Week 3

 In week 3, we successfully resolved the issue of accessing the Raspberry Pi desktop. Additionally, we started training a license plate detection model using YOLO, specifically YOLOv8, to improve our system's accuracy. This week's progress included preparing the dataset, training the model, analyzing the results, and implementing Non-Maximum Suppression (NMS) to refine detection


Lab Day:

Resolving Raspberry Pi Desktop Access

  1. SSH Connection: First, we connected to the Raspberry Pi system using SSH, allowing us to access its terminal remotely.

  2. VNC Viewer Setup: After setting up SSH access, we used VNC Viewer to remotely control the Raspberry Pi's desktop interface, which enable us to manage our training tasks efficiently.


Training the License Plate Detection Model

Choosing YOLO for Object Detection

We explored different object detection models and decided on YOLO due to its real-time efficiency and high detection accuracy. 

Data Preparation and Training

  • We collected and preprocessed a dataset containing labeled images of license plates.

  • Using YOLOv8, we initiated model training with our dataset.

  • The model's training progress was tracked across multiple epochs, and the best-performing YOLO weight file was saved based on validation performance.

  • A log was maintained for each epoch, recording loss values and accuracy metrics.

Performance Analysis and Loss Curve Visualization

  • We drew the loss curves over the training epochs to monitor convergence and detect potential overfitting.

  • The model was evaluated on a separate test set to analyze its generalization capability.

Challenges Encountered

Overlapping Bounding Boxes Issue

  • Reduced confidence scores: Since multiple overlapping boxes diluted the confidence of each detection, the model was less certain about the correct identification.

  • Redundant detections: The presence of multiple bounding boxes for a single license plate increased processing time and complexity.

Applying Non-Maximum Suppression (NMS)

  • Eliminates redundant bounding boxes.

  • Retains only the most confident detection for each license plate.

  • Improves detection accuracy by reducing false positives and ensuring a single box per license plate.


Next Steps

With the Raspberry Pi system fully accessible and the license plate detection model trained and optimized, our next steps will focus on:

  • Further improving recognition accuracy by fine-tuning YOLO's hyperparameters.

  • Integrating the trained model into a real-time application.

评论

发表评论

此博客中的热门博文

Week 4

图片
In Week 4, we faced major challenges in Optical Character Recognition (OCR) for license plates. The initial OCR strategy delivered poor results due to multiple factors, including motion blur, occlusion, perspective distortion, and character misalignment. To overcome these issues, we implemented license plate character correction, refined segmentation techniques, and integrated more advanced text detection models. Eventually, we selected the CRAFT model for robust text detection and further optimized our pipeline to ensure high accuracy. Additionally, we extended our work to video processing to enable real-time application. Challenges in OCR and Initial Solutions 1. Factors Affecting OCR Performance Our initial OCR implementation struggled due to: Blurriness and Occlusion : Vehicles at varying distances caused blurred license plate images, while partial occlusions further degraded recognition accuracy. Perspective Distortion : Different camera angles led to skewed and curved ...
阅读全文

Week 2

图片
 In week 2, we finally get the equipment(Raspberry PI, LCD screen. camera, SD card) for our project, but we encountered multiple challenges, including SD card burning failures and difficulties accessing the Raspberry Pi system desktop. At the same time, we began transitioning from traditional computer vision methods to deep learning for improved license plate recognition accuracy. Lab Day Using Raspberry PI pin map and LCD pin layout as guide, we built the circuit. Challenges and Issues Faced SD Card Burning Failure One of the major obstacles was the failure to burn the operating system onto the SD card. Several attempts were made using different SD cards and flashing software such as Balena Etcher and Raspberry Pi Imager. The common issues encountered included: SD card not being recognized by the system. Flashing process getting stuck midway. Errors upon booting the Raspberry Pi after flashing. To address these issues, we tested multiple SD cards, reformatted them using different ...
阅读全文

Week 1

图片
Allocation of Roles and Analysis: During the initial week of our project, our team held a discussion to delegate specific roles and responsibilities to each member. After that, we defined the project requirements and outlined the software architecture. Lab Day: Today, our original task was to solve the hardware problem, but when we arrived at the lab, we found that we had no materials for our group, so we gave priority to the research and discussion of the software problem. The components we needed included: So we shifted our focus this week to software. We successfully implemented the core functionality of the recognition algorithm and completed the coding phase. However, during testing, we encountered some problems. In the first week of our project, we focused on developing a license plate recognition system using traditional computer vision techniques. Our initial approach relied on OpenCV, a widely used library for image processing. Instead of utilizing deep learning models, we opt...
阅读全文