Our client is a business that is part of the top 100 companies on Forbes with over 100 billion EUR in annual revenues. They are the world’s leading Consumer Electronics, Home Appliances, Home Entertainment, Mobile Communications, Solar, and Display Technologies.
The task was to create an object detection system suitable for automobile use. The application must detect obstacles and calculate the distances between obstacles with only one video source from a moving camera.
The target platform is an embedded mobile device that can see monocularly. This will be effective in real-time.
We designed a solution that uses sensor data for ground plane estimation.
- We aggregate information about obstacles by using a variety of methods, and various algorithms
- All results are combined into one final output.
- We solved various problems: Stabilization of input (vibration), image registration, and dynamic object recognition.
- Our solution comprises lots of optimization and parallelization work on GPUs SIMD/MIMD processors.
- It can work in a variety of conditions (day/night, it can work in sunny or cloudy weather)
We are adamant about Agile methods in all instances of our work. Most often, our clients and partners want us to implement Agile methodologies (Scrum, Kanban) from the beginning of the work. If our engineers make the decision, we examine the project’s specifics. Then we suggest a project management method based on Scrum or Kanban that best suits the project’s specific requirements and context.
The development team comprises five R&D developers working in our headquarters, one researcher, and a development team located at our client premises. We are in close contact with the Client Service Team and the Infrastructure team to provide technical support and maintenance when required.
Architecture & Technologies
This software application is integrated with an automotive hardware module.
- C/C++, Python, Java
- Open CL, Open CV
- Mobile and embedded platform optimization