Calibration Aware Learning Method And System For Compressive Imaging Using Focal Plane Array
Abstract:
The present invention provides a method and a system to acquire high resolution images using low resolution focal plane arrays. The system (100) consists of a spatial light modulator (103) that optically modulates the light from a scene (101) and projects the modulated light onto a focal plane array (105) which is then communicated to an Analog to Digital (ADC) (106) to convert the optical light signal to digital signal for reconstruction and output a high-resolution image using a reconstruction module (107). The system consists of a method to learn an optimal modulation pattern which takes into consideration the discrepancies in the spatial light modulator. The recovery method utilizes the knowledge of an optimal sparsifying domain, in which the signals passing through the spatial light modulator are sparse and recovers an image. The image thus recovered will be of higher resolution than the resolution of the detector.
Get Free WhatsApp Updates!
Notices, Deadlines & Correspondence
Outer Ring Road, Nagavara, Bangalore – 560045, Karnataka, India
Inventors
1. Gayatri Parthasarathy
Central Research Laboratory, Bharat Electronics Ltd., Jalahalli Post, Bangalore – 560013, Karnataka, India
2. Kishore Bachina
Central Research Laboratory, Bharat Electronics Ltd., Jalahalli Post, Bangalore – 560013, Karnataka, India
Specification
Claims:
A method for obtaining high resolution images, the method comprising:
dividing a Spatial Light Modulator (SLM) into plurality of blocks;
determining a calibration matrix for the Spatial Light Modulator (SLM);
receiving a signal from a scene to be captured and focusing the signal onto the spatial light modulator (SLM);
modifying the Spatial Light Modulator (SLM) according to a modulation pattern and modulating the received signal;
focusing the modulated signal from the Spatial Light Modulator (SLM) onto a Focal Plane Array detector having a plurality of detector elements;
digitizing the detected data using an Analog to Digital Converter;
recovering the signals of each block from the detector data using the modulation pattern and a domain in which the received signal is sparse; and
stitching the recovered blocks together for obtaining the high-resolution image.
The method as claimed in claim 1, modifying the Spatial Light Modulator further comprises:
determining training signals for each block;
generating a sparsifying basis using the training signals;
generating a binary measurement matrix using the training signals and the generated sparsifying basis; and
modifying Spatial Light Modulator (SLM) based on the generated binary measurement matrix.
The method as claimed in claim 2, wherein determining training signals comprises:
determining ? matrices corresponding to each detector element of the Focal Plane Array (FPA); and
dividing the training scene into plurality of blocks and then applying the ? matrices onto the blocks.
The method as claimed in claim 1 and 2, wherein recovering the signals of each block, comprises:
recovering a sparse signal from the modulated signal using the measurement matrix; and
recovering the signal from the sparse signal using a sparsifying transform and calibration matrix.
The method as claimed in claim 2, wherein generating the measurement matrix comprises:
initializing a set of training signals;
calculating sparse coefficients for the set of training signals;
identifying desired set of measurements of each signal in the set of training signals such that the entropy of measurements is maximized, and the measurements satisfy the Restricted isometry property;
determining a measurement matrix which is closer to the desired set of measurements; and
thresholding the values of measurement matrix to obtain a binary measurement matrix.
The method as claimed in claim 1, wherein determining a calibration matrix includes ensuring that only one element of the Spatial Light Modulator is made transparent to receive the signal from the scene.
The method as claimed in claim 1, wherein the Spatial Light Modulator (SLM) is divided into plurality of blocks based on the size of the Focal Plane Array, as per requirement.
8.An imaging system (100) for obtaining high resolution images, the imaging system comprising:
an optical set up (102) configured to focus light from a scene (101);
a spatial light modulator (103) configured to receive light from scene (101) and output a signal to a relay optical system (104);
a Focal Plane Array (105) configured to receive the signal from the relay optical system;
an analog to digital converter (106) configured to convert the signal from the Focal Point Array to a digital signal; and
a processing unit (107) configured to convert the received digital signal to a high- resolution image.
The imaging system (100) as claimed in claim 7, wherein the spatial light modulator (103) is one of:
a digital micro-mirror device (DMD);
one or more LCDs; and
array of independently-controllable mechanical shutters that include an array of apertures.
The imaging system for obtaining high resolution images as claimed in claim 8, wherein the processing unit is configured to:
divide a Spatial Light Modulator (SLM) into plurality of blocks;
determine a calibration matrix for the Spatial Light Modulator (SLM);
receive a signal from a scene to be captured and focus the signal onto the spatial light modulator;
modify the Spatial Light Modulator (SLM) according to a modulation pattern and modulating the received signal;
focus modulated signal from the Spatial Light Modulator (SLM) onto a Focal Plane Array detector having a plurality of detector elements;
digitize the detected data using an Analog to Digital Converter;
recover the signals of each block from the detector data using the modulation pattern and a domain in which the received signal is sparse; and
stitch the recovered blocks together for obtain the high-resolution image.
, Description:FORM – 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003
COMPLETE SPECIFICATION
(SEE SECTION 10, RULE 13)
CALIBRATION AWARE LEARNING METHOD AND SYSTEM FOR COMPRESSIVE IMAGING USING FOCAL PLANE ARRAY
BHARAT ELECTRONICS LIMITED
WITH ADDRESS: OUTER RING ROAD, NAGAVARA, BANGALORE 560045, KARNATAKA, INDIA
THE FOLLOWING SPECIFICATION PARTICULARLY DESCRIBES THE INVENTION AND THE MANNER IN WHICH IT IS TO BE PERFORMED.
TECHNICAL FIELD
The present invention relates generally to resolution of images. The invention, more particularly, relates to system and method for performing compressive imaging for providing high resolution images using low resolution sensor.
BACKGROUND
Compressive sensing or compressed sampling is a method to sample signals at an average rate much below the Nyquist rate and recover the signals with high probability from the reduced set of measurements. Compressed sensing relies on the idea that signals sampled at Nyquist rate contain a large number of redundancies. In other words, every signal sampled at Nyquist rate contains a sparse representation in some transform. The technique aims at capturing only the vital information from these sparse signals.
The problem of compressive sensing can be modeled as
(1)
where the matrix , with M<
Documents
Application Documents
#
Name
Date
1
202141005717-STATEMENT OF UNDERTAKING (FORM 3) [10-02-2021(online)].pdf
2021-02-10
2
202141005717-FORM 1 [10-02-2021(online)].pdf
2021-02-10
3
202141005717-FIGURE OF ABSTRACT [10-02-2021(online)].jpg
2021-02-10
4
202141005717-DRAWINGS [10-02-2021(online)].pdf
2021-02-10
5
202141005717-DECLARATION OF INVENTORSHIP (FORM 5) [10-02-2021(online)].pdf