
+Datasheet
+##########
+
+This chapter describes the performance measurements of the Edge AI Inference
+demos.
+
+Performance data of the demos can be auto generated by running following
+command on target:
+
+.. code-block:: bash
+
+    debian@beaglebone:/opt/edge_ai_apps/tests# ./gen_data_sheet.sh
+
+The performence measurements includes the following
+
+#. **FPS** : Effective framerate at which the application runs
+#. **Total time** : Average time taken to process each frame, which includes
+   pre-processing, inference and post-processing time
+#. **Inference time** : Average time taken to infer each frame
+#. **CPU loading** : Loading on different CPU cores present
+#. **DDR BW** : DDR read and write BW used
+#. **HWA Loading** : Loading on different Hardware accelerators present
+
+Following are the latest performance numbers of the C++ demos:
+
+Source             : **USB Camera**
+====================================
+
+Capture Framerate  : **30 fps**
+Resolution         : **720p**
+format             : **JPEG**
+
+.. figure:: ./images/edgeai_object_detection.png
+   :scale: 60
+   :align: center
+
+   GStreamer based data-flow pipeline with USB camera input and display output
+
+.. csv-table::
+    :header: "Model", "FPS", "Total time (ms)", "Inference time (ms)", "A72 Load (%)", "DDR Read BW (MB/s)", "DDR Write BW (MB/s)", "DDR Total BW (MB/s)", "C71 Load (%)", "C66_1 Load (%)", "C66_2 Load (%) ", "MCU2_0 Load (%)", "MCU2_1 Load (%)", "MSC_0 (%)", "MSC_1 (%)", "VISS (%)", "NF (%)", "LDC (%)", "SDE (%)", "DOF (%)"
+
+    ONR-CL-6150-mobileNetV2-1p4-qat,30.80,33.22,3.02,21.60,1596,619,2215,9.0,20.0,9.0,6.0,1.0,22.17,0,0,0,0,0,0
+    TFL-CL-0000-mobileNetV1-mlperf,30.69,33.19,1.04,15.93,1425,563,1988,5.0,22.0,9.0,6.0,1.0,21.90,0,0,0,0,0,0
+    TFL-OD-2020-ssdLite-mobDet-DSP-coco-320x320,30.69,33.25,5.00,10.24,1534,570,2104,15.0,29.0,9.0,6.0,1.0,22.67,0,0,0,0,0,0
+    TVM-CL-3410-gluoncv-mxnet-mobv2,30.58,33.21,2.02,22.80,1522,617,2139,6.0,20.0,9.0,6.0,1.0,21.84,0,0,0,0,0,0
+
+
+Source             : **Video**
+==============================
+
+Video Framerate    : **30 fps**
+Resolution         : **720p**
+Encoding           : **h264**
+
+.. figure:: ./images/edgeai_video_source.png
+   :scale: 60
+   :align: center
+
+   GStreamer based data-flow pipeline with video file input source and display output
+
+.. csv-table::
+    :header: "Model", "FPS", "Total time (ms)", "Inference time (ms)", "A72 Load (%)", "DDR Read BW (MB/s)", "DDR Write BW (MB/s)", "DDR Total BW (MB/s)", "C71 Load (%)", "C66_1 Load (%)", "C66_2 Load (%) ", "MCU2_0 Load (%)", "MCU2_1 Load (%)", "MSC_0 (%)", "MSC_1 (%)", "VISS (%)", "NF (%)", "LDC (%)", "SDE (%)", "DOF (%)"
+
+    ONR-CL-6150-mobileNetV2-1p4-qat,30.52,33.46,3.03,14.28,990,403,1393,2.0,7.0,4.0,1.0,1.0,10.27,0,0,0,0,0,0
+    TFL-CL-0000-mobileNetV1-mlperf,30.77,33.47,1.07,30.76,746,97,843,2.0,2.0,1.0,1.0,1.0,15.76,0,0,0,0,0,0
+    TFL-OD-2020-ssdLite-mobDet-DSP-coco-320x320,30.56,33.54,5.06,22.58,736,92,828,2.0,2.0,1.0,1.0,1.0,16.9,0,0,0,0,0,0
+    TVM-CL-3410-gluoncv-mxnet-mobv2,30.64,33.47,2.01,33.33,712,110,822,1.0,1.0,0.0,1.0,1.0,15.3,0,0,0,0,0,0
+
+Source             : **CSI Camera (ov5640)**
+============================================
+
+Capture Framerate  : **30 fps**
+Resolution         : **720p**
+format             : **YUYV**
+
+.. figure:: ./images/edgeai_ov5640_camera_source.png
+   :scale: 60
+   :align: center
+
+   GStreamer based data-flow pipeline for with CSI camera (OV5640) input and display output
+
+.. csv-table::
+    :header: "Model", "FPS", "Total time (ms)", "Inference time (ms)", "A72 Load (%)", "DDR Read BW (MB/s)", "DDR Write BW (MB/s)", "DDR Total BW (MB/s)", "C71 Load (%)", "C66_1 Load (%)", "C66_2 Load (%) ", "MCU2_0 Load (%)", "MCU2_1 Load (%)", "MSC_0 (%)", "MSC_1 (%)", "VISS (%)", "NF (%)", "LDC (%)", "SDE (%)", "DOF (%)"
+
+    ONR-CL-6150-mobileNetV2-1p4-qat,29.57,34.09,3.02,12.21,1671,699,2370,8.0,45.0,9.0,6.0,1.0,21.35,0,0,0,0,0,0
+    TFL-CL-0000-mobileNetV1-mlperf,29.41,34.15,1.01,10.27,1502,645,2147,5.0,47.0,9.0,6.0,1.0,20.96,0,0,0,0,0,0
+    TFL-OD-2020-ssdLite-mobDet-DSP-coco-320x320,29.36,34.65,5.00,10.5,1610,655,2265,14.0,53.0,9.0,6.0,1.0,21.47,0,0,0,0,0,0
+    TVM-CL-3410-gluoncv-mxnet-mobv2,29.38,34.17,2.01,11.66,1596,698,2294,6.0,45.0,9.0,5.0,1.0,21.10,0,0,0,0,0,0
+
+Source             : **CSI Camera with VISS (imx219)**
+======================================================
+
+Capture Framerate  : **30 fps**
+Resolution         : **1080p**
+format             : **SRGGB8**
+
+.. figure:: ./images/edgeai_rpi_camera_source.png
+   :scale: 60
+   :align: center
+
+   GStreamer based data-flow pipeline with IMX219 sensor, ISP and display
+
+.. csv-table::
+    :header: "Model", "FPS", "Total time (ms)", "Inference time (ms)", "A72 Load (%)", "DDR Read BW (MB/s)", "DDR Write BW (MB/s)", "DDR Total BW (MB/s)", "C71 Load (%)", "C66_1 Load (%)", "C66_2 Load (%) ", "MCU2_0 Load (%)", "MCU2_1 Load (%)", "MSC_0 (%)", "MSC_1 (%)", "VISS (%)", "NF (%)", "LDC (%)", "SDE (%)", "DOF (%)"
+
+    ONR-CL-6150-mobileNetV2-1p4-qat,30.64,33.19,3.01,15.72,1781,853,2634,9.0,16.0,9.0,13.0,1.0,31.78,0,22.37,0,0,0,0
+    TFL-CL-0000-mobileNetV1-mlperf,30.59,33.14,1.04,12.78,1612,798,2410,5.0,18.0,9.0,13.0,1.0,31.65,0,22.31,0,0,0,0
+    TFL-OD-2020-ssdLite-mobDet-DSP-coco-320x320,30.56,33.07,5.00,13.30,1730,809,2539,15.0,25.0,9.0,13.0,1.0,32.6,0,22.19,0,0,0,0
+    TVM-CL-3410-gluoncv-mxnet-mobv2,30.48,33.14,2.01,12.91,1708,852,2560,7.0,16.0,9.0,13.0,1.0,31.83,0,22.26,0,0,0,0
+
+Source             : **IMX390 over FPD-Link**
+=============================================
+
+Capture Framerate  : **30 fps**
+Resolution         : **1080p**
+format             : **SRGGB12**
+
+.. figure:: ./images/edgeai_imx390_camera_source.png
+   :scale: 60
+   :align: center
+
+   GStreamer based data-flow pipeline with IMX390 sensor, ISP, LDC and display
+
+.. csv-table::
+    :header: "Model", "FPS", "Total time (ms)", "Inference time (ms)", "A72 Load (%)", "DDR Read BW (MB/s)", "DDR Write BW (MB/s)", "DDR Total BW (MB/s)", "C71 Load (%)", "C66_1 Load (%)", "C66_2 Load (%) ", "MCU2_0 Load (%)", "MCU2_1 Load (%)", "MSC_0 (%)", "MSC_1 (%)", "VISS (%)", "NF (%)", "LDC (%)", "SDE (%)", "DOF (%)"
+
+    ONR-CL-6150-mobileNetV2-1p4-qat,30.59,33.15,3.09,25.18,2207,1102,3309,10.0,16.0,9.0,14.0,1.0,31.73,0,22.94,0,10.8,0,0
+    TFL-CL-0000-mobileNetV1-mlperf,30.53,33.15,1.21,16.20,2019,1040,3059,5.0,18.0,9.0,15.0,1.0,32.80,0,23.34,0,10.10,0,0
+    TFL-OD-2020-ssdLite-mobDet-DSP-coco-320x320,30.43,33.13,5.02,23.7,2201,1067,3268,15.0,25.0,9.0,14.0,1.0,32.80,0,22.88,0,9.95,0
+    TVM-CL-3410-gluoncv-mxnet-mobv2,30.44,33.16,2.12,21.50,2111,1100,3211,7.0,16.0,9.0,15.0,1.0,32.28,0,22.88,0,10.6,0,0

beaglebone-ai-64/edge_ai_apps/docker_environment.rst

+.. _ai_64_edgeai_docker_env:
+
+Docker Environment
+###################
+
+Docker is a set of "platform as a service" products that uses the OS-level
+virtualization to deliver software in packages called containers.
+Docker container provides a quick start environment to the developer to
+run the out of box demos and build applications.
+
+The Docker image is based on Ubuntu 20.04.LTS and contains different
+open source components like OpenCV, GStreamer, Python and pip packages
+which are required to run the demos. The user can choose to install any
+additional 3rd party applications and packages as required.
+
+.. _ai_64_edgeai_docker_build_ontarget:
+
+Building Docker image
+======================
+
+The `docker/Dockerfile` in the edge_ai_apps repo describes the recipe for
+creating the Docker container image. Feel free to review and update it to 
+include additional packages before building the image.
+
+.. note::
+    Building Docker image on target using the provided Dockerfile will take
+    about 15-20 minutes to complete with good internet connection. 
+    Building Docker containers on target can be slow and resource constrained.
+    The Dockerfile provided will build on target without any issues but if
+    you add more packages or build components from source, running out of memory
+    can be a common problem. As an alternative we highly recommend trying
+    QEMU builds for cross-compiling the images for arm64 architecture on a PC
+    and then load the compiled image on target.
+
+Initiate the Docker image build as shown,
+
+.. code-block:: bash
+
+    debian@beaglebone:/opt/edge_ai_apps/docker#./docker_build.sh
+
+Running the Docker container
+============================
+
+Enter the Docker session as shown,
+
+.. code-block:: bash
+
+    debian@beaglebone:/opt/edge_ai_apps/docker#./docker_run.sh
+
+This will start a Ubuntu 20.04.LTS image based Docker container and the prompt
+will change as below,
+
+.. code-block:: bash
+
+    [docker] debian@beaglebone:/opt/edge_ai_apps#
+
+
+The Docker container has been created in privilege mode, so that it has root
+capabilities to all devices on the target system like Network etc.
+The container file system also mounts the target file system of /dev, /opt to
+access camera, display and other hardware accelerators the SoC has to offer.
+
+.. note::
+
+    It is highly recommended to use the docker_run.sh script to launch the
+    Docker container because this script will take care of saving any changes
+    made to the filesystem. This will make sure that any modifications to
+    the Docker filesystem including new package installation, updates to
+    some files and also command history is saved automatically and is
+    available the next time you launch the container. The container will
+    be committed only if you exit from the container explicitly. If you restart
+    the board without exiting container, any changes done from last saved state
+    will be lost.
+
+.. note::
+
+    After building and running the docker container, one needs to run 
+    ``setup_script.sh`` before running any of the demo applications.
+    Please refer to :ref:`pub_edgeai_install_dependencies` for more details.
+
+.. _ai_64_edgeai_docker_additional_commands:
+
+Handling proxy settings
+=======================
+
+If the board running the Docker container is behind a proxy server, the default
+settings for downloading files and installing packages via apt-get will not work.
+If you are running the board from TI network, docker build and run scripts will
+automatically detect and configure necessary proxy settings
+
+For other cases, you need to modify the script ``/usr/bin/setup_proxy.sh``
+to add the custom proxy settings required for your network.
+
+Additional Docker commands
+==========================
+
+.. note::
+    This section is provided only for additional reference and not required to
+    run out-of-box demos
+
+**Commit Docker container**
+
+Generally, containers have a short life cycle. If the container has any local
+changes it is good to save the changes on top of the existing Docker image.
+When re-running the Docker image, the local changes can be restored.
+
+Following commands show how to save the changes made to the last container.
+Note that this is already done automatically by ``docker_run.sh`` when you exit
+the container.
+
+.. code-block:: bash
+
+    cont_id=`docker ps -q -l`
+    docker commit $cont_id edge_ai_kit
+    docker container rm $cont_id
+
+
+For more information refer:
+`Commit Docker image <https://docs.docker.com/engine/reference/commandline/commit/>`_
+
+**Save Docker Image**
+
+Docker image can be saved as tar file by using the command below:
+
+.. code-block:: bash
+
+    docker save --output <pre_built_docker_image.tar>
+
+For more information refer here.
+`Save Docker image <https://docs.docker.com/engine/reference/commandline/save/>`_
+
+**Load Docker image**
+
+Load a previously saved Docker image using the command below:
+
+.. code-block:: bash
+
+    docker load --input <pre_built_docker_image.tar>
+
+For more information refer here.
+`Load Docker image <https://docs.docker.com/engine/reference/commandline/load/>`_
+
+**Remove Docker image**
+
+Docker image can be removed by using the command below:
+
+.. code-block:: bash
+
+    Remove selected image:
+    docker rmi <image_name/ID>
+
+    Remove all image:
+    docker image prune -a
+
+For more information refer
+`rmi reference <https://docs.docker.com/engine/reference/commandline/rmi/>`_ and
+`Image prune reference <https://docs.docker.com/engine/reference/commandline/image_prune/>`_
+
+**Remove Docker container**
+
+Docker container can be removed by using the command below:
+
+.. code-block:: bash
+
+    Remove selected container:
+    docker rm <container_ID>
+
+    Remove all container:
+    docker container prune
+
+For more information refer here.
+`rm reference <https://docs.docker.com/engine/reference/commandline/rm/>`_ and
+`Container Prune reference <https://docs.docker.com/engine/reference/commandline/container_prune/>`_
+
+Relocating Docker Root Location
+===============================
+The default location for Docker files is **/var/lib/docker**. Any Docker images
+created will be stored here. This will be a problem anytime the SD card is
+updated with a new targetfs. If a secondary storage (SSD or USB based storage)
+is available, then it is recommended to relocate the default Docker root
+location so as to preserve any existing Docker images. Once the relocation
+has been done, the Docker content will not be affected by any future targetfs
+updates or accidental corruptions of the SD card.
+
+The following steps outline the process for Docker root directory relocation
+assuming that the current Docker root is not at the desired location. If the
+current location is the desired location then exit this procedure.
+
+1. Run 'Docker info' command and inspect the output. Locate the line with
+   content **Docker Root Dir**. It will list the current location.
+
+2. To preserve any existing images, export them to .tar files for importing
+   later into the new location.
+
+3. Inspect the content under /etc/docker to see if there is a file by name
+   **daemon.json**. If the file is not present then create **/etc/docker/docker.json**
+   and add the following content. Update the 'key:value' pair for the key "graph"
+   to reflect the desired root location. If the file already exists, then make
+   sure that the line with "graph" exists in the file and points to the desired
+   target location.
+
+.. code-block:: json
+
+  {
+    "graph": "/run/media/nvme0n1/docker_root",
+    "storage-driver": "overlay",
+    "live-restore": true
+  }
+
+In the configuration above, the key/value pair
+**'"graph": "/run/media/nvme0n1/docker_root"'** defines the root location
+**'/run/media/nvme0n1/docker_root'.**
+
+4. Once the daemon.json file has been copied and updated, run the following
+   commands
+
+.. code-block:: bash
+
+   $ systemctl restart docker
+   $ docker info
+
+Make sure that the new Docker root appears under **Docker Root Dir** value.
+
+5. If you exported the existing images in step (2) then import them and they
+   will appear under the new Docker root.
+
+6. Anytime the SD card is updated with a new targetfs, steps (1), (3), and
+   (4) need to be followed.
+
+**Additional references**
+
+| https://docs.docker.com/engine/reference/commandline/images/
+| https://docs.docker.com/engine/reference/commandline/ps/

beaglebone-ai-64/edge_ai_apps/getting_started.rst

+.. _ai_64_edgeai_getting_started:
+
+Getting Started
+#################
+
+.. _ai_64_edgeai_getting_started_harware:
+
+Hardware setup
+===============
+
+BeagleBone® AI-64 has TI's TDA4VM SoC which houses dual core A72, high performance vision
+accelerators, video codec accelerators, latest C71x and C66x DSP, high bandwidth
+realtime IPs for capture and display, GPU, dedicated safety island and security
+accelerators. The SoC is power optimized to provide best in class performance
+for perception, sensor fusion, localization and path planning tasks in robotics,
+industrial and automotive applications.
+
+For more details visit https://www.ti.com/product/TDA4VM
+
+.. _ai_64_edgeai_hw_requirements_eaik:
+
+BeagleBone® AI-64
+-----------------
+
+BeagleBone® AI-64 brings a complete system for developing artificial intelligence (AI) 
+and machine learning solutions with the convenience and expandability of the BeagleBone® 
+platform and the peripherals on board to get started right away learning and building 
+applications. With locally hosted, ready-to-use, open-source focused tool chains and 
+development environment, a simple web browser, power source and network connection 
+are all that need to be added to start building performance-optimized embedded 
+applications. Industry-leading expansion possibilities are enabled through 
+familiar BeagleBone® cape headers, with hundreds of open-source hardware examples 
+and dozens of readily available embedded expansion options available off-the-shelf.
+
+To run the demos on BeagleBone® AI-64 you will require,
+
+    - BeagleBone® AI-64
+    - USB camera (Any V4L2 compliant 1MP/2MP camera, Eg. Logitech C270/C920/C922)
+    - Full HD eDP/HDMI display
+    - Minimum 16GB high performance SD card
+    - 100Base-T Ethernet cable connected to internet
+    - UART cable
+    - External Power Supply or Power Accessory Requirements
+
+        a. Nominal Output Voltage: 5VDC
+        b. Maximum Output Current: 5000 mA
+
+Connect the components to the SK as shown in the image.
+
+.. figure:: ./images/board_connections_bbai_64.jpg
+   :align: center
+
+   BeagleBone® AI-64 for Edge AI connections
+
+.. _ai_64_edgeai_usb_camera:
+
+USB Camera
+----------
+
+UVC (USB video class) compliant USB cameras are supported on the BeagleBone® AI-64.
+The driver for the same is enabled in linux image. The linux image has been tested with
+C270/C920/C922 versions of Logitech USB cameras. Please refer to
+:ref:`pub_edgeai_multiple_usb_cams` to stream from multiple USB cameras
+simultaneously.
+
+.. _ai_64_edgeai_imx219_sensor:
+
+IMX219 Raw sensor
+------------------
+
+**IMX219 camera module** from **Raspberry pi / Arducam** is supported by BeagleBone® AI-64. 
+It is a 8MP sensor with no ISP, which can transmit raw SRGGB8 frames over CSI lanes at 1080p 60 fps.
+This camera module can be ordered from
+https://www.amazon.com/Raspberry-Pi-Camera-Module-Megapixel/dp/B01ER2SKFS
+The camera can be connected to any of the 2 RPi zero 22 pin camera headers on BB AI-64 as
+shown below
+
+.. figure:: 
+   :scale: 20
+   :align: center
+
+   TODO: IMX219 CSI sensor connection with BeagleBone® AI-64 for Edge AI
+
+Note that the headers have to be lifted up to connect the cameras
+
+.. note:: To be updated
+    By default IMX219 is disabled. After connecting the camera you can enable it
+    by specifying the dtb overlay file in
+    ``/run/media/mmcblk0p1/uenv.txt`` as below,
+
+    ``name_overlays=k3-j721e-edgeai-apps.dtbo k3-j721e-sk-rpi-cam-imx219.dtbo``
+
+    Reboot the board after editing and saving the file.
+
+Two RPi cameras can be connected to 2 headers for multi camera usecases
+
+Please refer :ref:`pub_edgeai_camera_sources` to know how to list all the cameras
+connected and select which one to use for the demo.
+
+By default imx219 will be configured to capture at 8 bit, but it also supports
+10 bit capture in 16 bit container. To use it in 10 bit mode, below steps are
+required:
+
+    - Modify the ``/opt/edge_ai_apps/scripts/setup_cameras.sh`` to set the
+      format to 10 bit like below
+
+    .. code-block:: bash
+
+           CSI_CAM_0_FMT='[fmt:SRGGB8_1X10/1920x1080]'
+           CSI_CAM_1_FMT='[fmt:SRGGB8_1X10/1920x1080]'
+
+    - Change the imaging binaries to use 10 bit versions
+
+    .. code-block:: bash
+
+           mv /opt/imaging/imx219/dcc_2a.bin /opt/imaging/imx219/dcc_2a_8b.bin
+           mv /opt/imaging/imx219/dcc_viss.bin /opt/imaging/imx219/dcc_viss_8b.bin
+           mv /opt/imaging/imx219/dcc_2a_10b.bin /opt/imaging/imx219/dcc_2a.bin
+           mv /opt/imaging/imx219/dcc_viss_10b.bin /opt/imaging/imx219/dcc_viss.bin
+
+    - Set the input format in the ``/opt/edge_ai_apps/configs/rpiV2_cam_example.yaml``
+      as ``rggb10``
+
+
+Software setup
+==============
+
+.. _ai_64_edgeai_prepare_sd_card:
+
+Preparing SD card image
+-----------------------
+
+Download the ``bullseye-xfce-edgeai-arm64`` image from the links below and
+flash it to SD card using `Balena etcher <https://www.balena.io/etcher/>`_ tool.
+
+- To use via SD card: `bbai64-debian-11.4-xfce-edgeai-arm64-2022-08-02-10gb.img.xz <bbai64-debian-11.4-xfce-edgeai-arm64-2022-08-02-10gb.img.xz>`_
+- To flash on eMMC: `bbai64-emmc-flasher-debian-11.4-xfce-edgeai-arm64-2022-08-02-10gb.img.xz <https://rcn-ee.net/rootfs/bb.org/testing/2022-08-02/bullseye-xfce-edgeai-arm64/bbai64-emmc-flasher-debian-11.4-xfce-edgeai-arm64-2022-08-02-10gb.img.xz>`_
+
+
+The Balena etcher tool can be installed either on Windows/Linux. Just download the
+etcher image and follow the instructions to prepare the SD card.
+
+.. figure:: ./images/balena_etcher.png
+   :scale: 100
+   :align: center
+
+   Balena Etcher tool to flash SD card with Processor linux image Linux for Edge AI
+
+The etcher image is created for 16 GB SD cards, if you are using larger SD card,
+it is possible to expand the root filesystem to use the full SD card capacity
+using below steps
+
+.. code-block:: bash
+
+   #find the SD card device entry using lsblk (Eg: /dev/sdc)
+   #use the following commands to expand the filesystem
+   #Make sure you have write permission to SD card or run the commands as root
+
+   #Unmount the BOOT and rootfs partition before using parted tool
+   umount /dev/sdX1
+   umount /dev/sdX2
+
+   #Use parted tool to resize the rootfs partition to use
+   #the entire remaining space on the SD card
+   #You might require sudo permissions to execute these steps
+   parted -s /dev/sdX resizepart 2 '100%'
+   e2fsck -f /dev/sdX2
+   resize2fs /dev/sdX2
+
+   #replace /dev/sdX in above commands with SD card device entry
+
+.. _ai_64_edgeai_poweron_boot:
+
+Power ON and Boot
+-----------------
+Ensure that the power supply is disconnected before inserting the SD card.
+Once the SD card is firmly inserted in its slot and the board is powered ON,
+the board will take less than 20sec to boot and display a wallpaper as
+shown in the image below.
+
+.. figure:: 
+   :scale: 25
+   :align: center
+
+   TODO: BeagleBone® AI-64 wallpaper upon boot
+
+You can also view the boot log by connecting the UART cable to your PC and
+use a serial port communications program.
+
+For **Linux OS minicom** works well.
+Please refer to the below documentation on 'minicom' for more details.
+
+https://help.ubuntu.com/community/Minicom
+
+When starting minicom, turn on the colors options like below:
+
+.. code-block:: bash
+
+   sudo minicom -D /dev/ttyUSB2 -c on
+
+For **Windows OS Tera Term** works well.
+Please refer to the below documentation on 'TeraTerm' for more details
+
+https://learn.sparkfun.com/tutorials/terminal-basics/tera-term-windows
+
+.. note::
+    Baud rate should be configured to 115200 bps in serial port communication
+    program. You may not see any log in the UART console if you connect to it
+    after the booting is complete or login prompt may get lost in between boot
+    logs, press ENTER to get login prompt
+
+As part of the linux systemd ``/opt/edge_ai_apps/init_script.sh`` is executed
+which does the below,
+
+    - This kills weston compositor which holds the display pipe. This step will
+      make the wallpaper showing on the display disappear and come back
+    - The display pipe can now be used by 'kmssink' GStreamer element while
+      running the demo applications.
+    - The script can also be used to setup proxies if connected behind a
+      firewall.
+
+Once Linux boots login as ``root`` user with no password.
+
+.. _ai_64_edgeai_connecting_remotely:
+
+Connect remotely
+----------------
+If you don't prefer the UART console, you can also access the device with the
+IP address that is shown on the display.
+
+With the IP address one can ssh directly to the board, view the contents and run
+the demos.
+
+For best experience we recommend using VSCode which can be downloaded from
+here.
+
+https://code.visualstudio.com/download
+
+You also require the "Remote development extension pack" installed in VSCode
+as mentioned here:
+
+https://code.visualstudio.com/docs/remote/ssh
+
+.. figure:: 
+   :scale: 90
+   :align: center
+
+   TODO: Microsoft Visual Studio Code for connecting to BeagleBone® AI-64 for Edge AI via SSH

beaglebone-ai-64/edge_ai_apps/images/edge_ai_demos.drawio

+<mxfile host="Electron" modified="2021-05-19T15:45:06.473Z" agent="5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) draw.io/14.6.13 Chrome/89.0.4389.128 Electron/12.0.7 Safari/537.36" etag="dAgc-ZEs3_cQbep0zx-v" version="14.6.13" type="device" pages="3"><diagram name="CPP Demo Data Flow" id="XUTmI-InrPcYLR0GOj1Y">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</diagram><diagram name="Python Demo Data Flow" id="OihEl4gLeUUpuXvin1QQ">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</diagram><diagram id="ZmDZ4aD7BjtF5CznldHT" name="Page-3">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</diagram></mxfile>
\ No newline at end of file