<title>Focal Plane Spatial Response And Modulation Transfer Function (MTF) Measurements Using A Computer-Aided Flying Spot Scanner</title>

Abstract

AbstractIn the measurement and characterization of the performance of infrared and visible focal plane arrays, the determination of the detailed spatial response of individual detectors as well as the entire array can be extremely important in many electro-optic sensor applica­ tions such as the detection of targets in the presence of clutter and passive acquisition and tracking. This paper describes a computer-controlled flying spot scanning technique for the measurement of detailed focal plane detector responses as well as detector-to-detector cross talk and spurious responses. The technique uses a computer-controlled flying spot scanner and online data processing. A simple deconvolution is used to remove the known temporal responses of the detector and electronics followed by a two-dimensional decorrela- tion of the blur spot from output signal to obtain focal plane spatial response. As a natural result of this process the individual detector MTFs can be obtained. This technique has been implemented in a low-background focal plane test facility, which is also described. Several examples of actual test data are shown to demonstrate the use of this spot-scanning facility and its utility as a diagnostic tool.IntroductionIn the design and development of sophisticated electro-optic sensors for infrared (IR) and visible applications, both detailed focal plane performance characterizations and end- to-end simulations using complex computer programs are essential to the design of the sen­ sor and the optimization of its performance for a given range of scenarios and operating conditions. In earlier papers we have described our computer-aided facilities for detailed testing and characterization of infrared and visible focal plane arrays1' 2' 3 and our low- background brassboard test facility designed specifically to evaluate the end-to-end per­ formance of both scanning and staring IR sensors over a wide range operating condi­ tions.4 The brassboard test facility presents a wide range of target and background sig­ nals and image scenes to infrared focal planes which are interfaced a dedicated online minicomputer. The minicomputer acts as a programmable realtime or near-realtime signal processor through which the sensor performance can be measured and signal processing algor­ ithms can be evaluated using real signals from the focal plane detectors. This paper describes a technique for measurement of the detailed spatial responses of individual detectors and detector arrays using these facilities.In the analysis and characterization of sensors, important assumptions are made about the spatial responsivity variation of the individual detectors. Often in performance analyses (such as tracking performance analysis), the detector responsivity is assumed to be flat over the area of detector and zero outside. For some detectors, however, actual responsivity deviates significantly from this hypothetical shape. The boundaries may com­ prise a significant portion of the detector real estate. The responsivity in central area of the detector may not be flat but may have significant nonuniformities. The actual shape of the detector responsivity functions may have significant effects on the results of performance analyses for both individual detectors and focal plane arrays.In many cases, such as a fine tracker, sensor performance requirements dictate that more accurate and detailed estimate of the detector responsivity function must be obtained and utilized in further signal processing. A powerful technique for obtaining the detailed spatial responsivity function of individual detectors and focal plane arrays uses a computer-controlled flying spot scanner. However, because the blur spot of a typical spot- scan system is of the same order magnitude as detector cell size, true respon­ sivity shape cannot be obtained by a simple point-by-point spatial measurement without accounting for and removing the effect of finite blur-spot size and shape. In order to do this we utilize a recently developed two-dimensional decorrelation technique^ to obtain the detector responsivity function from detector spot-scan data obtained in our test facil­ ities. Furthermore, the modulation transfer function (MTF) of individual detectors (the magnitude of the response function in the spatial frequency domain) is also obtained.41