ObjectDetectingWithDrone
- 0 Collaborators
Object Detecting with drone using Intel's CNN algorithm & Implement Android Application ...learn more
Project status: Concept
Mobile, Artificial Intelligence
Intel Technologies
Other
Overview / Usage
It was held at 'AI Object Detecting Hackathon', an engineering festival held in Ilsan, Korea. The subject was to detect the object using the CNN algorithm provided by Intel and apply it to the actual case through its output. After we made the cage in accordance with the scenarios we actually made, the videos taken by the drones were used as inputs.
As automobiles become widespread, traffic accidents are very much associated with cars. Traffic accidents are increasing every year, and many animals such as cats, dogs, and other people die frequently as well as people. So the scenario that we made is as follows. From the driver 's point of view, when the driver pulls the car, if there is a cat or dog on the street, it will output a warning message and a voice. Also, from the standpoint of pedestrians, when the pedestrian is going to cross the road, if there is a car or a bus, a warning message and a voice are outputted. As a result, the driver will be able to reduce the number of traffic accidents for small animals and small children due to the small size of the vehicle.
In the above scenario, we shot the video, detected the object using the CNN algorithm created by Intel, sent the result to the server made in Java, and the server actually implemented the method by sending the result to the Android app.
Methodology / Approach
The method used for Object Detecting was provided by Intel. Since the server was created in java and we created a demo version of the application, we implemented the server through a simple socket network. In addition, Application uses Android Studio and does not implement many functions. When data such as 'cat, person, dog' etc. comes in, application is implemented to execute by dividing into driver's position and pedestrian's position.
Technologies Used
Intel provided the CNN algorithm and NUC to use this algorithm. There are no external libraries used to build the server, all of which are basically libraries. I also used the Glide library to insert gifs in Android Studio, but all the rest used only basic libraries.