This paper presents a unified framework for license plate detection and character recognition on low-power edge hardware. The proposed network uses a shared CSP-Darknet and PANet feature encoder with an anchor-free localization head and a RoIAlign-based sequence-recognition branch. Character decoding is performed directly on aligned feature maps, avoiding image-space plate cropping and a second recognition backbone. The model is trained on an AI-assisted dataset of approximately 37,000 manually verified images covering license plate layouts from all 50 United States. The dataset includes variations in illumination, viewpoint, weather, blur, occlusion, and plate design. The framework is converted to an INT8 TensorRT engine and evaluated on an NVIDIA Jetson Orin Nano with 8 GB memory. The INT8 model shows an mAP50 of 97.8% and a full-sequence accuracy of 90.7%. It records 147 FPS at 6.8 ms end-to-end latency, with 315 MB inference memory and 9.4 W average power consumption. Relative to the YOLOv8s+CRNN baseline, the proposed deployment provides a 7.1x throughput increase and a 77.8% reduction in allocated inference memory
Figures & tables
Fig. 1 : End-to-end architecture and runtime execution pipeline of the proposed unified ALPR framework, organized into pre-processing, model inference, and post-processing stages. Bilinear RoIAlign samples plate features directly from the shared PANet pyramid F using bounding box coordinates B generated by the anchor-free localization head, eliminating duplicate backbone evaluations and image-space cropping.
Fig. 2 : An overview of results and data saved from detection
Property
Value
Images
37,000
States
50
Labels
Box, state, plate text
Split
70:15:15
TABLE I : Summary of the dataset.
Architecture
mAP50 (%)
Precision (%)
Char. Acc. (%)
Seq. Acc. (%)
YOLOv8s + CRNN
98.2
97.6
94.1
88.4
YOLOv8s + PARSeq
98.4
97.9
96.8
92.1
YOLO-LPR without RoIAlign
96.5
95.8
91.2
83.7
Proposed framework, FP32
98.1
97.4
96.3
91.5
Proposed framework, INT8
97.8
97.1
95.8
90.7
TABLE II : Localization and recognition results on the held-out test set.
Configuration
Precision
E2E Latency (ms)
Throughput (FPS)
Memory (MB)
Power (W)
YOLOv8s + CRNN
FP16
48.3
20.7
1,420
14.2
YOLOv8s + PARSeq
FP16
112.6
8.8
2,150
14.8
Proposed framework
FP32
22.4
44.6
840
13.1
Proposed framework
FP16
11.2
89.2
510
11.8
Proposed framework
INT8
6.8
147.0
315
9.4
TABLE III : Batch-one edge Inference on the Jetson Orin Nano.
Extraction Method
Backbone Passes
Intermediate Copy (MB/s)
Latency (ms)
Seq. Acc. (%)
RGB image crop
2
412.5
18.4
91.2
Discrete RoIPool
1
0.0
5.1
84.6
Bilinear RoIAlign
1
0.0
5.6
91.5
TABLE IV : Effect of plate-feature extraction method.
Fig. 3 : Allocated inference memory and batch-one throughput across the evaluated ALPR configurations.
Fig. 4 : End-to-end latency decomposition of the proposed INT8 deployment into frame preparation, model execution, and output decoding.
Precision
Calibration
Images
mAP50 (%)
Seq. Acc. (%)
FP32
Not required
–
98.1
91.5
FP16
Not required
–
98.1
91.4
INT8
MinMax
100
94.2
78.3
INT8
Entropy
100
96.5
86.1
INT8
Entropy
1,000
97.8
90.7
TABLE V : Effect of numerical precision and INT8 calibration.
License plate character detection is a crucial component of intelligent transportation systems, where high accuracy and computational efficiency are required for real-time deployment. Although recent deep learning-based methods have substantially improved detection performance, many high-accuracy models rely on large-scale architectures that incur substantial computational overhead, limiting their applicability to resource-constrained devices. In this paper, we propose MicroCharNet, an ultra-lightweight model specifically designed for license plate character detection. The proposed architecture employs a compact backbone composed of C2f blocks, integrated with CoordAtt module to enhance feature extraction while preserving spatial information. A lightweight C3k2-based neck fuses multi-level features, followed by a single-level anchor-free detection head that enables end-to-end prediction. Experiments conducted on the UFPR-ALPR dataset demonstrate that MicroCharNet achieves competitive detection accuracy with only 0.08M parameters and 0.096 GFLOPs, while outperforming several recent YOLO-based baselines. Hardware-level evaluations further confirm its efficiency for real-time deployment on edge devices. These results indicate that carefully designed ultra-lightweight architectures can effectively balance accuracy and efficiency in license plate character detection. The source code is available at https://github.com/chequanghuy/MicroCharNet.
Huy Che, Dinh-Duy Phan, Duc-Lung Vu
University of Information Technology, Ho Chi Minh City, Vietnam · 2Vietnam National University, Ho Chi Minh City, Vietnam
Vehicle license plate recognition is an integral component of intelligent transportation systems. In this work, we present an embedded real-time license plate recognition system customized for developing countries. We address the challenge of handling complex, unstructured traffic scenes with diverse vehicle types while implementing the system on an embedded platform for low-cost deployment. Our method consists of license plate detection on a multi-vehicle image, followed by character recognition on the detected license plates. Both steps use lightweight convolutional neural networks to balance accuracy and efficiency. We also introduce the SL-LPR dataset of Sri Lankan road images, which contains a variety of vehicle types and traffic conditions typically seen in developing countries. On this dataset, the license plate detection and character recognition models achieved 93.6% mAP and 87.88% accuracy, respectively, and were competitive against larger models on several public datasets. To achieve real-time performance in a resource-constrained embedded environment, we applied low-bitwidth quantization using the Brevitas library and implemented FPGA acceleration for the models using the FINN framework. The end-to-end system can operate at 11.5~FPS when implemented on the Xilinx Kria KV260 platform. These results demonstrate that our system is effective for real-time license plate recognition on an embedded device, even in complex traffic scenarios. The SL-LPR dataset is available for research use at: https://github.com/sl-lpr-uom/SL-LPR.git.
In this paper, we present the first (to the best of our knowledge) demonstration of a low-power MCU-based edge device for Automatic License Plate Recognition (ALPR). The design leverages on a 9-core RISC-V processor, GAP8, coupled with a QVGA ultra-low-power greyscale imager. The proposed visual processing pipeline uses a multi-model inference approach based on SSDlite-MobilenetV2 for license plate detection and LPRNet for optical character recognition, reaching a 38.9% mAP score for the first task and a recognition rate of >99.13% for the latter on public datasets. On real-world data, the pipeline recognizes registration numbers when the size of LP crops is as small as 30x5 pixels. Thanks to the applied compression and optimization strategies, the multi-model inference (687 MMAC) achieves a throughput of 1.09 FPS at a power cost of 117 mW when running on GAP8. Our solution is the first MCU-class device embedding such a level of network complexity, resulting to be 73x more energy-efficient w.r.t. precedent mobile-class ALPR system featuring a Raspberry Pi3. The proposed design does not resort to any hardwired acceleration engines, thus retaining full flexibility for future algorithmic improvements.
Lorenzo Lamberti, Manuele Rusci, Marco Fariselli +2
Department of Electrical, Electronic and Information Engineering - University of Bologna, Italy · GreenWaves Technologies, Grenoble, France · Integrated System Laboratory - ETH Zurich, Switzerland