Snatching Detection System

Project Title

Snatching Detection System

Project Area of Specialization Artificial IntelligenceProject Summary

The increase of video surveillance in public spaces and the proliferation of body cameras for police can potentially be leveraged for gun detection systems. Video systems could alert police and surveillance personnel when a gun is detected in real-time, resulting in prompter action. In this project, we aim to develop a smart surveillance security system detecting weapons specifically guns. For this purpose, we have applied a few compute vision methods and deep learning for the identification of a weapon from the captured image. Recent work in the field of machine learning and deep learning particularly convolutional neural networks has shown considerable progress in the areas of object detection and recognition, exclusively in images. As the first step for any video surveillance application, object detection and classification are essential for further object tracking tasks. For this purpose, we trained the classifier model of YOLO v3, Yolo v5, i.e., “You Only Look Once”. This model is a state-of-the-art real-time object detection classifier. Furthermore, we are not just detecting the guns but also getting the frame of the incident and storing the data for future use.

Project Objectives

This project aims to design a system that is capable to detect weapons automatically. Thus, in order to achieve the goal, the project involves the following objectives and scope:

MAIN OBJECTIVE:

• We tackle realistic videos which include CCTV videos and movie videos. In addition, the waist shot of the video will be used as the input of our proposed algorithm as if the human and object are hard to be seen in the video, the system may provide a less accurate result

SUB-OBJECTIVE:

• We tackle realistic videos which include CCTV videos and movie videos. In addition, the waist shot of the video will be used as the input of our proposed algorithm as if the human and object are hard to be seen in the video, the system may provide a less accurate result

• We tackle realistic videos which include CCTV videos and movie videos. In addition, the waist shot of the video will be used as the input of our proposed algorithm as if the human and object are hard to be seen in the video, the system may provide a less accurate result

Project Implementation Method

In this work, we have attempted to develop an integrated framework for reconnaissance security that distinguishes the weapons progressively, if identification is positively true it will caution/brief the security personnel to handle the circumstance by arriving at the place of the incident through IP cameras. We propose a model that provides a visionary sense to a machine to identify the unsafe weapon and can also alert the human administrator when a gun or firearm is obvious in the edge.

The most important and crucial part of any application is to have a desired and suitable dataset in order to train the machine learning models. Therefore, we manually collected a huge amount of images from Google. One must save images in “.jpg” form; if the images are in different extensions, it will be a little troublesome and will generate errors when provided for training. Alternatively, since the images are processed in terms of batches, therefore prior to training, the sizes of all the images are transformed into the same width and height 416?×?416 pixels.

Benefits of the Project

Security is always a main concern in every domain, due to a rise in the crime rate in crowded events or suspicious lonely areas. Abnormal detection and monitoring have major applications of computer vision to tackle various problems. Due to the growing demand in the protection of safety, security, and personal properties, the need and deployment of video surveillance systems that can recognize and interpret the scene and anomaly events play a vital role in intelligence monitoring. This project implements automatic gun (or) weapon detection using a convolution neural network (CNN) based yolo algorithms.

Technical Details of Final Deliverable

1. Functional Requirements

• Testing of Weapon Videos on System.

• Collection of images for training

• Saving detected frames.

 

1. Non-Functional Requirements

1)Usability:

how easily a user can achieve their goal in a single page visit;

how quickly they perform the tasks in the store;

how memorable and intuitive the design is;

number and time of errors users make.

2)Security:

Only the system data administrator can assign roles and change access permissions to the system.

3)Performance:

The website’s homepage should load in less than 4 seconds on chrome amd Edge.

Final Deliverable of the Project Software SystemCore Industry ITOther Industries Security Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Quality EducationRequired Resources