Surveillance systems are image acquisition systems that use commercial security cameras fixed at specific locations in stores, houses, or warehouses to enhance safety and protect staff, assets, and buildings. Although these systems are good security options, a temporal secure area or nearby mobile needs an alternative personal monitoring solution that can be constructed by using the smartphone multimedia resources to access a PC's machine vision software, providing a tool for surveillance. In this chapter, an application is developed using only a smartphone camera and a pattern matching algorithm to detect presence of people in a secure area; when presence occurs, a signal is sent to start a recording and/or activate an alarm to warn the users. A virtual instrument algorithm application was able to detect objects in a room, recognizing who are residents and who are guests. The functionality of this project with encouraging results obtained provides a flexible security instrument.
Global Perspectives on Robotics and Autonomous Systems
TopIntroduction
Vision-based action recognition is one of the most challenging research and the most interesting topics of machine vision and pattern recognition. In (Aktı et al., 2019) an application for detecting fights were developed, this an important issue to be quickly under control before it escalates. However Convolutional Neural Networks (CNN) were used, this algorithm has a long learning curve to average users and in most of the cases lacks for a real-time application.
In (Thys & Ranst, 2019) CNN´s weakness is discovered, an adversary can fool a person detection system by hanging a small picture on a person, and also have some issues with elderly people.
Machine vision provides industrial teams or home applications with the ability to perceive actions as well as movements and make immediate decisions based on that visual input. The primary applications of machine vision are visual inspection and defect detection, part placement and measurement, as well as product classification and tracking. Machine vision constitutes one of the fundamental technologies in industrial automation. Machine Vision has helped improve product quality, increase production speed, and systematize manufacturing procedures and logistics for decades (López et al., 2008). And now, this proven technology is merging with artificial intelligence to lead the transition to Industry 4.0 (Intel, 2023). In this sense, machine vision principles, techniques and technologies have gained attention in recent years. And as a consequence, for achieving the purposes of machine vision applications, both, physical devices and software-dependent solutions need to be developed.
Image acquisition generally involves the capture of an image signal (it can be analog or digital) and the generation of a one-dimensional (1-D) or two-dimensional (2-D) array of integers representing pixel brightness levels (Whelan & Molloy, 2001), (Blasinski et al., 2018), (Farrell et al., 2012).
To be able to use Machine Vision tools an image acquisition system needs to be implemented, and this is a system that focuses on retrieving an image from a source, usually hardware systems like cameras and sensors, that we can use for security reasons. For example, they can be placed at important or valuable locations such as stores, houses, warehouses, etc. So the main purpose of this system is to keep these places or things under surveillance and this way to avoid theft (Blasinski et al., 2018). Although these systems are good security options, they can be improved by using artificial vision software, providing more and better tools for surveillance.
Image processing involves the use of algorithms to manipulate digital images. It is an interdisciplinary field that combines computer science, mathematics, and engineering. The field of image processing has evolved significantly over the years, thanks to advances in technology, mainly in computer hardware and software. These advances have led to the development of more sophisticated image processing techniques, such as machine learning and deep learning. One of the primary advantages of image processing is that it can be applied to various areas. For example, image processing is widely used in the field of medicine to analyze medical images. In the field of engineering, image processing can be used to analyze the structural integrity of buildings and other structures. One of the most popular applications of image processing is facial expression recognition. This is used in many fields, such as security, safe driving, and medical rehabilitation. However, image processing still has a lot of challenges. One of the most significant challenges is dealing with the variability in images caused by different lighting conditions, camera positioning, and other factors. These variations can significantly impact the results of image processing techniques. Image processing has improved data acquisition, it also has brought more efficient data analysis, and the ability to automate various tasks (Sajjad et al., 2023).