Pushbroom Imaging Research Articles

New methodology of hyperspectral information extraction and accuracy assessment based on a neural network

Spectral analysis is an important research topic in the field of hyperspectral remote sensing. In this paper, the pixel vectors of hyperspectral imagery are taken as research object. A new method of spectral analysis and hyperspectral image information extraction were studied based on a Neural Network (NN). An application of this method in disaster measurement of wheat stripe rust disease is described. The experimental data was hyperspectral imagery of wheat-infected with stripe rust and came from the Push-broom hyperspectral imager (PHI). The NN model of Self-organizing Feature Maps (SOFM) was introduced to cluster and analysis of the disease severity. By means of computing the spectral index data (SID) and spectral angle data (SAD) as well as spectral average reflectance data (ARD) of hyperspectral imagery in experimental area, the three kinds of two-dimensional data matrixes were used as the inputs of the SOFM model. After iterative learning and self-organized clustering, the outputs of the model were mapped in 3-dimensional severity space of wheat stripe rust. The computed data and image data in the experimental area were simulated through net-training. The trees and soil in the experimental area are easily interpreted by comparison of sample’s spectral curves. The disease and healthy sampling points were identified by “Red-edge” positions. Then, the degrees of disease severity were established by spectral threshold analysis of samples in near-infrared bands. Finally, a cross correlogram spectral matching technique was used to assess the accuracy. The study results show that the disease of wheat stripe rust was severe. The whole trial spot was divided into six feature classes, including four severity degrees of wheat stripe rust disease, they are healthy, light, mid and heavy. These results are reliable and accurate.

DEM matching for bias compensation of rigorous pushbroom sensor models

DEM matching is a technique to match two surfaces or two DEMs, at different reference frames. It was originally proposed to replace the need of ground control points for absolute orientation of perspective images. This paper examines DEM matching for precise mapping of pushbroom images without ground control points. We proved that DEM matching based on 3D similarity transformation can be used when model errors are only on the platform’s position and attitude biases. We also proposed how to estimate bias errors and how to update rigorous pushbroom sensor models from DEM matching results. We used a SPOT-5 stereo pair at ground sampling distance of 2.5 m and a reference DEM dataset at grid spacing of 30 m and showed that rigorous pushbroom models with accuracy better than twice of the ground sampling distance both in image and object space have been achieved through DEM matching. We showed further that DEM matching based on 3D similarity transformation may not work for pushbroom images with drift or drift rate errors. We discussed the effects of DEM outliers on DEM matching and automated removal of outliers. The major contribution of this paper is that we validate DEM matching, theoretically and experimentally, for estimating position and attitude biases and for establishing rigorous sensor models for pushbroom images.

Operational Atmospheric Correction for Imaging Spectrometers Accounting for the Smile Effect

Hyperspectral pushbroom imagers are affected by a number of artifacts, such as pixel nonuniformity, spectral smile, and keystone. These have to be taken into account during system correction, orthorectification, or atmospheric correction, as performed in processing and archiving facilities (PAFs). This contribution is presenting an efficient and accurate smile correction method integrated in the atmospheric correction. The proposed technique will be used in the PAF of the German hyperspectral Environmental Mapping and Analysis Program mission. The spectral smile shift across the detector array is parametrized with a fourth-order polynomial function for each channel based on the instrument optical design model or measured laboratory data. Alternatively, spectral smile shifts can be calculated from image data using channels in atmospheric absorption regions. The concept for the time-optimized processor is outlined, and the results are presented for simulated EnMAP data and existing pushbroom imagery [HYPERION, AISA (Airborne Imaging Spectrometer for Applications), and HYSPEX (Hyperspectral Camera)].

On-Orbit Calibration of Interior Orientation for HJ1B-CCD Camera

In order to acquire the on-orbit value of interior orientation and improve the orientation precision of remote sensing image,a method for the on-orbit calibration of HJ1B-CCD camera interior orientation was studied with linear array push-broom sensor imaging model in this paper.Firstly,the ground control points are selected and colinearity equations are constructed between ground control points and relevant image points.Secondly,the error equations based on colinearity equations are built and the initial value of unknown parameters are estimated.Finally,the error equations are solved by special ridge estimation to acquire the interior and exterior orientation.The results showed that this method could calibrate the HJ1B-CCD camera interior orientation on-orbit with good stability and precision.

推扫型干涉成像光谱仪去除条带非均匀性的方法

The structure and work principle of push-broom interferential imaging spectrometer was analyzed.It was found that the brightness of interferential data was non-uniform along the slit direction by the processed error of slit,the stripe noises along the push-broom direction were added on the restored spectral-image cube,image quality was degraded,and spectral information was difficult to evaluated.The corrected coefficient method was used to reduce stripes,and several coefficient-extracted methods were discussed.The non-uniformity of interferential data was extracted from the calibration of instruments,and the corrected coefficient was multipled by the interferential data to reduce the stripes.The results show that numerous stripes are reduced.Without the on-orbit calibration data,the coefficient is modified by the spectral-image cube contained uniform targets,and the remanent stripes are reduced efficiently by the modified coefficient.

A Piecewise Approach to Epipolar Resampling of Pushbroom Satellite Images Based on RPC

Epipolar line determination and image resampling are important steps for stereo image processing. Unlike frame cameras that have well-known epipolar geometry, the pushbroom camera does not produce straight epipolar lines and the epipolar pair does not exist for the entire scene. These properties make it difficult to establish epipolar geometry of the pushbroom camera for epipolar image resampling. Therefore, some researchers have adopted approximate models to avoid these problems. In this study, a new method for the conjugate epipolar curve pair determination and epipolar resampling of spaceborne pushbroom images based on RPC is proposed. The proposed method assumes that the conjugate epipolar curve pairs exist approximately for the local scene area and the global epipolar pairs can also exist if the local pairs are sequentially linked. Then, epipolar image resampling is established by reassigning the generated conjugate epipolar curve pair points to satisfy the epipolar resampling image condition. Ikonos stereo images are tested for the evaluation, and the proposed method showed a maximum y-parallax of 1.25 pixels for manually measured tie points, while the resampling method by the parallel projection model showed a maximum of 4.59 pixels.

Determination of the Indirect Orientation of Orbital PUSHBROOM Images Using Control Straight Lines

Abstract The aim of this paper is to present an experimental assessment of two models that use “control lines” for the indirect orientation of pushbroom imagery. Since pushbroom image acquisition is not instantaneous, six exterior orientation parameters (EOPs) must be estimated for each scanned line. The sensor position and attitude parameters are modelled with a time‐dependent polynomial. The relationship between a straight line in the image space and its homologous form in the object space is established in the first model, based on the principle that the position vector containing an image point (projection ray) and the vector normal to the projection plane in the object space are orthogonal. The second model is based on the equivalence between the vector normal to the projection plane in the image space and the vector normal to the rotated projection plane in the object space. The equivalence property between planes was adapted to consider the pushbroom geometry. A model based on collinearity equations using points adapted to the pushbroom geometry was also implemented, aiming at a comparison of the methodologies. Six experiments using different sets of observations for indirect estimation of EOPs of images from the China–Brazil Earth Resources Satellite (CBERS) were carried out, by varying the geometric distribution and the number of straight lines. Also, experiments combining points and straight lines were accomplished. The results showed that an accuracy of around twice the ground sample distance (GSD) in the check points can be achieved with the models studied, which can then be used to estimate the EOPs of pushbroom images. Several other factors affecting the accuracy, such as the distribution and number of control features, were also assessed.

In-flight spectral performance monitoring of the Airborne Prism Experiment

Spectral performance of an airborne dispersive pushbroom imaging spectrometer cannot be assumed to be stable over a whole flight season given the environmental stresses present during flight. Spectral performance monitoring during flight is commonly accomplished by looking at selected absorption features present in the Sun, atmosphere, or ground, and their stability. The assessment of instrument performance in two different environments, e.g., laboratory and airborne, using precisely the same calibration reference, has not been possible so far. The Airborne Prism Experiment (APEX), an airborne dispersive pushbroom imaging spectrometer, uses an onboard in-flight characterization (IFC) facility, which makes it possible to monitor the sensor's performance in terms of spectral, radiometric, and geometric stability in flight and in the laboratory. We discuss in detail a new method for the monitoring of spectral instrument performance. The method relies on the monitoring of spectral shifts by comparing instrument-induced movements of absorption features on ground and in flight. Absorption lines originate from spectral filters, which intercept the full field of view (FOV) illuminated using an internal light source. A feature-fitting algorithm is used for the shift estimation based on Pearson's correlation coefficient. Environmental parameter monitoring, coregistered on board with the image and calibration data, revealed that differential pressure and temperature in the baffle compartment are the main driving parameters explaining the trend in spectral performance deviations in the time and the space (across-track) domains, respectively. The results presented in this paper show that the system in its current setup needs further improvements to reach a stable performance. Findings provided useful guidelines for the instrument revision currently under way. The main aim of the revision is the stabilization of the instrument for a range of temperature and pressure conditions to be encountered during operation.

Concatenation of HRSC colour and OMEGA data for the determination and 3D-parameterization of high-altitude CO 2 clouds in the Martian atmosphere

We used Mars Express HRSC and OMEGA data to investigate mesospheric cloud features observed in the equatorial belt of Mars from December 2007 until early March 2008. This period corresponds to early northern spring of Martian year 29. The reflection peak at 4.26 μm in OMEGA data identifies the clouds as CO 2 ice clouds. HRSC observed the clouds together with OMEGA in five orbits. Cloud features are most prominent in the shortwave HRSC colour channels with wavelength centers at 440 and 530 nm, but rarely visible in all other channels. In the period of Ls 0–36°, OMEGA and HRSC together detected mesospheric CO 2 ice clouds in 40 orbits. They occur in a latitude belt of ±20° around the equator and at longitudes between 240°E (Tharsis) in the West and 30°E (Sinus Meridiani) in the East. The clouds were observed between 3 and 5 p.m. local time with mainly ripple-like to filamentary cloud forms. The viewing angles of the HRSC blue and green colour channels differ by 6.6° and the resulting parallax can be used to directly measure cloud heights by means of ray intersection. 17 HRSC data takes were found to exhibit clouds with heights from 66 to 83 km with an accuracy of 1–2 km. The pushbroom imaging technique also yields a time delay for the two observations in the order of 5–15 s close to periapsis, and therefore time-related cloud movements can be detected. A method was developed to determine the across-track cloud displacements, which can directly be translated to wind velocities. Zonal cloud movements could be measured in 13 cases and were oriented from East to West. Related wind speeds range between 60 and 93 m/s with an accuracy of 10–13 m/s.

Detection and Restoration of Defective Lines in the SPOT 4 SWIR Band

This paper presents the categorization and restoration of defective lines developed in pushbroom images. About 100 of the 3000 SPOT 4 SWIR detectors malfunction, which degrades image quality. Conventional methods have difficulties in effectively detecting and restoring defective lines, because they ignore the heterogeneity of the ground surface and the presence of sporadically unstable detectors with gain and offset that vary during a scan. While all defective lines have previously been considered as a single type, here they are categorized into three types according to the variation pattern in the scanning direction: constant defective lines, irregular defective lines, and irrecoverable defective lines. The detection procedure utilizes summed data and standard deviation data that consist of abnormal peaks originating from defective lines and a slowly varying baseline reflecting the surface characteristics within the image. The defective lines are detected by finding abnormal peaks, and classified and restored by using either a moment-matching method or interpolation, depending upon their types. Three SPOT 4 images were used to test and evaluate the performance of the proposed method. From the test results, the constant defective line was the most common type, comprising about 60%, while the irregular defective lines caused serious image degradation because of the difficulty of detecting and classifying them. Commission and omission errors were less than 10% and detection accuracy was higher than 90%. The analysis of signal-to-noise ratio (SNR) showed that the low SNR created by the defective lines was effectively removed. Our method gave a significant improvement of the detection and restoration capability.

Indirect Detection of Fusarium Verticillioides in Maize (Zea mays L.) Kernels by near Infrared Hyperspectral Imaging

Near infrared (NIR) hyperspectral imaging and hyperspectral image analysis were evaluated for their potential to distinguish between Fusarium verticillioides infected and sound whole maize ( Zea mays L.) kernels. Hyperspectral images of infected and sound kernels were acquired using a Spectral Dimensions MatrixNIR camera with a spectral range of 960–1662nm and a sisuChema hyperspectral pushbroom imaging system with a spectral range of 1000–2498 nm. Background, bad pixels and shading were removed using exploratory principal component analysis (PCA) on absorbance images. PCA could be used effectively on the cleaned images to identify classes including infected and non-infected regions on individual kernels. A distinct difference between infected and sound kernels along principal component (PC) one with two distinguishable clusters was found. The loading line plot of the first PC of the sisuChema hypercube showed important absorbance peaks for the two classes, i.e. 1960nm and 2100nm for the infected class and U50 nm, 2300nm and 2350nm for the non-infected class. Partial least squares discriminant analysis (PLS-DA) was applied. The coefficient of determination was 0.73 for the MatrixNIR image and 0.86 for the sisuChema image, both after three PLS components. These PLS-DA models could be used to calculate predictions on a test set image. The predictions could be shown as prediction images and an acceptable root mean square error of prediction was obtained (0.23). NIR hyperspectral imaging has the potential to be used as a rapid, objective means of indentifying fungal infected maize kernels and infected regions.

Design and fabrication of a 900–1700 nm hyper-spectral imaging spectrometer

This paper presents a 900–1700 nm hyper-spectral imaging spectrometer which offers low distortions, a low F-number, a compact size, an easily-fabricated design and a low cost (is presented in this paper). The starting point for its optical design is discussed according to the geometrical aberration theory and Rowland circle condition. It is shown that these methods are useful in designing a push-broom hyper-spectral imaging spectrometer that has an aperture of f/2.4, modulation transfer functions of less than 0.8 at 25 cycles/mm, and spot sizes less than 10 μm. A prototype of the optimized hyper-spectral imaging spectrometer has been fabricated using a high precision machine and the experimental demonstration with the fabricated hyper-spectral imaging spectrometer is presented.

Maize kernel hardness classification by near infrared (NIR) hyperspectral imaging and multivariate data analysis

The use of near infrared (NIR) hyperspectral imaging and hyperspectral image analysis for distinguishing between hard, intermediate and soft maize kernels from inbred lines was evaluated. NIR hyperspectral images of two sets (12 and 24 kernels) of whole maize kernels were acquired using a Spectral Dimensions MatrixNIR camera with a spectral range of 960–1662 nm and a sisuChema SWIR (short wave infrared) hyperspectral pushbroom imaging system with a spectral range of 1000–2498 nm. Exploratory principal component analysis (PCA) was used on absorbance images to remove background, bad pixels and shading. On the cleaned images, PCA could be used effectively to find histological classes including glassy (hard) and floury (soft) endosperm. PCA illustrated a distinct difference between glassy and floury endosperm along principal component (PC) three on the MatrixNIR and PC two on the sisuChema with two distinguishable clusters. Subsequently partial least squares discriminant analysis (PLS-DA) was applied to build a classification model. The PLS-DA model from the MatrixNIR image (12 kernels) resulted in root mean square error of prediction (RMSEP) value of 0.18. This was repeated on the MatrixNIR image of the 24 kernels which resulted in RMSEP of 0.18. The sisuChema image yielded RMSEP value of 0.29. The reproducible results obtained with the different data sets indicate that the method proposed in this paper has a real potential for future classification uses.

A Fast Approach to Best Scanline Search of Airborne Linear Pushbroom Images

The linear pushbroom camera has become one of the most important imaging sensors in today's photogrammetry and remote sensing practices. Airborne digital sensors (ADS) or three-line scanner (TLS) imaging system such as the ADS40 from Leica Geosystems and STARIMAGER from STAR-LABO Corporation use the pushbroom technique to collect high-resolution, multi-channel seamless image strips. As we know, the object-to-image coordinate computation serves as a core step during the process of photogrammetric images. However, each scanline captured by a linear pushbroom sensor has six exterior orientation ( E O) parameters at the corresponding instant of exposure. The image point coordinates will not be accurately calculated through colinearity equations unless reasonable EO parameters are determined. Therefore, the best scanline search (BSS) has direct effects on efficiency of object-to-image coordinate computation during image processing. This paper addresses a fast BSS method based on the novel central perspective plane of scanline (CPPS) constraints. The search process is simply performed through analytical geometric calculations, which can significantly improve the efficiency of the object-to-image coordinate computation. The feasibility and robustness of the proposed method have been verified using ADS40 and STARIMAGER images. Experimental results show that the proposed method improves scanline search speed considerably with the time cost decreased by nearly 85 percent compared with the traditional methods.

Precision Forestry Using Airborne Hyperspectral Imaging Sensor

Universiti Putra Malaysia in collaboration with a private company, Aeroscan Precision (M) Sdn Bhd based in UPMSerdang conducted a research and commercial applications of airborne hyperspectral sensing data in precision forestry.The UPM-APSB’s AISA sensor was flown over a representative series of forested areas in Peninsular Malaysia from12-20th. July 2004. It is a pushbroom imaging spectrometer recording remote sensing images over a large spectrum ofwavelengths from the visible (400 nm) to near infrared (1 000 nm). Images have a ground pixel size of 1m by 1 m at aflight altitude of 1 000 m a.s.l and a constant flight speed of 120 knots. The ground validation segment of the projectswas focused around hill/montane dipterocarps,. In order to characterize the properties and status of the forests, a numberof images and field spectrum were developed. Prior to and after the flight, field spectral reflectance measurements usinga handheld FieldSpec spectroradiometer were taken over the timber species of interest. The use of spectral unmixingmethods for the discrimination of individual timber species image components leading to a more accurate identificationof timber species, timber inventory and volume estimates were evaluated. The results imply that UPM-APSB airbornehyperspectral imaging technology would enable the development of a rapid forest resources assessment, especially inthe sustainable forest management in Malaysia and other tropical countries.

ORIENTAÇÃO INDIRETA DE IMAGENS CBERS: AVALIAÇÃO DE TÉCNICAS QUE USAM LINHAS RETAS E SUA COMBINAÇÃO COM PONTOS

The aim of this paper is to present and assess techniques for orientation of pushbroom sensors that allow the estimation of the polynomial coefficients describing the platform trajectory, using linear features and linear features combined with points as ground control. The pushbroom image acquisition is not instantaneous and, as a consequence, six EOP (Exterior Orientation Parameters) for each scanned line must be estimated. The sensor position and attitude parameters are modeled with a time dependent polynomial. The proposed mathematic model is based on the equivalence property between planes, considering the pushbroom geometry. The relationship between object and image space is established through a mathematical model based on the equivalence between the vector normal to the projection plane in the image space and to the vector normal to the rotated projection plane in the object space. The model based on collinearity equations using points adapted to the pushbroom geometry was also implemented aiming at the comparison of the techniques, as well as to evaluate the use of control points and straight lines simultaneously. Some experiments using a CBERS scene were accomplished in order to test the developed techniques. The obtained results showed that is possible estimate the EOP of pushbroom images using straight lines and that the combination of straight lines with points is a more flexible alternative that allows good results.

Uniformity of Imaging Spectrometry Data Products

The increasing quantity and sophistication of imaging spectroscopy applications have led to a higher demand on the quality of Earth observation data products. In particular, it is desired that data products be as consistent as possible (i.e., ideally uniform) in both spectral and spatial dimensions. Yet, data acquired from real (e.g., pushbroom) imaging spectrometers are adversely affected by various categories of artifacts and aberrations including as follows: singular and linear (e.g., bad pixels and missing lines), area (e.g., optical aberrations), and stability and degradation defects. Typically, the consumer of such data products is not aware of the magnitude of such inherent data uncertainties even as more uncertainty is introduced during higher level processing for any particular application. In this paper, it is shown that the impact of imaging spectrometry data product imperfections in currently available data products has an inherent uncertainty of 10%, even though worst case scenarios were excluded, state-of-the-art corrections were applied, and radiometric calibration uncertainties were excluded. Thereafter, it is demonstrated how this error can be reduced (<5%) with appropriate available technology (onboard, scene, and laboratory calibration) and assimilation procedures during the preprocessing of the data. As a result, more accurate, i.e., uniform, imaging spectrometry data can be delivered to the user community. Hence, the term uniformity of imaging spectrometry data products is defined for enabling the quantitative means to assess the quality of imaging spectrometry data. It is argued that such rigor is necessary for calculating the error propagation of respective higher level processing results and products.

Formulation of distortion error for the line-of-sight (LOS) vector adjustment model and its role in restitution of SPOT imagery

A recently developed line-of-sight (LOS) vector adjustment model is an effective geopositioning method for pushbroom images. A requirement for this method is that a predicted orbit be close to a true orbit. Although the method has proven very effective for SPOT 5, which has an accurate orbital control system, it was an open question whether or not the method could be effective for a satellite system with poor orbital accuracy. In this paper, a distortion error is defined and formulated as a common criterion for testing orbital requirements of a pushbroom system from the perspective of the LOS vector adjustment model. Using the distortion error, it is shown that a satellite system, even with poor locational accuracy comparable to SPOT 3–4, meets the orbital requirements as long as the maximum topographic height in the landscape is less than 4 km against the height level of the control points. A test was performed using three SPOT images and only five ground control points (GCPs). Results indicated that the root mean squared errors (RMSEs) of horizontal residual errors calculated from 30 check points were less than 10 m. The achieved accuracy of three-dimensional object-point determination was 6 m in the X -dimension and 7 m in both Y - and Z -dimensions. The conclusion can be drawn that the orbital requirements of this method can accommodate most space-borne systems, and therefore the LOS vector adjustment model is effective for SPOT and comparable pushbroom imagery.

Automatic Georeferencing of Airborne Pushbroom Scanner Images With Missing Ancillary Data Using Mutual Information

We describe a methodology that is used for the automatic georeferencing of multispectral images acquired from airborne pushbroom imaging cameras. This methodology is based upon the dense registration of flight strips data onto a reference orthoimage and uses the mutual information criterion to make the raw (source) image and the orthoimage (target) aligned. By taking into account a rigorous modeling of the pushbroom imaging process, we show how the mutual information between the raw image and the orthoimage can be used to estimate unknown or inaccurate flight attitude parameters [(tilt bias angles, instantaneous yaw, and height above ground level (AGL)] without using any ground control nor tie points. Moreover, we show how a coarse digital elevation model can be incorporated in the proposed procedure to improve the geocorrection and even estimate a digital surface model of the surveyed area. Two experiments using Compact Airborne Spectrographic Imager (CASI) and Airborne Imaging Spectroradiometer for Applications (AISA) Eagle sensor data are described and show the effectiveness of our approach, with planimetric root mean square errors being in the range of two to four pixels, depending on the flatness and land cover, and altimetric root mean square errors being less than 1% of the flight altitude AGL.

A Generic Model for Along-track Stereo Sensors Using Rigorous Orbit Mechanics

In this paper a generic, rigorous sensor model for highresolution optical satellite sensors with along-track stereoscopic capabilities is introduced. The idea is to determine the orbit of the satellite platform covering the time acquisition of all images, using satellite photogrammetry in combination with astrodynamics, trying to find common exterior orientation parameters for all images directly or indirectly. As a result, the number of unknown parameters is reduced and also the correlation between them, thus giving a more stable solution. Great effort is made in order to define the essential forces which are involved in the acquisition of the pushbroom images, according to the needed accuracy and the data provided. The fundamental assumption is that Keplerian motion is maintained along the acquisition time of all the along-track images. Various versions of the model are developed based on different orbit determinationpropagation methods. An accuracy assessment is made of the above different orbit determination-propagation methods. It is possible to extract the exterior orientation of all images together directly, without ground control points and using the metadata information, with acceptable accuracy. The model is evaluated using SPOT5-HRS imagery with precision close to pixel size. Moreover, the accuracy of the along-track model is compared with the accuracy of single image sensor model.