Skip to content

Commit

Permalink
Update README.md
Browse files Browse the repository at this point in the history
Hint to MediaPipe Modelmaker added
  • Loading branch information
robodhhb authored Nov 29, 2023
1 parent 85e9f7a commit d708b84
Showing 1 changed file with 8 additions and 0 deletions.
8 changes: 8 additions & 0 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -35,6 +35,10 @@ neue picamera2-Bibliothek. Die Implementierung und die Installationsanweisung is
Die Objekterkennung erfolgt auf der CPU des Raspberry Pi und ist daher etwas langsamer als die Implementerung
mit dem Coral USB Accelerator (4-4.5 fps statt 6.5 bis 7 fps auf Raspberry Pi 4B).

Um andere Zugelemente zu erkennen, muss ein Modell mit dem MediaPipe Modelmaker neu trainiert werden. Der
[Object detection model customization guide](https://developers.google.com/mediapipe/solutions/customization/object_detector/)
beschreibt den Vorgang im Detail.

### Welcome to the Smart-Modelrailway-Cam
We monitor track sections of a model railway system with artificial intelligence.
We use a Raspberry Pi 4B with the Picamera V2.
Expand Down Expand Up @@ -69,3 +73,7 @@ the installation description can be accessed in the folder "40_SMRC_MediaPipe_Ve
Object-detection runs on the CPU of the raspberry pi and performs therefore slower than on the
Coral USB Accelerator (4-4.5 fps instead of 6.5 - 7 fps on Raspberry Pi 4B).

To detect other train elements you have to retrain a model with the MediaPipe ModelMaker.
See [Object detection model customization guide](https://developers.google.com/mediapipe/solutions/customization/object_detector/) for explanations.


0 comments on commit d708b84

Please sign in to comment.