AutoTurret - Arduino controlled Laser Turret with Computer Vision

Project Overview

The security of individuals and communities has become increasingly important in recent times. This project aimed to design and build an automated laser turret to enhance security measures in various settings, using computer vision and an Arduino microcontroller. The laser turret is capable of detecting human poses and aiming at potential threats, while also incorporating a “surrender” mode to prevent unnecessary harm. The project was designed with the belief that security is of paramount importance in modern society, and that technological advancements can help to improve safety measures. The laser turret was designed using servos and lasers, and the computer vision aspect was achieved through the implementation of OpenCV software. The project was tested and refined to improve accuracy and speed, resulting in a successful final product. The survey conducted as part of this project aimed to gather information on the most important features of a laser turret, its potential uses, and the public perception of such a technology. The survey results indicated that accuracy, cost, and range were the most important factors, while the legality and privacy concerns of the technology were also deemed significant. Overall, this project presents a viable solution for security purposes, incorporating advanced technology while also considering societal and cultural factors. The laser turret has the potential to enhance security measures in various settings, including election centers, public areas, and other locations where security is a concern.

Features

The automated turret system is designed to detect and track human targets using computer vision and laser diode technology. The system operates using the following step-by-step working principle:

• Human detection: The system uses computer vision algorithms to detect and locate human targets within its field of view. This is achieved by analyzing the video stream from a camera mounted on the turret system.
• Human pose analysis: The system performs human pose analysis to determine whether the detected human is in a surrender pose or not. This is done to avoid unnecessary use of force and ensure the safety of non-threatening individuals.
• Target tracking: Once a human target is detected and identified as a potential threat, the system tracks the target’s movements using servo motors to adjust the position of the camera and laser diode.
• Laser aiming: The laser diode is used to aim at the detected human target. This is done by adjusting the position of the laser using the servo motors to ensure that the laser is pointing directly at the target.
• Warning countdown: The system issues a warning countdown to the detected human target, giving them an opportunity to surrender and avoid the use of force.
• Shooting: If the detected human target does not comply with the warning countdown, the system can be programmed to automatically shoot the laser diode at the target.

The flowchart below describes the overall process of the auto-targeting turret. The pre-defined processes in the flowchart are discussed in detail in the upcoming sections. In general, the automated turret system uses computer vision and laser diode technologies to detect, track, and possibly eliminate human threats. The use of servo motors to modify the position of the camera and laser diode in the system allows for precision aiming, while the warning countdown function allows the system to avoid the application of superfluous force

Auto-Aiming

The turret receives video feed from a stationary camera. Each frame of the video feed are analyzed to detect intruders and move the servo accordingly so that the lasers can be aimed at the target. This process is visualized in Figure 4. But a single frame is in pixel values. To calculate the servo degrees from the pixel values, we first need to determine the range of the servo motor or the Field-of-View (FOV). The field of view of the camera is the area that the camera can see. This is determined by the camera lens and sensor size. Once we know the field of view, we can calculate the number of pixels per degree of rotation. This value will be unique for each camera and lens combination. In this experiment, we figured it out by manually moving the servos in a controlled manner.

Then, when the camera detects a human target, it calculates the position of the target within the image in terms of the pixel values. These pixel values are then converted into degrees of rotation for the servo motor using the previously calculated pixels per degree value. From the figure above, we get the following equations.

Since, “d” is common, we can get the value of β by the following equations:

This β is the number of degrees the servos need to move to aim at the target. This calculated degree is fed to the servo motors so that it can accurately point the lasers at the target. In similar fashion, vertical movement is also calculated. These two calculations are done for every frame, giving the turret the ability to automatically aim at any intruder visible in the camera feed.

Measuring Distance & Firing Effect

To measure the distance to a human target, we use a method called shoulder width calculation. This involves measuring the width of the target’s shoulders in pixels within the camera image, and using this measurement to estimate their distance from the turret. To begin, we use the landmark detection method described earlier to identify the target’s shoulders within the camera image. Once we have identified the shoulder landmarks, we calculate the distance between them in pixels. Next, we use the known width of an average human shoulder to estimate the actual distance to the target. This is based on the assumption that the target’s shoulder width will be proportional to their actual distance from the turret. By continuously measuring the target’s shoulder width as they move within the camera’s field of view, we can track their distance and adjust the aim of the turret accordingly. This allows the system to accurately target human subjects at a distance and respond appropriately to their movements. Although the turret does not actually fire bullets, we can simulate the shooting of a gun by alternating the laser light and playing sound effects.

Final Product

Cost

Conclusion

I had a lot of fun working on this project. Partly because I got to work with something tangible, rather than some pixel on a screen. I also learned a lot about computer vision and how to use it in real life applications.