Projects Sharing Researchers
- Uniform, Non-Disruptive, and Radiometrically Accurate Calibration of Infrared Focal Plane Arrays Using Global Scene Motion
- Method for Skin Cancer Detection Based on the Application of Statistical Decision Theory to Dynamic Infrared Image Sequences
- A Method and System for Feature Extraction and Decision Making from Series of Images
- Method for Sensitivity Optimization of Optical Receivers Using Avalanche Photodiodes Operating Under a Dynamic Reverse Bias
- High Performance Hyperspectral Detectors Using Photon Controlling Cavities
- Majeed Hayat
- Sanjay Krishna
- Sebastian Godoy
|Project Title||Method to Fuse Material Classification with Spatio-Spectral Edge Detection in Spectral Imagery|
A comprehensive edge-detection algorithm for spectral imagery by using real long-wave spectral imagery.
The technology is based on identifying joint spatial and spectral features through statistical analysis. The ASRC algorithm fuses thematic classification with spectral edge detection, thus the edge detection threshold process is adaptively changed depending on the spatial diversity of the materials within the imaged scene.
|Tags||spectral imagery, imaging|
|Posted Date||Sep 25, 2013|
The Adaptive Spectral Ratio Contrast (ASRC) algorithm is an extension of the Spectral Ratio Contrast (SRC) algorithm that fuses and utilizes the information from multispectral classification of the type of materials within the very spectral image whose edges are being identified. Edge detection is an important tool in analyzing and interpreting spectral imagery. According to IEEE, image segmentation is the single most important and difficult task in digital image processing. One of the key challenges, in detecting edges among multispectral and hyperspectral images, is that the edges can be defined by wavelength (color) changes rather than intensity changes. Several attempts have been made in defining this edge detector and recently a new approach was taken which uses the concept of spectral ratio signatures, which when fused with classification, achieved outstanding results for color-based edges.
Researchers at the University of New Mexico have developed a comprehensive edge-detection algorithm for spectral imagery by using real long-wave spectral imagery. The technology is based on identifying joint spatial and spectral features through statistical analysis. The ASRC algorithm fuses thematic classification with spectral edge detection, thus the edge detection threshold process is adaptively changed depending on the spatial diversity of the materials within the imaged scene.
- Reduces false edges due to unwanted changes in the intensity or noise
- Algorithms utilize spectral library information
- Dramatic reduction in the number of operations used with respect to other algorithms
- Integral in emerging spectral imaging sensors that are bias tunable
- ASRC has increased tolerance to noise
- Enhances the capabilities of the previously reported SRC algorithm
- Enhances the detection of edges that are solely due to changes in color, wavelength or type of material without exhibiting changes in the intensity
- Adaptively changes its edge detection capability depending on the spectral diversity and features of the materials
|9,430,842 Issued Patent||None||Download|
|Patent Number||Issue Date||Type||Country of Filing|
|9,430,842||Aug 30, 2016||Utility||United States|