A Raspberry Pi-Based Camera for Determining the Maximum Distance of Face Recognition Using HLBPH as the Algorithm

In this paper, the researchers focus on determining the maximum distance of the Raspberry Pi Camera using the Haar-Local Binary Pattern Histogram as its algorithm for face recognition. This research aims to create details on how face recognition would respond as the subject gets farther from the camera and if the algorithm can detect the face from a distance between the camera and the subject's face. The testing will be in a living room with 16.55 m2 of space. The first part will be testing the height of the camera's placement on a standard 5-shelf bookcase, starting from 0 ft on the ground and up until 6 ft which is the height of the mentioned bookcase. The second part is that at every difference in distance from the camera, there will be an increase of 0.5 ft until the algorithm cannot determine a face being detected. As the distance between the target and the camera increases, the reliability increases as the camera can see the face properly. However, as the distance gets farther than 4.5 ft, the reliability of the detectability of the face in front of the camera greatly decreases to 0%. Another test is then conducted to test the delay of the detection of the algorithm with its most reliable distance determined, which is 2.5 ft to 4.5 ft. based on this study, the researchers recommend that future researchers use a better-quality camera since the camera used in this study limits how the algorithm recognizes faces.

I. Introduction

The camera allows us to see what the naked eye cannot see. Additionally, the camera's sensor is one of the important parts since it is accountable for registering the incoming light and turning it into an image; this sensor can be found behind the lens in a camera. In this paper, the researchers focus on determining the maximum distance of the Raspberry Pi Camera using the Haar-Local Binary Pattern Histogram (HLBPH) as its algorithm for face recognition [1]–[3].