Roadkill is a kind of accidents that an animal is killed by motor vehicles on the highway and it has become a worldwide problem. Tasmania in Australia has one of the highest roadkill rates, and the roadkill is the second biggest threat to the survival of the Tasmanian devil after Devil Facial Tumour Disease (DFTD). The government encourages the public to report the roadkill so they can understand the high-risk area of roadkill and take actions such as installing virtual fences. The government has developed a free mobile App for roadkill reporting however normally it is hard and unsafe to stop and record the roadkill in a highway.

Problem Statements

Intentionally killing animals on the road has become a worldwide issue. Roadkill is a serious problem in Tasmania. There are 500,000 native animals are killed on the state’s roads each year. In every 2 minutes, an animal is killed by a car somewhere on Tasmanian roads which giving the state one of the highest roadkill rates in the world. It is not only harmful to wildlife animals but also a threat to drivers. There is 8.8 million dollars insurance claim for a car accident caused by animals according to an insurance company in South Australia. Also, hitting large size animals such as kangaroos will cause fatal consequences. More importantly, no one wants to hurt these innocent animals, it feels really bad.

Although the Government has a reporting application for roadkill, the problems are that it requires drivers to stop somewhere on the road to take the picture and write lots of information to upload which is difficult in reality.

What it does

To solve this problem, our idea is to design an intelligent Wildlife Road Detection System which can automatically detect and report any roadkill on the road. Which means, with our system and device, everyone can contribute to data collection in a more efficient and safer way, and therefore to reduce the roadkill accidents.

This system includes a dashcam and an associate App. When a driver is driving, the dashcam automatically recognizes wildlife animals by using object detection technology, takes photos of them and send them to the App installed on the user's mobile phone. Then the photo will be uploaded and analyzed on our cloud platform. These analyzed data will be used to create heatmaps of roadkill displayed on the App where user can check where are the high accidental areas for prevention. Also, our App has the function that can warn the driver with voice alerts when they are getting close to the high accidental area (such as 200 meters away and reminding drivers to slow down speed) at the time when animals are most active such as during dawn and dusk.

How I built it

Our group design the wildlife road detection system including a dashcam and an App. After the discussion among group members, we use the dashcam that can detect wildlife animals and capture their photos. During the detection of animals, the dashcam will use object recognition technology to recognize animals and store related data such as species, date, location, time on the picture. Then, these pictures will be transmitted to the App through Bluetooth or wireless Internet and stored in the database. Taking privacy into consideration, the pictures will be edited with solid format such as no display of human face in photos, and drivers’ traceability will be encrypted and stored in the database.

According to our design, the App will display the data in the map with different color dots representing different wildlife animal species. The users can use the filter to choose which kinds of wildlife animal presented in the map. The App also includes the voice prompt to warn drivers the high accidental area in advance, if users are familiar with road conditions, they can switch off the voice prompt. Moreover, the App has two picture uploading modes such as annual mode and automatic mode. All uploaded picture information can be seen in the record interface of App, user can annually upload pictures and delete uploaded pictures. The data will be analyzed in the App and presented with a pie graph in different periods such as a week, month and year in the fact interface.

How can we make a Profit

In the first stage, the camera could be sold with the app as a bundle. Insurance Company can promote our bundle to customers to reduce the claims related to roadkill. Once we have enough data, drivers who don’t want to buy the camera can subscribe our App to get a warning and improve their driving safety. At last, governments and wildlife rescue organization can use our data to install virtual fences, clean the roads and do other research projects.

Challenges I ran into

This is the first time for all of the team members to Uhack, so we do not have experience in planning Business Canvas, especially in how to make money. We have discussed for several rounds and think several modes of making profits, but things seem not to be so easy. However, thanks to the help of the mentors, we can focus attention on the possible direction to work out this problem (After getting to research more, we found that the insurance companies in Australia have demand for our product), and also thanks for the workshops and the speakers we quickly got the skill of writing Business Canvas. Besides, another issue challenge us is privacy. When the dash cam takes photos, it inevitably captures some personal privacy which should be not open to the public. To solve the privacy issue, we came up with ideas such as cropping the objects (animals) only or mosaic human faces and vehicle registration plates.

Accomplishments that I'm proud of

Our teams have a strong will to win the Uhack game, so we put enough energy and time to the competition. During these days, we finished our Business Canvas after several modifications and update to a real business level which is realistic and profitable. We successfully deliver a pitch presentation and produce excellent pitch video which uses professional software. Besides, we proudly finish our product’s prototype which introduces the main functions, business model, and future plan with vivid animations.

What I learned

Through Uhack, we worked as a real business team and separate tasks into small parts and every team member has a clear responsibility. During the competition, we gather together to brainstorm ideas, pitch video, design the system and improve, and we have built a strong partnership with all the team members and realized the importance of working together and the trust in each other. This is the basis for us to highly engaged in Uhack and work toward positive outcomes. Besides, from Uhack we practice what we have learned in university and put theoretical knowledge into practice.

What's next for Wildlife Road Detection - uDetect uShare uSave

In the next stage, we first plan to improve the basic function such as the display of the map and the organization of the data we collected, and then we may put more emphasis on the design details such as how to deal with those pictures captured at the same venues. In the future, the system aims to analyze the distribution of certain animal for advanced prevention of accident and also can be extended to detect other road hazards such as bushfire, road defects and damage to public property. Moreover, this technology can be used on cameras for motorbike and bicycle, or even be integrated into the vehicles. We also plan to integrate other application or service such as weather conditions into the analysis so that we can have a multi-face and more accurate analysis. Finally, we hope this application can be promoted in Australia nationwide even around the world.

We also plan to development reward mechanism which can give drivers more incentives to upload photos, and certificate will be rewarded to people based on the numbers of photo they upload, the more picture an account uploads, the more points and rewards they would get. The purpose of this strategy is to attract more people to use this app and allow app to upload more pictures (even users choose upload manually) to facilitate us to collect image data.

uDetect, uShare, uSave. Together, we can create a better driving environment and save a lot of lives.

Built With

Share this project: