This PDF file contains the front matter associated with SPIE Proceedings Volume 11130 including the Title Page, Copyright information, Table of Contents, Introduction, and Conference Committee listing.

We aim to find a non-invasive methodology to evaluate meat freshness and beginning of spoilage through visible and near-infrared spectroscopy. We have used two configurations capable to sense at different probing depths. A table-top spectrophotometer equipped with an integrating sphere was utilized for a shallow probing depth (80 μm) and covered 400-1700 nm spectral range. In another configuration, a fiber-optic linear array was coupled to a portable spectrophotometer (400-1000 nm) for increasing the average probing depth up to 570 μm. According to the results, it is possible to observe the decreasing trend in the light absorbance in both visible and NIR spectral ranges showing loss of freshness and meat spoilage over time. In the visible wavelength range, absorbance changes at 540 nm and 580 nm wavelengths allow for monitoring oxymyoglobin degradation, which is associated with loss of freshness. In the NIR region, monitoring changes in the absorbance of fat (1200 nm), water (1450 nm), and proteins (1525 and 1600 nm) show promise to detect spoilage. Specifically, absorbance of the superficially located oxymyoglobin decays immediately but only after 4 hours at the depth of 0.57 mm, while absorbance of surface water and protein components experiences a steep decrease only after about 2.5 hours that could be interpreted as a beginning of the spoilage process. The fiber-optic approach capable for real-time, non-destructive, inexpensive, and multi-component detection using a wide range of studied wavelengths is a great promise to design a compact, portable device for a variety of users at the meat supply chain.

The purpose of this paper is to discuss some recent approaches to composite hypothesis testing in the context of hyperspectral target detection applications. The primary tool for the development of practical detection algorithms is the generalized likelihood ratio test (GLRT), which does not have any “built-in” optimality criterion. The GLRT computes maximum likelihood estimates of the unknown parameters and uses them in the place of the true parameters. In this paper we review the asymptotic optimality properties of GLRTs and use them to discuss a family of universal tests known as competitive Bayes and Neyman-Pearson tests. Then, based on this background, we review the family of clairvoyant fusion algorithms and their applicability to hyperspectral target detection.

Most of the techniques used for medical diagnosis, apply methods of analysis to identify certain substances called biomarkers. These techniques generally have the disadvantages of being laborious, invasive, and dependent on the physician’s experience. Raman spectroscopy is projected as a technique capable of identifying biomarkers in a noninvasive, simple and economical way. The analysis of the spectroscopic results by means of multivariable mathematical techniques would allow to eliminate the subjective interpretation of the results and therefore contribute to objective and more reliable diagnoses. The tumor suppressor wild type p53 protein is considered a cancer biomarker. Present in the human body is activated when cellular damage is detected. The p53 protein acts to protect DNA integrity: repairing the damage or inducing cellular death. When p53 do not respond correctly, the damage is not arrested, and tumor growth is developed. Mutations in p53 are related to inactivation of the wild type and therefore the presence of tumors. In this work, Raman spectra of wild type and mutants p53 were obtained through a micro-spectrometer. The spectra were analyzed by multivariate methods. Principal component analysis and support vector machine algorithms showed that it is possible to discriminate between the wild and mutant type of this biomarker with an accuracy of 94%. Raman spectra of wild type p53 at different concentrations were used to estimate the limit of the detection of this protein by means of partial least squares regression. The limit of detection was found as low as 0.946 μM without additional reagents.

We investigate deep neural networks to reconstruct and classify hyperspectral images from compressive sensing measurements. Hyperspectral sensors provide detailed spectral information to differentiate materials. However, traditional imagers require scanning to acquire spatial and spectral information, which increases collection time. Compressive sensing is a technique to encode signals into fewer measurements. It can speed acquisition time, but the reconstruction can be computationally intensive. First we describe multilayer perceptrons to reconstruct compressive hyperspectral images. Then we compare two different inputs to machine learning classifiers: compressive sensing measurements and the reconstructed hyperspectral image. The classifiers include support vector machines, K nearest neighbors, and three neural networks (3D convolutional neural networks and recurrent neural networks). The results show that deep neural networks can speed up the time for the acquisition, reconstruction, and classification of compressive hyperspectral images.

Data fusion algorithms help extract information from “asynchronous” time series satellite data whereas data merging data help extract information from “synchronous” time series satellite data into a series of synthetic images by using the temporal, spatial, or even spectral properties. Such data fusion algorithms including Bayesian maximum entropy (BME) and spatial and temporal adaptive reflectance fusion model (STARFM) have greatly improved the coverage, enhancing data application potential with higher spatiotemporal resolution via multi-sensor earth observations. The goal of this study is to assess the utility of BME and modified BME algorithm with the aid of a data merging algorithm called Modified Quantile-Quantile Adjustment (MQQA), in comparison with STARFM for the retrieval of Aerosol Optical Depth in an urban environment. MQQA heavily counts on big data to support the systematic bias correction from “synchronous” time series satellite data. Such assessment of algorithmic efficiency needs to be carried out for both top of atmosphere reflectance and ground reflectance levels in support of the deep blue method for the retrieval of atmospheric optical depth at the ground level.

As the Unmanned aerial vehicle (UAV) become ubiquitous, many early studies utilizing the UAV for three-dimensional (3D) modeling system are successively proposed as well. The advantage of using UAV successfully advances efficiency and reduces the cost time for simplifying the process of collecting data. Moreover, the characteristics of the light sized and suspension of new designed UAV could allow the users to gather the precise data from the rugged terrain and through a crowd of trees and may build a detail 3D and hyperspectral images. By those high-resolution 3D model data, the texture and shape of observation can be seen and brings research into further analysis. However, there are some limits that the information of 3D modeling data in RGB wide bands is not adequate to investigate vegetation research. The hyper spectrum collects the information of tens of bands with the narrow bandwidth could analyze the detail difference between abnormal and normal situations when the plant is growing. Therefore, how to combine hyperspectral data into a 3D modeling system would be important when researches want to attain the location data and spectral image simultaneously. In this study, the UAV will be loaded with two kinds of systems, and try to build the model include 3D modeling information and spectral information. As a result, it can monitor the growth condition of plants, their environment, also the precise location at the same time so that it can make vegetation analysis more complete.

A Compressive Sensing Snapshot Imaging Spectrometer (CSSIS) and its performance are described. The number of spectral bins recorded in a traditional tiled array spectrometer is limited to the number of filters. By properly designing the filters and leveraging compressive sensing techniques, more spectral bins can be reconstructed. Simulation results indicate that closely-spaced spectral sources that are not resolved with a traditional spectrometer can be resolved with the CSSIS. The nature of the filters used in the CSSIS enable higher signal-to-noise ratios in measured signals. The filters are spectrally broad relative to narrow-line filters used in traditional systems, and hence more light reaches the imaging sensor. This enables the CSSIS to outperform a traditional system in a classification task in the presence of noise. Simulation results on classifying in the compressive domain are shown. This obviates the need for the computationally-intensive spectral reconstruction algorithm.

A CLARREO (Climate Absolute Radiance and Refractivity Observatory) Pathfinder (CPF) mission has been funded to demonstrate retrieval of SI-traceable spectral reflectance with absolute uncertainty <0.3% (k=1). The mission consists of a Reflected Solar spectrometer that will be hosted on the International Space Station (ISS) in the 2023 timeframe and rely on a ratioing radiometer approach to retrieve the unprecedented accuracy. Demonstrating that the accuracy is achieved through an Independent Calibration effort similar in philosophy to the efforts in metrology laboratories relying on multiple, independent measurements to improve credibility for a sensor’s absolute and relative uncertainty error budgets. These measurements use different traceability paths and multiple instrument approaches and CPF’s Independent Calibration will be similar in this regard. The Independent Calibration relies on a pre-launch absolute radiometric calibration obtained from additional testing done after instrument thermal-vacuum (TVAC) testing. The added radiometric calibration is combined with a high fidelity instrument model to provide an on-orbit radiometric calibration independent from the ratioing radiometer approach. The current work describes the post-TVAC testing portion of the CPF Independent Calibration Plan and the Independent Model Development as well as planned on-orbit Evaluation of the Independent Calibration.

The Mapping Imaging Spectrometer for Europa (MISE) is a high-throughput pushbroom imaging spectrometer designed for NASA’s planned flyby mission to Jupiter’s moon Europa. The MISE design utilizes heritage from previously demonstrated instruments on airborne platforms, while advancing the state of the art to operate within Europa’s challenging environment. The instrument operates at F/1.4 and covers a spectral range from 0.8 to 5 microns with 10 nm spectral sampling. Through high resolution mapping, MISE is designed to identify distributions of organics, salts, acid hydrates, water ice phases, altered silicates, radiolytic compounds, and warm thermal anomalies at global, regional, and local scales. Such distribution maps will help study surface and subsurface geologic processes, and assess the habitability of Europa’s ocean. We discuss the optical specifications and baseline performance of the MISE optical design.

This paper will present an overview of compressive sensing for channeled polarimetry. We frame the reconstruction of the Stokes parameters as an underdetermined problem, where we solve for 3N unknowns from N measurements. We discuss two types of polarimeters: channeled spectropolarimeters and channeled linear imaging polarimeters. The polarimeters may differ in a few aspects: the output may be signals or images, the optical elements may vary, and the dimensions may be spatial or spectral. Our algorithms work with existing polarimeters and require no change in optical elements or measurement procedure. The purpose of this work is to present this framework and describe how it applies across different types of polarimeters. Both simulations and experiments show that our algorithms produce more accurate reconstructions with less artifacts than frequency domain filtering

Here we report progress in the fabrication, calibration, and testing of a compact spectral imaging camera. The camera uses a micro-array of Fabry-Perot etalons bonded directly to a broadband focal plane array sensor. The array of etalons adds negligible size and weight to the system compared to a panchromatic imager. Other recent demonstrations of compact spectral imagers in the visible and near infrared (VNIR) have commonly used arrays of etalons in a single order, thereby reducing the system bandwidth and sensitivity to achieve the required spectral resolution. Here, we demonstrate a camera utilizing multiple etalon orders in a spectral multiplexing scheme known as Multiple Order Staircase Etalon Spectrometry (MOSES). An important characteristic of the MOSES system is that there is a relaxed tradeoff between spectral resolution and sensitivity (or etalon throughput). Unlike single-order etalon techniques, MOSES allows for the reconstruction of the spectrum to the bandwidth limit of the detector and reflecting layers. This is important in coastal environmental sensing, where IR spectral features may be desired at the same time as UV light transmitted through shallow water layers. This VNIR system demonstration indicates the feasibility of MOSES devices in other wavebands.

A mid-infrared spectroscopic imager needs to be portable and tough for mounting on drones or small satellites. Meanwhile, an extremely compact and tough hyperspectral camera with mass less than 2 kg is required for mounting on drones to observe nutritive components like nitrogen and phosphorus. We proposed a near-common-path wavefrontdivision phase-shift interferometer as an imaging-type two-dimensional Fourier spectrometer. Because the proposed interferometer has strong robustness against mechanical vibrations, a palm-sized Fourier spectroscopic imager can be realized without an anti-mechanical vibration system. We developed a palm-sized (80-mm cube weighing 0.5 kg) and tough hyperspectral camera for mid-infrared light (wavelength of 8–14 μm) that can be operated using only a notebook personal computer. Furthermore, the field of view of a conventional hyperspectral camera is narrow (e.g., 6.4 deg × 5.1 deg). However, employing a proposed field angle correlation method and using a fisheye lens as the objective lens, the field of view can be expanded to 180 deg. The total price of the mid-infrared two-dimensional spectroscopic imager is no more than several thousand USD because a low-price microbolometer (Vision Sensing, VIM-80G2, wavelength range: 8–14 μm, 80 × 80 pixels, price: 300 USD) is used. Additionally, a long-stroke (10 mm) and high-resolution (Optical encoder resolution: 100 nm) impact-drive actuator (Technohands XCWT70-10 weighing 30 g) is introduced as a lowprice (1,000 USD) and tough phase-shift stage with cross-roller linear-motion guides.

Prescribed burns are being considered as a management tool for the prevention of forest fires in many countries that have important firefighting problems. Knowledge of fire intensity and eliminated vegetation fuel are of great interest to evaluate their effectiveness. Both parameters are directly related to burn severity, so their evaluation is fundamental to predict the post-fire evolution of burned area. In this study we evaluated two prescribed burnings carried out in North of Spain during October 2017 by using multispectral data from an Unmanned Aerial Vehicle (UAV). In particular, four surface reflectance images were obtained in green (550 nm), red (660 nm), red-edge (735 nm) and near infrared (790 nm) at very high spatial resolution (GSD 20 cm) from which different spectral indexes were computed. Additionally, vegetation and soil burn severity was measured in 153 field plots and an analysis of variance (ANOVA) between each spectral index and burn severity (both in vegetation and soil) was performed. A Fisher’s least significant difference test determined that three vegetation burn severity levels and two soil burn severity levels could be statistically distinguished. The identification of such burn severity levels is sufficient and useful to forest managers. We conclude that multispectral data from UAVs may be considered as a valuable indicator of burn severity for prescribed burnings.