Making Dashcams Smarter for Road Safety
Using smartphones to enhance dashcam safety features and real-time analysis.
Seyul Lee, Jayden King, Young Choon Lee, Hyuck Han, Sooyong Kang
― 8 min read
Table of Contents
- Why Are Dashcams Not Used to Their Full Potential?
- The Big Challenge of Video Analysis
- Keeping It Close to Home: What Is Edge Computing?
- Why Use Smartphones and Tablets?
- The Technical Challenges
- 1. Heavy Workload
- 2. Device Connectivity
- 3. Different Device Capabilities
- 4. Different Video Streams
- The Solution: A Smart video Analysis System
- Key Features of Our System
- Putting It to the Test
- How Did It Perform?
- The Benefits of This System
- Lessons Learned
- Looking Ahead
- Conclusion
- Original Source
- Reference Links
Dashcams are the cameras that sit on your car's dashboard or windscreen and record everything happening in front of you. They are mainly used to collect evidence if you get into a car accident. Sounds great, right? But here’s the catch: most of the video recorded isn’t actually about accidents. Instead, it just sits there, waiting for someone to hit “delete.”
But what if we could use dashcam video for something more than just plugging the gaps in your memory after a wild drive? What if we could use it to keep you safer on the road?
Why Are Dashcams Not Used to Their Full Potential?
Most dashcams capture hours of video during your drives. But because most of it doesn’t involve accidents, it gets tossed away like last week's salad. So, we lose some potentially useful data. This is a common problem: how to turn this data into something beneficial, especially regarding safety.
This is where video analysis comes in. By analyzing dashcam footage, we can identify potential dangers on the road, like pedestrians, other cars, or even potholes. But there’s a catch: analyzing video in real time is like asking your grandma to race in the Olympics—she just doesn’t have the resources (sorry Grandma!).
The Big Challenge of Video Analysis
To analyze video in Real-time, we need a lot of computing power. The problem is, most dashcams don't have that kind of muscle. You might as well be trying to lift weights with a toothpick.
One solution many people think about is sending the video to the cloud for processing. But, this comes with its own issues. Imagine trying to send your home videos to the cloud while your neighbor is streaming his favorite show. The internet could slow down, and you’d be waiting longer for your dashcam to process a video than it takes to finish a season of your favorite TV show.
Keeping It Close to Home: What Is Edge Computing?
Here’s where edge computing enters the picture. Instead of sending the video footage off to the cloud, we can process it right here at the edge—think of it as using a home printer instead of going to a print shop.
In this case, the “printers” are the Smartphones and tablets that people often carry around. These devices are on hand in almost every car, meaning we can tap into their computing power. That way, we can analyze video data while avoiding the cloud waiting game.
Why Use Smartphones and Tablets?
Smartphones are everywhere, and they usually have more computing power than a regular dashcam. Plus, they’re charged and ready to go with built-in cameras, making them a perfect match for the dashcam gig.
With a smartphone, not only do we have the computer power, but we can also use the phone’s camera to act as an additional dashcam. It’s like having a backup quarterback—always ready if the starter is having a tough day.
The Technical Challenges
Even with all that power, we still run into some bumps in the road.
1. Heavy Workload
Real-time video analysis can be very demanding, requiring quick processing of a lot of video data. If you try to squeeze all that work onto just one device, it’s like trying to shove a fully stocked refrigerator into a tiny cupboard. It just won’t fit.
2. Device Connectivity
With multiple smartphones in play, we can encounter connectivity issues. If someone decides to turn their phone off or the battery dies, it can disrupt everything. It’s like a game of musical chairs, but with devices that might just refuse to cooperate.
3. Different Device Capabilities
Not all smartphones are created equal. Some are like speedy race cars, while others are more like comforting old sedans. Each phone has varying power levels, which makes scheduling the work a little trickier.
4. Different Video Streams
If you're using multiple cameras, the video streams can require different kinds of analysis. It’s like trying to juggle while also coordinating a dance—challenging, to say the least.
System
The Solution: A Smart video AnalysisWe are introducing a distributed system that can analyze dashcam videos in real time using these smartphones. The system breaks down the workload into manageable tasks, dividing them between all the devices in the car. It’s like organizing a potluck dinner—everyone brings a dish, but we make sure no one is trying to bring everything themselves.
Key Features of Our System
-
Pipelines for Efficiency
The system works by breaking down the video analysis process into steps that can be done simultaneously. This is called pipelining. If one task is busy, another can take over, keeping everything moving smoothly. It’s like having a well-organized kitchen—lots of chefs working together without bumping into each other.
-
Smart Frame Scheduling
The system uses a scheduling method that makes sure video frames are assigned to devices based on their available capacity. So rather than just throwing all the work at one device, we look at what each device can handle. Think of it like assigning tasks based on ability—every chef takes the job they're best at!
-
Dynamic Frame Rate Control
The system constantly checks the device’s resources and adjusts the frame rate as needed. If a device is getting overwhelmed, the system can lower the frame rate to prevent a backlog. If it has extra capacity, it can increase the frame rate. It’s like balancing your workload at a party—if you’re having too much fun on one game, you might want to slow down, or if the guests are eager for more, let them play!
Putting It to the Test
We designed and tested this system using smartphones and a dashcam emulation app. This app mimics the functionality of a regular dashcam, allowing us to test how our system works in different scenarios without needing every car to have actual dashcams.
How Did It Perform?
In our tests, the system showed that it could process videos from two different sources while maintaining low latency. This means that the alerts for potential dangers can be provided almost immediately—like instant notifications on your smartphone!
We also tested various environments, from stable situations to those where devices frequently joined or left the system. The analysis worked efficiently, even when the strength of the devices varied.
The Benefits of This System
-
Extra Value from Unused Video Data
Instead of tossing the non-accident footage away, we can use it to improve safety.
-
Low-Latency Video Analytics
The system allows for real-time analysis, meaning we can help drivers react to potential hazards quickly.
-
A Practical Mobile Solution
The entire system functions through mobile applications, making it accessible for anyone with a smartphone.
Lessons Learned
-
Individual Device Power vs. Number of Devices
Our tests showed that the power of individual devices significantly impacts speed. While the number of devices can help, having a strong primary device is essential for better performance.
-
Device Connectivity Matters
Maintaining a solid connection among devices is crucial. Without it, performance may lag, just like a bad Wi-Fi signal can frustrate your streaming.
-
Network Bandwidth Considerations
The system can consume significant bandwidth when transferring video, so it uses a strong local network to avoid issues.
Looking Ahead
We plan to make more improvements to the system. These include:
- Choosing Analysis Models by Device Temperature
The hotter a device gets, the slower it can process data. We can develop a system that selects less demanding analysis models if devices are running hot.
- Dropping Overdue Frames
Sometimes frames take too long to analyze and could miss their window of usefulness. We could create a feature to discard these outdated frames to keep the process swift and efficient.
Conclusion
In conclusion, transforming dashcams into smart safety tools is not only possible—it’s already happening! By using smartphones and tablets as part of our proposed system, we can take real-time video analysis to the next level. This means not only can we prevent accidents before they happen, but we can also use previously discarded footage for better safety on the road.
And who knows? The next time you step into a car, you may just find that little camera working hard behind the scenes to keep you safe without you even knowing it. Now, that’s some clever road safety!
Title: In-Vehicle Edge System for Real-Time Dashcam Video Analysis
Abstract: Modern vehicles equip dashcams that primarily collect visual evidence for traffic accidents. However, most of the video data collected by dashcams that is not related to traffic accidents is discarded without any use. In this paper, we present a use case for dashcam videos that aims to improve driving safety. By analyzing the real-time videos captured by dashcams, we can detect driving hazards and driver distractedness to alert the driver immediately. To that end, we design and implement a Distributed Edge-based dashcam Video Analytics system (DEVA), that analyzes dashcam videos using personal edge (mobile) devices in a vehicle. DEVA consolidates available in-vehicle edge devices to maintain the resource pool, distributes video frames for analysis to devices considering resource availability in each device, and dynamically adjusts frame rates of dashcams to control the overall workloads. The entire video analytics task is divided into multiple independent phases and executed in a pipelined manner to improve the overall frame processing throughput. We implement DEVA in an Android app and also develop a dashcam emulation app to be used in vehicles that are not equipped with dashcams. Experimental results using the apps and commercial smartphones show that DEVA can process real-time videos from two dashcams with frame rates of around 22~30 FPS per camera within 200 ms of latency, using three high-end devices.
Authors: Seyul Lee, Jayden King, Young Choon Lee, Hyuck Han, Sooyong Kang
Last Update: 2024-11-29 00:00:00
Language: English
Source URL: https://arxiv.org/abs/2411.19558
Source PDF: https://arxiv.org/pdf/2411.19558
Licence: https://creativecommons.org/licenses/by/4.0/
Changes: This summary was created with assistance from AI and may have inaccuracies. For accurate information, please refer to the original source documents linked here.
Thank you to arxiv for use of its open access interoperability.