Characterizing and Mitigating Sensor Generated Spatial Correlations in Airborne Hyperspectral Imaging Data

Abstract

In hyperspectral imaging (HSI), the spatial contribution to each pixel is non-uniform and extends past the traditionally square spatial boundaries designated by the pixel resolution, resulting in sensor-generated blurring effects. The spatial contribution to each pixel can be characterized by the net point spread function, which is overlooked in many airborne HSI applications. The objective of this study was to characterize and mitigate sensor blurring effects in airborne HSI data with simple tools, emphasizing the importance of point spread functions. Two algorithms were developed to (1) quantify spatial correlations and (2) use a theoretically derived point spread function to perform deconvolution. Both algorithms were used to characterize and mitigate sensor blurring effects on a simulated scene with known spectral and spatial variability. The first algorithm showed that sensor blurring modified the spatial correlation structure in the simulated scene, removing 54.0%–75.4% of the known spatial variability. Sensor blurring effects were also shown to remove 31.1%–38.9% of the known spectral variability. The second algorithm mitigated sensor-generated spatial correlations. After deconvolution, the spatial variability of the image was within 23.3% of the known value. Similarly, the deconvolved image was within 6.8% of the known spectral variability. When tested on real-world HSI data, the algorithms sharpened the imagery while characterizing the spatial correlation structure of the dataset, showing the implications of sensor blurring. This study substantiates the importance of point spread functions in the assessment and application of airborne HSI data, providing simple tools that are approachable for all end-users.

Spectral Field Campaigns: Planning and Data Collection

“…If the size of a structural cell is larger than the image pixel plus its accuracy buffer, structural cells should be preferred; their size can vary depending on the variation of the environmental feature of interest. It is also worth mentioning that pixels as seen by a sensor are not square but elliptic and that surrounding pixels contribute substantially to the signal detected per focus pixel (Inamdar et al 2020). Theoretically, elliptic or hexagonal sampling units, representing shapes that are frequent in nature, should capture local environmental conditions better than square plots.…”

Deciphering the many maps of the Xingu – an assessment of land cover classifications at multiple scales

“…It comprises the majority of the area on the ground contributing signal to a pixel (see Figure 3). Most often used to describe an image after it has been geometrically corrected and resampled (to square pixels) but can also refer to the raw data with unaltered geometry (see Inamdar et al 2020 for an example). Importantly, neighboring pixels are not independent…”

“…Pixels are most commonly represented as squares following geocorrection and resampling. See Inamdar et al (2020) for additional details. Images subsets are an illustration of different pixel sizes for the differentiation of land covers suchs water, rocks and vegetation.…”

The Quality of Remote Sensing Optical Images from Acquisition to Users

“…Continuing the presentation of the papers of the special issue, two of them [3,4] consider the blurring effects and noise introduced into the images by the acquisition system.…”

“…In particular, in [3], the authors develop two simple algorithms to characterize and mitigate sensor-generated spatial correlations while emphasizing the implications of sensor point spread functions, i.e., considering the characteristics of the modulation transfer functions (MTF). In hyperspectral imaging (HSI), in fact, the spatial contribution to each pixel is nonuniform and extends past the traditionally square spatial boundaries designated by the pixel resolution, resulting in sensor-generated blurring effects.…”