Day night imaging application requires high dynamic range optical imaging system to detect targets of interest covering mid-day (>32000 Lux), and moonless night (∼1mLux) under clear sky- (visibility of >10km, atmospheric loss of <1dB/km) and hazy (visibility of >500m, atmospheric loss of >15dB/Km) conditions. Major governing factors for development of such camera systems are (i) covert imaging with ability to identify the target, (ii) imaging irrespective to the scene background, (iii) reliable operation , (iv) imaging capabilities in inclement weather conditions, (v) resource requirement vs availability power & mass, (vi) real-time data processing, (vii) self-calibration, and (viii) cost. Identification of optimum spectral band of interest is most important to meet these requirements. Conventional detection systems sensing in MWIR and LWIR band has certain draw backs in terms of target detection capabilities, susceptibility to background and huge thermo-mechanical resource requirement. Alternatively, range gated imaging camera system sensing in NIR/SWIR spectrum has shown significant potential to detect wide dynamic range targets. ToF Camera configured in NIR band has certain advantages in terms of Focal Plane Assembly (FPA) development with large format detectors and thermo-mechanical resource requirement compared to SWIR band camera configuration. In past, ToF camera systems were successfully configured in NIR spectrum using silicon based Electron Multiplying CCD (EMCCD), Intensifier CCD (ICCD) along with Gating device and pulsed laser source having emission in between 800nm to 900nm. However, these systems have a very low dynamic range and not suitable for clear sky mid-day conditions. Recently silicon based scientific grade CMOS image sensors have shown significant improvement in terms of high NIR responsivity and available in bigger formats (5MP or more), adequate Full well capacity for day time imaging (>30Ke), very low readout noise (<2e) required for night imaging and higher frame rate (more than 100fps). Taking advantage of these, laser based camera system configuration was worked out and presented in this paper using scientific grade CMOS sensor and NIR Laser. Camera can image target range from 4km to 5km with resolution of 5cm. Camera can have instantaneous coverage of 100mx100m (at 5km). Scientific grade CMOS sensor could also be used for clear sky day time imaging conditions with Laser off condition. To reduce the laser energy requirement, FPA required to be operated in multi-integration mode where multiple low energy pulses could be thrown within given integration time and detector and its associated electronics will collect and accumulate only those photons which are reflected back from the target of interest using appropriate gating control mechanism. Paper will bring out system engineering aspects for finalization of imaging spectrum, optical parameters in terms of aperture & focal length, required laser energy, highlighting advantage of pulse mode operation of laser compared to continuous mode operation in terms of laser energy & back-scattered light, silicon based optical detector performance results and post processing aspects for target detection. Paper will also discuss achieved performance of proto-model camera.
The push-broom sensors in bands meant to study oceans, in general suffer from residual non uniformity even after radiometric correction. The in-orbit data from OCM-2 shows pronounced striping in lower bands. There have been many attempts and different approaches to solve the problem using image data itself. The success or lack of it of each algorithm lies on the quality of the uniform region identified. In this paper, an image based destriping algorithm is presented with constraints being derived from Ground Calibration exercise. The basis of the methodology is determination of pixel to pixel non-uniformity through uniform segments identified and collected from large number of images, covering the dynamic range of the sensor. The results show the effectiveness of the algorithm over different targets. The performance is qualitatively evaluated by visual inspection and quantitatively measured by two parameters.