Safety hemant detection by using yolo m.pptx

Project

1. www.cambridge.edu.in Department of Computer Science & Engineering Phase-2 Project Presentation On “Safety Helmet Detection Model Based On Improved YOLO M” By Gangadhara M N Nagendra K P Naresh Prajwal B 1CD20CS050 1CD20CS099 1CD20CS103 1CD20CS119 Under the Guidance of Ms.Maria Kiran L Assistant Professor

2. Contents Introduction Literature Survey Proposed Method Objectives Motivation Challenges Applications Hardware and Software Requirements Architecture Flow Diagram/Methodology Works to be completed. References

3. Introduction We'll use an improved YOLO-M (You Only Look Once for Multi-Object Detection) architecture. To do this, we'll gather a dataset of images showing people with and without helmets, preprocess the data, modify YOLO-M for helmet detection, and train the model. After training, we'll evaluate its performance and deploy it for real-time or static image helmet-wearing detection. “The goal is to enhance safety monitoring in various environments”. Creating a safety helmet detection system with an improved YOLO-M model means training a computer to recognize if people are wearing safety helmets in pictures or videos.

4. Literature Survey Sl. No Paper Name Author Name Year Advantages Disadvantages 1 A Real-Time Safety Helmet Wearing Detection Approach Based on CSYOLOv3 Haikuan Wang Zhaoyan Hu Yuanjun Guo Feixiang Zhou 2020 Backbone network of darknet53 is improved by applying the cross stage partial network ( CSPNet ), which reduces the calculation cost and improves the training speed. Challenging to detect safety helmet wearing (SHW) in real time on the premise of ensuring low precision performance. 2 A Novel Implementation of an AI-Based Smart Construction Safety Inspection Protocol in the UAE Mohammad Z. Shanti, Chung- suk Cho, Young-ji Byon, Song- kyoo Kim 2021 Convolutional Neural Network (CNN) model in order to detect two main components of the PFAS that are safety harness and life-line in addition to a standard safety measure of using a safety helmet. 1. Despite having some errors and mis detections the model can be considered reliable.

5. 3 Research on application of helmet wearing detection improved by YOLOv4 algorithm. 1.Haoyang Yu, 2.Ye Tao, 3.Wenhua Cui, 4.Tianwei Shi. 2023 The model was tested improves the detection accuracy of small targets by adding multi-scale prediction and improving the structure of PANet network. Too many parameters and the detection effect of small targets is poor. 4 Multiple Complex Weather Tolerant and Low Cost Solution for Helmet Detection Chang Hao Xu Yong Huang Shuqin 2021 K-means++ clustering algorithm is utilized to change the dimension of anchor box for better detection performance, and MSRCR algorithm is used to filter complex weather conditions. Reduces the rates of missing detection and misdetection of small target detection in original network. Literature Survey (cont.)

6. 5 SHEL5K: An Extended Dataset and Benchmarking for Safety Helmet Detection. Munkh -Erdene Otgonbold Munkhjargal Gochoo Fady Alnajjar Luqman Ali 2022 The proposed dataset was tested on multiple state-of-the-art object detection models, i.e., YOLOv3, YOLOv4,YOLOv5. It contains only fewer images with better labels and Too many parameters, substantial detection interferences Literature Survey (cont.)

7. Proposed Method Dataset Collection Data Preprocessing Model Selection Evaluation Metrics Iterative Improvement

8. Objectives Real-time Detection Enhanced Accuracy Minimized False Positives/Negatives Adaptability to Varied Helmets

9. Motivation Accident prevention Workplace safety Enhancement Cost reduction Employee welfare

10. Challenges Real Time processing demands Privacy concerns Adaptation to new helmets False positive/negative Trade-off

11. Applications Enforce safety protocols on construction Sites Manufacturing Plants Mining Operations Oil and Gas industry Automotive manufacturing Construction Vehical Operation

12. Hardware- Software Requirements Hardware Requirements: System: Intel core i5 Minimum and 2GHz Minimum RAM: 8GB Hard Disk: 10GB or above Input Device: Keyboard and Mouse Output Device: Monitor Software Requirements: Language: python Software: VS code

13. Architecture Input(640,640,3) Focus(320,320,12) Conv2D_BN_SiLU(320,320,64) Conv2D_BN_SiLU(160,160,128) CspLayer (160,160,128) Conv2D_BN_SiLU(80,80,256) CspLayer (80,80,256) Conv2D_BN_SiLU(40,40,512) CspLayer (40,40,512) Conv2D_BN_SiLU(20,20,1024) SPPBottleneck (20,20,1024) CspLayer (20,20,1024) Yolohead Yolohead Yolohead Concat+CspLayer Concat+CspLayer Concat+CspLayer Concat+CspLayer Upsampling Upsampling Downsampling Downsampling Conv2D Conv2D

14. Architecture (Cont..) High-level architecture for single-stage object detectors : Single-stage object detectors (like YOLO ) architecture are composed of three components: Backbone, Neck and a Head to make dense predictions. Model Backbone: The backbone is a pre-trained network used to extract rich feature representation for images. This helps reducing the spatial resolution of the image and increasing its feature (channel) resolution.

15. Architecture (Cont..) Model Neck : The model neck is used to extract feature pyramids. This helps the model to generalize well the objects on different sizes and scales. Model Head: The model head is used to perform the final stage operations. It applies anchor boxes on feature maps and render the final output: classes , objectness scores and bounding boxes.

16. Flow Diagram/ Methodology

17. Flow Diagram/ Methodology (Cont..) A flow chart of helmet wearing monitoring system. It is mainly divided into two modules: helmet wearing identification module and personnel tracking module. The detection module’s function is to train the detection module with a large amount of data, detect real-time surveillance video captured by the camera, analyze the situation of employees wearing helmets, and obtain target information for employees in the video for use by the follow-up tracking module. The personnel tracking module sets up the tracking set using the detection target box sent by the front detection module and then uses the target tracking algorithm to track the target set.

18. Implementation Front End: User Interface Video Upload Display Results Test and debug the frontend Back End: YOLO Model Dataset Training Test and debug the backend

19. Results

20. Results( Cont …)

21. Results( Cont …)

22. Results( Cont …)

23. Conclusion The safety helmet detection project leveraging YOLOv5 with MobileNetv3 as the backbone feature demonstrates a robust solution for ensuring safety compliance in various environments. By harnessing the power of deep learning and computer vision techniques, the system can accurately detect the presence or absence of safety helmets in real-time scenarios.

24. References LILI WANG , XINJIE ZHANG, AND HAILU YANG ” Safety Helmet Wearing Detection Model Based on Improved YOLO-M” 2020 Desu Fu1, Lin Gao1*, Tao Hu1, Shukun Wang1, Wei Liu ” Research on Safety Helmet Detection Algorithm of Power Workers Based on Improved YOLOv5”2021 KUN HAN AND XIANGDONG ZENG” Deep Learning-Based Workers Safety HelmetWearing Detection on Construction SitesUsing Multi-Scale Features”2021 NING LI 1,2, XIN LYU 1, SHOUKUN XU 2, YARU WANG2,YUSHENG WANG 2, AND YUWAN GU2”Incorporate Online Hard Example Miningand Multi-Part Combination Into AutomaticSafety Helmet Wearing Detection”2020

25. References AYYOUB BOUHAYANE 1,2, ZAKARIA CHAROUH 3MOUNIR GHOGHO 2,4 AND ZOUHAIR GUENNOUN 1” A Swin Transformer-Based Approach forMotorcycle Helmet Detection”2023 MOHAMMAD Z. SHANTI 1, CHUNG-SUK CHO 1, YOUNG-JI BYON 1, CHAN YEOB YEUN 2 TAE-YEON KIM 1,SONG-KYOO KIM 2,3 AND AHMED ALTUNAIJI 2”A Novel Implementation of an AI-Based SmartConstruction Safety Inspection Protocolin the UAE”2021

26. Questions?

27. Thank you!