README.md

metadata

license: agpl-3.0
language:
  - en
pipeline_tag: object-detection
tags:
  - Axera
  - License Plate Recognition
  - NPU
  - OCR
  - Object Detection

plate-axera

This version of plate-axera has been converted to run on the Axera NPU using w8a16 quantization. There are two models included:

1. pld_650_npu3.axmodel: This model is trained to detect the license plate with label 'plate'.
2. plr_650_npu3.axmodel: This model is trained to recognize the characters of detected license plates, and also shows the color of the license plate.

Supported Classes

Detection model supports the following classes:

1. plate

Supported characters

Recognition model supports the following characters:

{"皖沪津渝冀晋蒙辽吉黑苏浙京闽赣鲁豫鄂湘粤桂琼川贵云藏陕甘青宁新警学港澳台使领挂OABCDEFGHJKLMNPQRSTUVWXYZ0123456789"}

Supported colors

Recognition model supports the following colors:

['blue', 'green', 'yellow', 'white', 'black']

Compatible with Pulsar2 version: 5.2.

Convert tools links:

For those who are interested in model conversion, you can try to export axmodel through:

• The repo of AXera Platform, where you can get the detailed guide.
• Pulsar2 Link, How to Convert ONNX to axmodel

Support Platform

https://docs.m5stack.com/zh_CN/ai_hardware/AI_Pyramid-Pro

• AX650N/AX8850
  • M4N-Dock(爱芯派Pro)
  • AI Pyramid
  • M.2 Accelerator card

How to use

Download all files from this repository to the device.

python env requirement

pyaxengine

https://github.com/AXERA-TECH/pyaxengine

wget https://github.com/AXERA-TECH/pyaxengine/releases/download/0.1.3.rc2/axengine-0.1.3-py3-none-any.whl
pip install axengine-0.1.3-py3-none-any.whl

Inference with AX650 Host, such as M4N-Dock(爱芯派Pro)

Plate Detection

Input image:

run

python3 axmodel_infer_pld.py

root@ax650:~/plate-axera# python3 axmodel_infer_pld.py
[INFO] Available providers:  ['AxEngineExecutionProvider', 'AXCLRTExecutionProvider']
[INFO] Using provider: AxEngineExecutionProvider
[INFO] Chip type: ChipType.MC50
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Engine version: 2.12.0s
[INFO] Model type: 2 (triple core)
[INFO] Compiler version: 5.2 eccb31f5
class: plate left:597 top:417 right:759 bottom:475 conf: 88%

Output image:

Plate Recognition

Input image:

run

python3 axmodel_infer_plr.py

root@ax650:~/plate-axera# python3 axmodel_infer_plr.py
[INFO] Available providers:  ['AxEngineExecutionProvider', 'AXCLRTExecutionProvider']
[INFO] Using provider: AxEngineExecutionProvider
[INFO] Chip type: ChipType.MC50
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Engine version: 2.12.0s
[INFO] Model type: 2 (triple core)
[INFO] Compiler version: 5.2 eccb31f5
Plate: [苏A8A68Y], score: 0.9997, color: [blue], score:1.0000

Plate det & Rec End2End

Input image:

run

python3 axmodel_infer_plate_end2end.py

root@ax650:~/plate-axera# python3 axmodel_infer_plate_end2end.py
[INFO] Available providers:  ['AxEngineExecutionProvider', 'AXCLRTExecutionProvider']
[INFO] Using provider: AxEngineExecutionProvider
[INFO] Chip type: ChipType.MC50
[INFO] VNPU type: VNPUType.DISABLED
[INFO] Engine version: 2.12.0s
[INFO] Model type: 2 (triple core)
[INFO] Compiler version: 5.2 eccb31f5
[INFO] Using provider: AxEngineExecutionProvider
[INFO] Model type: 2 (triple core)
[INFO] Compiler version: 5.2 eccb31f5
Det---class:[plate], bbox:[597,417,759,475], conf:0.88
Rec---Plate:[川A2E7V7], score:0.9991, color:[blue], score:1.0000
Result saved to: ./plate_end2end_res.jpg

Output image: