1. Introduction

All types of existing imaging spectrometers acquire two-dimensional spatial information and one-dimensional spectral information (i.e., a three-dimensional data cube) of an object or a scene [1–3]. The spectral image can be obtained by the following methods: (1) Wavelength-scan methods that measure the images one wavelength at a time. Typical representative of this method is color filter imaging spectrometer that mainly uses either a circular-variable filter [4], a liquid-crystal tunable filter [5–8] or an acousto-optical tunable filter [9–15] to capture a full spectral image by measuring one image at a time but each time at a different wavelength. (2) Spatial-scan methods that measure the whole spectrum of a portion of the image at a time and scan the image (e.g., line by line). Typical representative of this method is dispersive imaging spectrometer that uses a prism, a grating, or both to achieve the dispersion of light and to collect the entire spectral image by measuring only one line of the object at a time and by scanning the object line-by-line (e.g., whisk-broom scanning [16], push-broom scanning [17]). (3) Time-scan methods that measure a set of images where each one of them is a superposition of spectral or spatial image information and therefore require a mathematical transformation of the acquired data to derive the spectral image (e.g., by Fourier transform or Hadamard transform [18]). Typical representative of this method is interferometric imaging spectrometer (i.e., Fourier transform imaging spectrometer) that is based on temporal interferometry or spatial interferometry [19–34]. (4) Compromise methods that measure the whole spectral image simultaneously, but compromise on the number of points in the spectrum, the field of view (FOV) or spatial resolution. Typical representatives of this method are computed-tomography imaging spectrometer [35–41] and snapshot spectral imagers [42–49]. (5) Spatial-Time joint scan method that uses a time-scan method to measure the spectral image of each line of the object and also uses the spatial-scan method to obtain the entire spectral image by scanning the object line-by-line (i.e., push-broom scanning). Typical representative of this method is coherent-dispersion imaging spectrometer [50].

When measuring an infrared broad-band source or an ultraviolet-visible narrow-band source, interferometric spectrometry (especially temporal interferometry based on linear scanning, such as Michelson-type interferometer) can perform high spectral resolution measurements. Nevertheless, when measuring an ultraviolet-visible broad-band source, the interferometric spectrometry performance is weakened [27,51–60]. That is, interferometric spectrometry is only suitable for low and medium spectral resolution measurements of a broad-band source in the ultraviolet-visible spectral region. Interestingly, the coherent-dispersion spectrometry is an effective and unique method to perform high spectral resolution measurements of a broadband source in the ultraviolet-visible spectral region [50,59, 60].

On the other hand, three-dimensional imaging technologies can potentially capture the three-dimensional structure, range, and texture information of an object. According to the characteristics of the reconstructed images, three-dimensional imaging technologies can be roughly classified into four categories: binocular, multi-view, volumetric imaging, and spatial imaging [61,62]. Typical representatives of spatial imaging are holographic imaging and integral imaging [63–68]. It is a pity that all existing imaging spectrometers are based on two-dimensional imaging technologies.

This paper presents a coherent-dispersion stereo-imaging spectrometer, which can not only obtain three-dimensional stereoscopic image, but also perform high spectral resolution measurements of a broadband source in the ultraviolet-visible spectral region. After a detailed description of the principle, preliminary numerical calculations are illustrated by an example for the spectral range from 400 nm to 700 nm. Finally, the conclusion is given.

2. Principle

Figure 1 shows the equivalent orthographic view for the optics of the three-area-array coherent-dispersion stereo-imaging spectrometer (CDSIS), and Fig. 2 shows the equivalent light path diagram of the CDSIS. The CDSIS consists of a moving corner-cube-mirror interferometer, a plane transmission grating, an objective lens, a collimating lens, a collecting lens, three entrance slits, and three area-array detectors. The moving corner-cube-mirror interferometer comprises one moving corner-cube mirror (CCM), one fixed CCM, and one beam splitter. The optical path difference is created by the linear scanning of the moving CCM driven by a linear actuator. The interferometer has neither tilt nor shearing problems. The back focal plane of the objective lens is coincident with the front focal plane of the collimating lens. Three entrance slits (S1, S2 and S3) are located at the front focal plane of the collimating lens (i.e., back focal plane of the objective lens) and are parallel to each other. The entrance slit S2 is located in the center of the field of view of the optical system, so it intersects the optical axis of the optical system. The entrance slits S1 and S3 are symmetric about the entrance slit S2. The three entrance slits are perpendicular to the flight direction of the instrument platform (e.g., satellite), but parallel to the linear scanning direction of the moving CCM. Three area-array detectors (A, B and C) are located at the back focal plane of the collecting lens. The area-array detector B is located in the center of the field of view of the optical system, so it receives the orthographic view light passing through the entrance slit S2. The area-array detector A receives the front view light passing through the entrance slit S1. The area-array detector C receives the back view light passing through the entrance slit S3. In one scan period of the moving CCM, the CDSIS simultaneously records the spectral images of three scan lines on the scene: the detector A records the front view spectral image of one scan line on the scene, the detector B records the orthographic view spectral image of another scan line on the scene, and the detector C records the back view spectral image of another scan line on the scene. When the CDSIS is spatially scanned perpendicular to the entrance slits: the detector A records the front view consecutive spectral image of the scene, the detector B records the orthographic view consecutive spectral image of the scene, and the detector C records the back view consecutive spectral image of the scene. Namely, the CDSIS records three consecutive spectral images of front view, orthographic view and back view of the scene when it is spatially scanned perpendicular to the entrance slits.

 

Fig. 1 Equivalent orthographic view for the optics of the three-area-array coherent-dispersion stereo-imaging spectrometer (CDSIS). ZPD: zero path difference.

Download Full Size | PDF

 

Fig. 2 Equivalent light path diagram of the three-area-array CDSIS. A, B and C: area-array detectors located in the back focal plane of the collecting lens. S1, S2 and S3: entrance slits located at the front focal plane of the collimating lens (back focal plane of the objective lens).

Download Full Size | PDF

The CDSIS combines interferometric spectroscopy with dispersive spectroscopy to get spectral information of an object or a scene. The CDSIS integrates a moving corner-cube-mirror interferometer and a plane transmission grating to obtain spectral information. For each unit of the scene, the CDSIS simultaneously produces multiple interferograms in one scan period of the moving CCM, each interferogram with a separate wavelength range and located in a separate pixel of a given column of one area-array detector. More specifically, for each scene unit, not only the transmission grating spreads the spectral information onto a given column of one area-array detector, but also the interferometer simultaneously generates multiple interferograms in one scan period of the moving CCM, each interferogram covering a separate wavelength range controlled by the transmission grating and detector geometry, each interferogram located in a separate pixel of the given column above. In other words, the transmission grating divides a broad-band spectrum into a number of narrow-band spectra, each narrow-band spectrum is located on a separate pixel of a given column of the detector, and the interferometer simultaneously produces multiple interferograms for the broad-band spectrum in one scan period of the moving CCM, each interferogram corresponds to a different narrow-band spectrum. Thus any noise present in an optical signal is limited to the pixels where this noisy signal impinges on, and these noise signals have no effect on the interferograms with different narrow-band spectra. Moreover, the center fringes are spread into multiple separate pixels (interferograms), each with a separate narrow-band spectrum, thus the dynamic range requirement of the detector is reduced [69,70]. Therefore, the CDSIS has the advantage of providing the high spectral resolution of interferometry, while avoiding the multiplex disadvantage of interferometry for a broadband source in the ultraviolet-visible spectral region, and reducing the dynamic range requirement of the detector. The CDSIS can perform high spectral resolution measurements of a broadband source in the ultraviolet-visible spectral region.

Throughout this paper, rows of each area-array detector are parallel to the y-axis of the detector plane, columns of each area-array detector are parallel to x-axis of the detector plane, ${\lambda }_o$ is the center wavelength of a source spectra covering a wavelength range from ${\lambda }_1$ to ${\lambda }_k$ (i.e., a wavenumber range from ${\sigma }_{\min }=1/{\lambda }_k$ to ${\sigma }_{\max }=1/{\lambda }_1$ if ${\lambda }_1<{\lambda }_k$), $l$ is the displacement of the moving CCM from the zero path difference (ZPD) position, $f$ is the focal length of the objective lens, $f_1$ is the focal length of the collimating lens, and $f_2$ is the focal length of the collecting lens.

In Fig. 1, typical orthographic view light rays emerging from two representative object points on a scan line (in Y-axis direction of the object coordinate system O-XYZ) of the scene are drawn. One point $P_i\left(0,0,Z_i\right)$ is located on the optical axis of the objective lens. The second point $P_j\left(0,Y,Z_j\right)$ is displaced by a distance $Y$ from the optical axis of the objective lens. Moreover, six representative image points are shown: $P_i{}^{\prime \text{}\prime }\left(0,0\right)$ is the image point of object point $P_i\left(0,0,Z_i\right)$ on the back focal plane of the objective lens, $P_j{}^{\prime \text{}\prime }\left(0,y^{\prime }\right)$ is the image point of object point $P_j\left(0,Y,Z_j\right)$ on the back focal plane of the objective lens, $P_{i1}{}^{\prime }\left(x_{B1},0\right)$ is the image point on area-array detector B of $P_i\left(0,0,Z_i\right)$ formed by wavelength ${\lambda }_1$, $P_{ik}{}^{\prime }\left(x_{Bk},0\right)$ is the image point on area-array detector B of $P_i\left(0,0,Z_i\right)$ formed by wavelength ${\lambda }_k$, $P_{j1}{}^{\prime }\left(x_{B1},y\right)$ is the image point on area-array detector B of $P_j\left(0,Y,Z_j\right)$ formed by wavelength ${\lambda }_1$, $P_{jk}{}^{\prime }\left(x_{Bk},y\right)$ is the image point on area-array detector B of $P_j\left(0,Y,Z_j\right)$ formed by wavelength ${\lambda }_k$. It can be obtained that

In Fig. 2, typical front view, orthographic view and back view light rays are received by the area-array detectors A, B and C, respectively. $x_{A1}$, $x_{Ao}$ and $x_{Ak}$ are the x-axis coordinates on the detector plane for the wavelengths ${\lambda }_1$, ${\lambda }_o$ and ${\lambda }_k$ of the front view light. $x_{B1}$, $x_{Bo}$ and $x_{Bk}$ are the x-axis coordinates on the detector plane for the wavelengths ${\lambda }_1$, ${\lambda }_o$ and ${\lambda }_k$ of the orthographic view light. $x_{C1}$, $x_{Co}$ and $x_{Ck}$ are the x-axis coordinates on the detector plane for the wavelengths ${\lambda }_1$, ${\lambda }_o$ and ${\lambda }_k$ of the back view light. ${x^{\prime }}_A$ is the x-axis coordinate of the entrance slit S1 on the back focal plane of the objective lens, ${x^{\prime }}_B$ is the x-axis coordinate of the entrance slit S2 on the back focal plane of the objective lens, and ${x^{\prime }}_C$ is the x-axis coordinate of the entrance slit S3 on the back focal plane of the objective lens. $\psi$ is the angle between the grating normal and the optical axis of the optical system. ${x^{\prime }}_B=0$, ${x^{\prime }}_A-{x^{\prime }}_B={x^{\prime }}_B-{x^{\prime }}_C$, so ${x^{\prime }}_A=-{x^{\prime }}_C$. It can be obtained that

Figure 3 shows the equivalent light path diagram from the transmission grating to the detector plane for the orthographic view light. The grating equation for a plane transmission grating is given by [71]

 

Fig. 3 Equivalent light path diagram from grating to detector for the orthographic view light.

Download Full Size | PDF

Let the optical axis of the collecting lens overlap with the 1-order diffracted light ray from the grating of the center wavelength ${\lambda }_o$ of the orthographic view light, thus the x-axis coordinate on the detector plane of the center wavelength ${\lambda }_o$ of the orthographic view light is $x_{Bo}=0$.

According to the grating equation, the characteristics of the lens and the geometry, it can be obtained that

For the front view light, the incidence angle measured from the grating normal is $\alpha =\psi -\rho$ (see Fig. 2), it can be obtained that

For the back view light, the incidence angle measured from the grating normal is $\alpha =\psi +\rho$ (see Fig. 2), it can be obtained that

The CDSIS is based on three-view stereo imaging to obtain three-dimensional spatial information (i.e., three-dimensional stereoscopic image) of a scene or an object. The CDSIS images each scene unit on a separate column of one area-array detector at each of the three views. More specifically, for each of the three views, each scene unit is imaged on a given column of one area-array detector, and different wavelengths are dispersed across different rows of that column. When the CDSIS is spatially scanned perpendicular to the entrance slits, it records three images of front view, orthographic view and back view of each scene unit. Figure 4 shows the equivalent schematic diagram of the three-view stereo imaging for the CDSIS. The instrument platform (e.g., satellite) flies from left to right along the X-axis direction of the object coordinate system O-XYZ. At time $t_A$, an object point $P_i\left(X,Y,Z\right)$ on a scan line (i.e., green thick line parallel to the Y-axis direction in Fig. 4) of the scene is imaged on the entrance slit S1, and the coordinates of this front view image point are $\left({x^{\prime }}_A,{y^{\prime }}_A\right)$. At time $t_B$, the same object point $P_i\left(X,Y,Z\right)$ is imaged on the entrance slit S2, and the coordinates of this orthographic view image point are $\left({x^{\prime }}_B,{y^{\prime }}_B\right)$. At time $t_C$, the same object point $P_i\left(X,Y,Z\right)$ is imaged on the entrance slit S3, and the coordinates of this back view image point are $\left({x^{\prime }}_C,{y^{\prime }}_C\right)$. With the flight of the instrument platform (e.g., satellite), three consecutive two-dimensional track images of front view, orthographic view and back view of the scene are recorded, respectively, by three area-array detectors (A, B and C). The orthographic view image is used to create a two-dimensional orthophoto image. The front view and back view images are used to reconstruct the three-dimensional stereoscopic image.

 

Fig. 4 Equivalent schematic diagram of the three-view stereo imaging for the CDSIS.

Download Full Size | PDF

_$H$ is the flight height of the instrument platform (e.g., satellite), ${\varphi }_0$ is the angle between the front view axis (back view axis) and the orthographic view axis, and $d$ is the ground distance of the scene corresponding to the front view and the orthographic view (i.e., the instantaneous imaging half width of the scene in the flight direction). It can be obtained that

The relationship between the coordinates of the image point on the back focal plane of the objective lens and the coordinates of the object point on the scene can be expressed by a matrix of image construction, if they are both defined in the object coordinate system O-XYZ. The matrix of image construction can be written as

Solving the matrix of image construction, the collinearity equations can be given by

According to Eq. (28) and Eq. (29) together with Eq. (13), Eq. (15), Eq. (17), Eq. (19), Eq. (21) and Eq. (22), for the front view and back view of an object point $P_i\left(X,Y,Z\right)$ on the scene, we can get four equations and use the least square method to solve three unknowns $X$, $Y$ and $Z$.

Assume that the pixel size of each area-array detector is $b$, and the number of pixels in each row (in y-axis direction) of each area-array detector is $N$. Based on Eqs. (22) and (23), the ground sample distance (GSD) across the flight direction (i.e., the spatial resolution of the CDSIS in Y-axis direction) is given by

The elevation uncertainty in Z-axis direction is given by

According to Eq. (11) and Eq. (12), the size of each column (in x-axis direction) of the area-array detector B must be greater than $\left|x_{B1}-x_{Bk}\right|$, and the number of pixels in each column of the area-array detector B used to acquire multiple interferograms should be

According to Eq. (14) and Eq. (16), the size of each column (in x-axis direction) of the area-array detector A must be greater than $\left|x_{A1}-x_{Ak}\right|$, and the number of pixels in each column of the area-array detector A used to acquire multiple interferograms should be

According to Eq. (18) and Eq. (20), the size of each column (in x-axis direction) of the area-array detector C must be greater than $\left|x_{C1}-x_{Ck}\right|$, and the number of pixels in each column of the area-array detector C used to acquire multiple interferograms should be

The optical path difference (OPD) for each interferogram recorded by area-array detector B and generated by the light emerging from the object point $P_j\left(0,Y,Z_j\right)$ is given by [21,28]

The interferogram $I_B\left(l,y\right)$ recorded by area-array detector B and generated by the light emerging from the object point $P_j\left(0,Y,Z_j\right)$ is given by

The OPD for each interferogram recorded by area-array detector A (or C) and generated by the light emerging from the object point $P_k\left(d,Y,Z_k\right)$ is given by

The interferogram $I_A\left(l,y\right)$ recorded by area-array detector A (or C) and generated by the light emerging from the object point $P_k\left(d,Y,Z_k\right)$ of the scene is given by

The theoretical spectral resolution $\delta {\sigma }_B$ for the orthographic view spectral image of the three-area-array CDSIS can be calculated by

The theoretical spectral resolution $\delta {\sigma }_A$ for the front view (or back view) spectral image of the three-area-array CDSIS can be calculated by

In practice, the designed spectral resolution for the front view (or back view) spectral image of the CDSIS can be given by

For a source spectra covering a wavelength range from ${\lambda }_1$ to ${\lambda }_k$, if ${\lambda }_1<{\lambda }_k$, the maximum wave number of the source spectra is ${\sigma }_{\max }=1/{\lambda }_1$. According to Nyquist criterion, for convenience, the sampling interval for each interferogram produced by the CDSIS can be $\chi \le 1/2{\sigma }_{\max }\text{=}{\lambda }_1/2$. The number of sampling points for each unilateral interferogram recorded by the area-array detector B and generated by the light emerging from the object point $P_j\left(0,Y,Z_j\right)$ can be given by

Table 1 shows the comparisons of the three-area-array CDSIS, the coherent-dispersion imaging spectrometer (CDIS) in [50] and the coherent-dispersion spectrometer (CDS) in [59]. Table 2 shows the comparisons of the CDSIS, the interferometric imaging spectrometer, the dispersive imaging spectrometer, the color filter imaging spectrometer, and the snapshot spectral imager. Table 3 shows the comparisons of the CDSIS, the broadband snapshot imaging spectrometer (BSIS) in [46], the infrared broadband snapshot imaging spectrometer (IBSIS) in [47], and the orthogonal-dispersion device in [72].

Table 1. Comparisons of the CDSIS, the CDIS in [50], the CDS in [59], and the UVS in [60]

View Table | View all tables in this article

Table 2. Comparisons of the CDSIS, interferometric, dispersive, and color filter imaging spectrometers

View Table | View all tables in this article

Table 3. Comparisons of the CDSIS, the BSIS in [46], the IBSIS in [47], and the device in [72]

View Table | View all tables in this article

Both the Andor iDus CCD and Newton CCD spectroscopy detectors of Oxford Instruments are suitable for the CDSIS. For example, the Andor iDus 420 Series 1024 x 255 pixel Spectroscopy CCD, the Andor Newton 920 Series 1024 x 255 pixel Spectroscopy CCD, the Andor Newton 940 Series 2048 x 512 pixel Spectroscopy CCD, and so on [73,74].

By dispersing the light with a plane transmission grating instead of a prism, mainly because the grating can provide higher and more linear dispersion of the wavelength than the prism [60]. However, the prism possesses a wider spectral range than the grating [75].

3. Preliminary theoretical calculation and numerical simulation

Assume that the source spectrum covers a wavelength range from 400 nm to 700 nm, i.e., a wavenumber range from 14285.7 cm⁻¹ to 25000 cm⁻¹. Table 4 shows the wavelength difference versus wavenumber difference for several wavelengths. If the desired spectral resolution of the CDSIS is 0.05 nm at 450 nm together with 0.1 nm at 700 nm, the designed spectral resolution (in wavenumber) of the CDSIS should be $\delta {\sigma }^{\prime }=2{\text{cm}}^{-1}$. From Eq. (45), the maximum displacement of the moving CCM from the ZPD position should be $l_{\max }=1/\left(2\delta {{\sigma }^{\prime }}_B\right)=2.5mm$. Based on Eq. (48), the number of sampling points for each unilateral interferogram recorded by the detector B and generated by the light emerging from the object point $P_i\left(0,0,Z_i\right)$ is $K_B\left(0\right)=4{\sigma }_{\max }{l}_{\max }=4\times \text{25000}c{m}^{-1}\times 2.5mm=\text{25000}$.

Table 4. Wavelength difference versus Wavenumber difference for several wavelengths

View Table | View all tables in this article

Suppose that the flight height of the instrument platform (e.g., satellite) is $H=200km$, the instantaneous imaging half width in the flight direction is $d=60km$, the ground imaging width across the flight direction is $W=120km$, the ground sample distance is $\text{GSD}=120m$, the pixel size of each area-array detector is $b=0.02mm$, the focal length of the objective lens is $f=50mm$, the focal length of the collecting lens is $f_2=100mm$. Thus the focal length of the collimating lens is $f_1=150mm$, the angle between the front view axis (back view axis) and the orthographic view axis is ${\varphi }_0\approx 16.7°$, the field of view across the flight direction is ${\text{FOV}}_Y\approx 16.7°$, the number of pixels in each row (in y-axis direction) of each area-array detector is $N\text{=}W/\text{GSD}\text{=1000}$, and the length of the entrance slit is $L_S=30mm$.

Fused silica is a proper material for the plane transmission grating, and its refractive index formula is given by [76]

Assume that the angle between the grating normal and the optical axis of the optical system is $\psi =10°$, the center wavelength is ${\lambda }_o\text{=550}nm$, and the transmission grating with 150 grooves/mm. According to Eq. (4), Eqs. (11)-(21) and Eq. (50), the x-axis coordinates on the detector plane of the wavelengths from ${\lambda }_1=400nm$ to ${\lambda }_k=700nm$ for front view light, orthographic view light and back view light are shown in Fig. 5. The number of pixels in each column of area-array detector B is $M_B\ge \left|x_{B1}-x_{Bk}\right|/b=\left|2.3136-\left(-2.3009\right)\right|/0.02=230.725$. Namely, for each scene unit, $M_B=231$ separate interferograms are simultaneously generated and recorded by a separate column of the detector B. The number of pixels in each column of area-array detector A should be $M_A\ge \left|x_{A1}-x_{Ak}\right|/b=\left|-15.1632-\left(-19.8078\right)\right|/0.02=232.23$. The number of pixels in each column (in x-axis direction) of the area-array detector C should be $M_C\ge \left|x_{C1}-x_{Ck}\right|/b\text{=}\left|\text{20}\text{.4176}-\text{15}\text{.3769}\right|/0.02=252.035$.

 

Fig. 5 The x-axis coordinates on the detector plane of different wavelengths for the front view light, orthographic view light and back view light.

Download Full Size | PDF

According to Eqs. (11)-(13) and Eq. (41), three representative interferograms recorded simultaneously by a separate column of area-array detector B in one scan period of the moving CCM and generated by the light emerging from object point $P_i\left(0,0,Z_i\right)$ are shown in Fig. 6. The first representative interferogram contains only wavelength 449.95 nm, 450 nm, 450.05 nm and 450.1 nm. The second representative interferogram contains only wavelength 549.93 nm, 550 nm, 550.07 nm and 550.14 nm. The third representative interferogram contains only wavelength 699.7 nm, 699.8 nm, 699.9 nm and 700 nm. Figure 7 shows the spectrum obtained from Fourier transform of the three interferograms shown in Fig. 6 and its distribution in a given column of area-array detector B.

 

Fig. 6 Several interferograms recorded simultaneously by area-array detector B in one scan period of the moving CCM for object point $P_i\left(0,0,Z_i\right)$ when the spectral resolution is 2 cm⁻¹.

Download Full Size | PDF

 

Fig. 7 Spectrum obtained from Fourier transform of the three interferograms in Fig. 6 and its distribution in a given column of area-array detector B.

Download Full Size | PDF

Figure 8 shows the detailed spectrum obtained from Fourier transform of the three interferograms shown in Fig. 6. In order to better show the detailed spectrum, there are three pictures in Fig. 8: the top picture is obtained from Fourier transform of the first interferogram (in Fig. 6) that contains only wavelength 449.95 nm, 450 nm, 450.05 nm and 450.1 nm; the middle picture is obtained from Fourier transform of the second interferogram (in Fig. 6) that contains only wavelength 549.93 nm, 550 nm, 550.07 nm and 550.14 nm; and the bottom picture is obtained from Fourier transform of the third interferogram (in Fig. 6) that contains only wavelength 699.7 nm, 699.8 nm, 699.9 nm and 700 nm. In the top picture of Fig. 8, from left to right, the four peaks correspond to the spectra of wavelengths 449.95 nm, 450 nm, 450.05 nm and 450.1 nm, respectively. In the middle picture of Fig. 8, from left to right, the four peaks correspond to the spectra of wavelengths 549.93 nm, 550 nm, 550.07 nm and 550.14 nm, respectively. In the bottom picture of Fig. 8, from left to right, the four peaks correspond to the spectra of wavelengths 699.7 nm, 699.8 nm, 699.9 nm and 700 nm, respectively. Twelve spectral peaks are clearly visible. It can be easily obtained that the spectral resolution of the CDSIS for the orthographic view spectral image is higher than 0.05 nm at 450 nm and 0.07 nm at 550 nm together with 0.1 nm at 700 nm. Figures 6-8 are the simulation results by MATLAB software.

 

Fig. 8 The detailed Spectrum obtained from Fourier transform of the three interferograms in Fig. 6.

Download Full Size | PDF

4. Conclusion

A three-area-array coherent-dispersion stereo-imaging spectrometer (CDSIS) was presented and preliminary theoretical calculations were given. The CDSIS records three consecutive spectral images of the front view, orthographic view and back view of the scene when it is spatially scanned perpendicular to the entrance slits. The orthographic view image is used to create a two-dimensional orthophoto image. The front view and back view images are used to reconstruct the three-dimensional stereoscopic image. Compared with all existing imaging spectrometers that only acquire two-dimensional spatial information and one-dimensional spectral information, the most prominent and important advantage of the CDSIS is that the CDSIS can acquire both three-dimensional spatial information and one-dimensional spectral information. The CDSIS has another important advantage of providing the high spectral resolution of interferometry, while avoiding the multiplex disadvantage of interferometry for a broadband source in the ultraviolet-visible spectral region. The third advantage of the CDSIS, compared with traditional interferometry, is that the dynamic range requirement of the area-array detector is reduced. There is also a tradeoff for the CDSIS, which mainly includes three aspects: (1) the spatial-time joint scan makes the data collection time long, (2) the presence of a moving part will reduce the stability against various disturbances, and (3) the presence of the entrance slit means the instrument does not benefit from the throughput advantage. The CDSIS is a unique concept that not only acquires three-dimensional spatial information and one-dimensional spectral information, but also performs high spectral resolution measurements of a broadband source in the ultraviolet-visible spectral region. The CDSIS will be suitable for hyperspectral earth remote sensing or space exploration.