Merge pull request #11 from jwillikers/hailo2

Add support for using the Raspberry Pi AI accelerator

.justfile

@@ -43,7 +43,7 @@ init-dev: && sync
     set -euxo pipefail
     distro=$(awk -F= '$1=="ID" { print $2 ;}' /etc/os-release)
     if [ "$distro" = "debian" ]; then
-        sudo apt-get --yes install ffmpeg libatlas-base-dev python3-dev python3-picamera2 python3-venv
+        sudo apt-get --yes install ffmpeg hailo-all libatlas-base-dev python3-dev python3-picamera2 python3-venv
     fi
     [ -d venv ] || python -m venv --system-site-packages venv
     venv/bin/python -m pip install --requirement requirements-dev.txt
@@ -66,11 +66,13 @@ install-system: init
     sudo cp xmp_utils.py /usr/local/bin/xmp_utils.py
     sudo chown root:root /usr/local/bin/xmp_utils.py
     sudo mkdir --parents /usr/local/etc/detectionator/
-    sudo cp config/fast-and-close-config.toml /usr/local/etc/detectionator/config.toml
+    sudo cp config/hailo-config.toml /usr/local/etc/detectionator/config.toml
     sudo chown --recursive root:root /usr/local/etc/detectionator
     sudo mkdir --parents /usr/local/share/detectionator/
-    sudo cp --recursive models /usr/local/share/detectionator/models
+    sudo cp models/* /usr/local/share/detectionator/models
     sudo chown --recursive root:root /usr/local/share/detectionator
+    sudo mkdir /var/log/detectionator
+    sudo chown detectionator:detectionator /var/log/detectionator
     sudo -H -u detectionator bash -c '[ -d /home/detectionator/venv ] || python -m venv --system-site-packages /home/detectionator/venv'
     sudo -H -u detectionator bash -c '/home/detectionator/venv/bin/python -m pip install --requirement requirements.txt'
     sudo systemctl enable detectionator.service
@@ -80,7 +82,7 @@ install-user: init
     [ -d venv ] || python -m venv --system-site-packages venv
     venv/bin/python -m pip install --requirement requirements.txt
     mkdir --parents {{ config_directory() }}/detectionator
-    cp config/fast-and-close-config.toml {{ config_directory() }}/detectionator/config.toml
+    cp config/hailo.toml {{ config_directory() }}/detectionator/config.toml
     mkdir --parents {{ config_directory() }}/systemd/user
     ln --force --relative --symbolic etc/systemd/user/* {{ config_directory() }}/systemd/user/
     systemctl --user daemon-reload

.vscode/settings.json

@@ -26,6 +26,7 @@
         "gpsdctl",
         "grayscale",
         "gstreamer",
+        "hailo",
         "hynix",
         "Hyperpixel",
         "ifdef",

LICENSE.adoc

@@ -1,7 +1,7 @@
 = GNU General Public License
 
 _Version 3, 29 June 2007_
-_Copyright © 2007 Free Software Foundation, Inc. &lt;<http://fsf.org/>&gt;_
+_Copyright © 2007 Free Software Foundation, Inc. <http://fsf.org/>_
 
 Everyone is permitted to copy and distribute verbatim copies of this license
 document, but changing it is not allowed.
@@ -584,11 +584,11 @@ for a GUI interface, you would use an “about box”.
 You should also get your employer (if you work as a programmer) or school, if any, to
 sign a “copyright disclaimer” for the program, if necessary. For more
 information on this, and how to apply and follow the GNU GPL, see
-&lt;<http://www.gnu.org/licenses/>&gt;.
+<http://www.gnu.org/licenses/>.
 
 The GNU General Public License does not permit incorporating your program into
 proprietary programs. If your program is a subroutine library, you may consider it
 more useful to permit linking proprietary applications with the library. If this is
 what you want to do, use the GNU Lesser General Public License instead of this
 License. But first, please read
-&lt;<http://www.gnu.org/philosophy/why-not-lgpl.html>&gt;.
+<http://www.gnu.org/philosophy/why-not-lgpl.html>.

README.adoc

@@ -41,6 +41,7 @@ The project is largely based off of the tensorflow example in the {picamera2} pr
 
 The Detectionator is based off of the Raspberry Pi 5 and the Camera Module 3.
 It incorporates a GPS HAT for the Raspberry Pi to obtain the geolocation data to embed in the photographs.
+The Raspberry Pi AI Kit is recommended, as it reduces the CPU usage dramatically on the Pi.
 
 [NOTE]
 ====
@@ -66,6 +67,7 @@ The https://www.samsung.com/us/computing/memory-storage/memory-cards/pro-ultimat
 This is 4 17mm M2.5 Hex socket-socket standoffs and 8 6mm M2.5 screws.
 * https://shop.pimoroni.com/products/booster-header?variant=47414520906[Booster Header]
 This header has a standard 8.5mm height with 5mm tall pins.
+* https://www.raspberrypi.com/products/ai-kit/[Raspberry Pi AI Kit] This is component is optional.
 
 [TIP]
 ====

config/fast-and-close-config.toml

@@ -1,6 +1,6 @@
 # This config template is tweaked for capturing fast-moving creatures, such as birds.
 # Note that without global shutter, there can be distortions in the image when the creature is moving quickly.
-# Rolling shutter also makes it more diffult to successfully detect fast-moving objects.
+# Rolling shutter also makes it more difficult to successfully detect fast-moving objects.
 [detectionator]
 
 # Capture several photographs upon a successful detection.

config/hailo-config.toml

@@ -0,0 +1,43 @@
+# This config template is tweaked for capturing fast-moving creatures, such as birds.
+# Note that without global shutter, there can be distortions in the image when the creature is moving quickly.
+# Rolling shutter also makes it more difficult to successfully detect fast-moving objects.
+[detectionator]
+
+# Capture several photographs upon a successful detection.
+# Re-running the detection algorithm before taking another photograph may lead to a considerable delay before another picture.
+burst = 1
+
+# Use the exact ratio for the Hailo chip.
+align_resolutions = false
+
+# Focus the lens on the closest part of the range.
+# todo Test how well the Macro range works.
+# autofocus-range = "macro"
+autofocus-range = "full"
+
+autofocus-mode = "auto"
+
+# Focus the lens as quickly as possible.
+autofocus-speed = "fast"
+
+detection-threshold = 0.4
+
+encoder-quality = "high"
+
+focal-detection-threshold = 0.25
+
+# capture-mode = "video"
+#stream = "mediamtx.lan.jwillikers.io:8554/detectionator"
+#rtsp-tcp = false
+
+# Don't use a gap after a successful detection but start searching again immediately.
+gap = 0
+
+hailo = true
+
+# Below are standard options.
+label = "/usr/local/share/detectionator/models/coco_labels_hailo.txt"
+log-level = "info"
+match = ["bird", "bear", "cat", "cow", "dog", "elephant", "giraffe", "horse", "sheep", "zebra"]
+model = "/usr/share/hailo-models/yolov8s_h8l.hef"
+output = "~/Pictures"