💰 Budget: 5000$
Project Setup and Features
We have a multi-camera setup that is calibrated, and the relationships between cameras are known very accurately. Each camera has detected features in the images. There are a few dozen up to 100s of features per image. The camera views are overlapping, so that the observed features overlap between cameras. The attached image shows an example of the features detected in the camera image.
Data and Information Provided
Once the project starts, we will provide several sets of sample images from the cameras, as well as information about each feature, and its position in camera pixel space. We will also provide the relationships between the cameras. We can also provide the camera lens parameters, focal length, and principle points.
NDA Requirement
It will be necessary to sign a nondisclosure agreement prior to starting project work.
Project Objective
The project objective is to correspond the features between cameras with a high degree of certainty across multiple test samples. Although a fast computation is desirable, it is not a requirement. An important part of this project is to ensure that the features are correctly matched, so that the correspondence between cameras have as few mismatches as possible. It is possible that there will be some false-positive features detected, and it is also possible that although the camera images will overlap, they may cover slightly different fields of view.
Technical Requirements
The matching should generalize for 2 cameras up to N cameras.
Project Deliverables
The project success will show correct matching of the features. The source code for the matching algorithm will be delivered as part the contract. Any relevant documentation will be delivered. This is a Work for Hire, so that we will own the source code and the copyrights to it.
Skills Required
Skills: C++, Python, C#, Java, 3D Modeling, JavaScript
Categories
Categories: Data Science & Analytics, Data Analysis & Testing