Passive Four-Dimensional Imaging and Recognition in Low Light Levels with Visible Range Image Sensors
Abstract
We proposed to investigate object classification with novel passive sensing based on 3D integral imaging in low light levels with CMOS cameras, and compare with LWIR cameras. The ability to detect, identify, and visualize targets and scenes passively in 3D in low light levels or with mobile devices will be very important for applications of interest to US NAVY, including soldier centric applications. We will demonstrate that there are substantial benefits in using 3D passive imaging in low light levels with CMOS cameras or with LWIR cameras for visualization, seeing through occlusion and obscuration (such as trees, foliage, etc.), and real-time 3D object identification and tracking. The availability of 3D information in low light levels can offer new tools for information exploitation for the US NAVY war fighter. Our 3D sensing method is based on multi-view 3D integral imaging to record parallax and to extract range and 3D profile of targets. 3D information obtained from passive integral imaging with a camera array is inferred based on the different positions of the cameras. This is a passive sensing acquisition technology, applied in cases where active sensing technologies are inadequate, and non-controlled illumination conditions/scenarios are necessary.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 06, 2023
- Accession Number
- AD1205224
Entities
People
- Bahram Javidi
Organizations
- University of Connecticut