APPARATUS AND METHODS FOR REMOTE MEASUREMENT OF SEA SURFACE TEMPERATURE

Abstract

Apparatus and methods are disclosed for highly accurate remote measurement of sea surface skin temperature. Thermal band 8 to 14 micron images of the surface of the ocean taken by a downward looking infrared camera are processed to determine the optimum segments of the image to utilize. The influence of contaminating reflection of the downwelling flux from the sky and other error sources are removed and from the data and/or otherwise corrected for making sea surface temperature accuracy within several tenths of a degree possible.

Claims

1. A method for remote measurement of sea surface temperature comprising the steps of: capturing a multi-pixel image in selected wavebands of a small area of sea surface at a selected incidence angle, each pixel of the image having a fraction of a degree field-of-view width; sorting blackbody thermal values of the pixels, from warmest to coolest; locating coordinates of the warmest pixels relative to position of image capture and nadir angle thereto; determining incidence and reflected angles from a selected warmest pixel; and determining downwelling infrared flux in the wavebands and at the incidence angle and utilizing determination of downwelling infrared flux to correct calculation of true sea surface temperature.

2. The method of claim 1 wherein the step of capturing a multi-pixel image includes positioning an infrared camera capable of image capture in the 8 to 14 micron thermal band above a sea surface.

3. The method of claim 2 further comprising: pointing the camera to a desired azimuth and elevation angle to the sea surface responsive to a processor; receiving GPS/GNSS, surface meteorological, point, azimuth and elevation angle data at the processor; receiving pitch roll and location of geometric horizon data at the processor; utilizing a processing algorithm at the processor to extract contaminated pixels; and repeatedly comparing a time series of calculated true sea surface temperatures to enhance system performance.

4. The method of claim 1 further comprising the step of processing the pixels to determine optimum segments of the image to utilize and correcting for influence of contaminating reflection of the downwelling flux from the sky.

5. The method of claim 1 further comprising the steps of: capturing a time series of pitch and roll angles from an inertial measurement unit; fitting sinusoids to the time series of pitch and roll angles and determining mean angles therefrom; applying Fourier transform to the time series of pitch and roll angles and inspecting the dc component of the Fourier transform; comparing and assessing sinusoid fit to Fourier transform results; and determining geometric horizon from the mean angles.

6. The method of claim 5 further comprising the steps of: mapping sky blackbody temperature to a selected high elevation angle; selecting a method of measuring sky blackbody temperature at the selected elevated angle; determining angle of the warmest pixel relative to the geometric horizon; determining or estimating Mean Radiating Temperature of the atmosphere; calculating atmospheric opacity; and mapping opacity to the desired high angle and using opacity and brightness temperature processing to determine brightness at the selected elevation angle.

7. The method of 5 further comprising determining a desired high angle downwelling flux from lower elevation angle observation.

8. A method for highly accurate measurement of physical temperature of sea surface skin comprising the steps of: calibrating a passive noncontact infrared thermal 8 to 14 micron band multi-pixel camera capable of capturing a plurality of pixels in each captured image of its field of view to express thermal blackbody temperature of each of the pixels; directing the field of view of the camera to the sea surface at a known incidence angle and capturing an image or images thereof and sending data indicative thereof to a processor; utilizing angle of a viewed wave surface in each pixel, view angle of the of the infrared camera onto the wave surface in each pixel, and angle of the reflected view of the sky in each pixel to determine the flux emitted from the sea surface and reflected sky flux in each pixel at the processor; and obtaining corrected sea surface skin temperature at the processor by separating the values of emitted flux from the first several microns of the sea surface skin and values of variable flux from sky reflected from the wave surface angles in the field of view of the camera in a selected pixel and processing accordingly.

9. The method of claim 8 further comprising controlling camera view position using real time data from an inertial measurement unit and a camera positioning unit responsive thereto received at the processor.

10. The method of claim 8 wherein determinations of sea surface skin temperature accurate to about 0.3 C are enabled.

11. The method of claim 8 wherein the camera field of view pixel width is a fraction of a degree thus spanning only a few millimeters of wave surface.

12. The method of claim 8 wherein the known incidence angle is between about an orthogonal value of 0.008 degrees to 20 degrees of normal to the wave surface.

13. The method of claim 9 further comprising the steps of receiving GPS/GNSS, surface meteorological, point, azimuth, elevation angle and Solar ephemeris data at the processor, and utilizing data received at the processor to extract contaminated pixels.

14. Apparatus for remote measurement of sea surface temperature from a selected viewing platform comprising: a multi-pixel infrared camera mounted in a hermetic enclosure; a signal storage and processing unit mounted in said enclosure and in communication with said camera; an inertial measurement unit mounted in said enclosure and in communication with said storage and processing unit and connected for receipt of GNSS position and velocity and determination of platform movement; a surface air temperature and relative humidity probe mounted with said enclosure and in communication with said storage and processing unit; and a pan-tilt positioner connectable to said enclosure and in communication with and operationally responsive to said storage and processing unit.

15. The apparatus of claim 14 where said camera is an 8 to 14 micron band non-contact thermal infrared camera with a 25 degree filed of view calibrated to express thermal blackbody temperature of each pixel in said field of view.

16. The apparatus of claim 15 wherein width of each pixel is a fraction of a degree.

17. The apparatus of claim 14 further comprising a germanium window mounted in an open of said enclosure at a location so that said camera views therethrough.

18. The apparatus of claim 14 wherein said storage and processing unit includes programming that performs system housekeeping functions, controls and performs operational process steps of sea surface temperature measurement, captures data from said camera, said probe, and said pan-tilt positioner, determines geometric horizon, controls said positioner operations, and communicates data via output ports.

19. The apparatus of claim 14 further comprising means for accommodating measurement of zenith downwelling blackbody temperature at said enclosure including one of a neutral density filter and blackbody mask of known temperature, losses and reflections.

20. The apparatus of claim 14 further comprising a GNSS antenna port at said enclosure, a breathing dissicator mount at said enclosure, a Schrader valve port for pressure testing at said enclosure, an onboard power supply in said enclosure and a visible band camera mounted in said enclosure adjacent a view port therefor and communicating with said storage and processing unit.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The accompanying drawings illustrate a complete embodiment of the invention according to the best mode so far devised for the practical application of the principles thereof, and in which:

[0029] FIG. 1 is a long wave infrared (LWIR) image of the ocean taken from the gunwale of a vessel with an infrared camera;

[0030] FIG. 2 contains LWIR infrared camera sea surface temperature measurements in a 1.21.2 degree FOV area (ISARFOV) at the bottom center of the image in FIG. 1;

[0031] FIG. 3 illustrates the apparatus of this invention mounted on a ship's bow and pointed seaward in front of the passage of the ship.

[0032] FIG. 4 illustrates selected data indicative of reflections from sea water for orthogonal polarizations of incident and reflected rays and the average thereof;

[0033] FIG. 5 illustrates data of Fresnel calculations of electromagnetic flux reflections and transmissions from a flat sea surface as a function of incidence angle;

[0034] FIG. 6 is a plot of the sensitivity of seawater reflection coefficient to errors in measurement of incidence angle;

[0035] FIG. 7 is a laboratory measure of index of refraction in the thermal LWIR band;

[0036] FIG. 8 is a mapping of the sensitivity and losses of the optical system against the index of refraction across the LWIR band;

[0037] FIG. 9 is a plot of the fractional sensitivity of seawater reflection coefficient R to sea water index of refraction for a 0.01 misestimation of seawater index of refraction n;

[0038] FIG. 10 is an illustration of the apparatus of this invention pointed to an elevated angle above the horizon;

[0039] FIG. 11 depicts the infrared camera system pointed to a near zenith viewing through a neutral density filter or, alternatively, a blackbody mask;

[0040] FIG. 12 is a transmission plot of a neutral density germanium filter;

[0041] FIG. 13 is an image of blackbody mask of about 35% transmission for use with this invention;

[0042] FIG. 14 is an image of the sea surface in the LWIR showing reflections of the Sun;

[0043] FIG. 15 is a flowchart illustrating steps in operational determination of sea surface temperature of this invention;

[0044] FIG. 16 is a flowchart illustrating steps to determining the position of the geometric horizon and the local gravitational vector by applying processing to data from an inertial measurement system associated with the apparatus of this invention;

[0045] FIG. 17 is a flowchart illustrating steps for mapping low angle infrared camera measured sky brightness temperatures to higher elevation angles;

[0046] FIG. 18 is a flowchart demonstrating steps for selection of pixel(s) within the infrared camera image for processing into sea surface temperature;

[0047] FIG. 19 is an exploded view of a preferred embodiment of the apparatus of this invention;

[0048] FIG. 20 is a perspective view of the preferred embodiment of the apparatus of this invention shown in FIG. 19 with the cover removed;

[0049] FIG. 21 is a second perspective view of the preferred embodiment of the apparatus of this invention shown in FIG. 19 without the cover;

[0050] FIG. 22 is a side view of the preferred embodiment of the apparatus of this invention shown in FIG. 19; and

[0051] FIG. 23 is an end view of the preferred embodiment of the apparatus of this invention shown in FIG. 19.

DESCRIPTION OF THE INVENTION

[0052] FIG. 3 depicts the sea-viewing measurement apparatus of this invention. Infrared camera system 31 is preferably located on a suitable location of a sea going vessel or structure. In use on ships, the bow area is preferred location to observe the sea surface undisturbed by the motion of the ship that stirs and mixes the water. Camera 33 is pointed downward to sea surface 35 to capture an image of the surface. The camera pixels are calibrated in equivalent blackbody temperature of the received flux. This flux includes a small contribution from reflected flux 37 from downwelling sky radiation 39. The pixels in the image originate from varying slopes of the waves 41 on the sea surface, and contain differing amounts of upwelling emission from the sea 43 and the reflected sky flux 37 contribution, as defined by Fresnel's Equations.

[0053] FIG. 4 illustrates data indicative of reflections from sea water for orthogonal polarizations in the plane of (45) and perpendicular to the plane of (47) the incident and reflected rays, and the average thereof (49). The X-axis represents the incidence angle while the Y-axis is the reflectance. The average reflection coefficient is constant from orthogonal out to 20 degrees of incidence angle.

[0054] Fresnel's Equations define the magnitude of the reflection and transmission of the vertical and horizontally polarized fluxes from and across the sea surface as a function of incidence angle. The observed thermal emission from the sea is dependent upon the viewing angle, and the downwelling flux from the sky is dependent upon the incident and reflected angles of this flux as defined by Fresnel's Equations for amplitude of the electromagnetic propagation (see FIG. 5 showing Fresnel calculation of fluxes (mapped sky temperature at 51, emission at 53, reflection at 55 and the total thereof at 57 in degrees K at the camera Y-axis) from a flat sea as a function of incidence angle (X-axis) for 100K zenith sky and seawater n=1.186, wherein the increased reflection and decreased emission with increasing incidence angle work in concert to yield colder images of the sea as the FOV goes from nadir to horizonsee FIG. 1, left side):

[00001]$R_{}\left(\right)={\left(\frac{n_1.Math.\cos .Math..Math.{}_i-n_2.Math.\cos .Math..Math.{}_t}{n_1.Math.\cos .Math..Math.{}_i+n_2.Math.\cos .Math..Math.{}_t}\right)}^2={\left[\frac{n_1.Math.\cos .Math..Math.{}_i-n_2.Math.\sqrt{1-{\left(\frac{n_1}{n_2}.Math.\sin .Math..Math.{}_i\right)}^2}}{n_1.Math.\cos .Math..Math.{}_i+n_2.Math.\sqrt{1-{\left(\frac{n_1}{n_2}.Math.\sin .Math..Math.{}_i\right)}^2}}\right]}^2={\left(\frac{\sin .Math..Math.\left({}_i-{}_t\right)}{\sin .Math..Math.\left({}_i+{}_t\right)}\right)}^2.Math..Math.R_{.Math.}\left(\right)={\left(\frac{n_1.Math.\cos .Math..Math.{}_t-n_2.Math.\cos .Math..Math.{}_i}{n_1.Math.\cos .Math..Math.{}_t+n_2.Math.\cos .Math..Math.{}_i}\right)}^2={\left[\frac{n_1.Math.\sqrt{1-{\left(\frac{n_1}{n_2}.Math.\sin .Math..Math.{}_i\right)}^2}-n_2.Math.\cos .Math..Math.{}_i}{n_2.Math.\sqrt{1-{\left(\frac{n_1}{n_2}.Math.\sin .Math..Math.{}_i\right)}^2}+n_2.Math..Math.\cos .Math.{}_i}\right]}^2={\left(\frac{\tan .Math..Math.\left({}_i-{}_t\right)}{\tan .Math..Math.\left({}_i+{}_t\right)}\right)}^2$$.Math.T_{}\left(\right)={\left(1-s.Math.q.Math.r.Math.t.Math.R_{}\left(\right)\right)}^2=\frac{4.Math.n_1.Math.n_2.Math.\cos .Math..Math.{}_i.Math.\cos .Math..Math.{}_t}{{\left(n_1.Math.\cos .Math..Math.{}_i+n_2.Math.\cos .Math..Math.{}_t\right)}^2}=\frac{2.Math.\sin .Math..Math.{}_t.Math.\cos .Math..Math.{}_i}{\sin .Math..Math.\left({}_i+{}_t\right)}$$.Math.T_{.Math.}\left(\right)=\frac{4.Math.n_1.Math.n_2.Math..Math.\cos .Math.{}_i.Math.\cos .Math..Math.{}_t}{{\left(n_1.Math.\cos .Math..Math.{}_t+n_2.Math.\cos .Math..Math.{}_i\right)}^2}=\frac{2.Math.\sin .Math..Math.{}_t.Math.\cos .Math..Math.{}_i}{\sin .Math..Math.\left({}_i+{}_t\right).Math.\cos .Math..Math.\left({}_i-{}_t\right)}$$.Math.{}_t={\sin }^{-1}\left(\frac{n_1.Math.\sin .Math..Math.{}_i}{n_2}\right)$

[0055] The sea slope and the look angle of the infrared camera must be determined from the incidence angle of the wave location of the warmest pixel in the infrared camera image and the camera system. Any thermal band narrow pixel infrared camera can be utilized, provided it accurately measures equivalent blackbody temperatures and observes with a very narrow FOV. Since a high degree of accuracy of sea surface temperature measurement is desired (0.3 C or better), a number of parameter uncertainties and measurement errors need to be addressed.

[0056] Accurate knowledge of the above-referenced incidence angles is somewhat relaxed if incidence angles are small (near orthogonal). FIG. 4 shows the amplitude of reflection from the surface from sea water as a function of zenith angle. At orthogonal incidence, about 0.008 of the flux is reflected. The horizontal and vertical polarizations of the electric vector in the electromagnetic flux differ in reflection values as the zenith angle is increased. The average of these two polarizations is within 1% of the orthogonal value out to about 20 degrees of zenith angle. Most observations utilized for the sea surface temperature determination are within this 20 degree angle.

[0057] The blackbody temperatures seen by a LWIR camera at incidence angles from normal to grazing for an assumed 100K sky and 300K sea is shown in FIG. 5. Also shown are the contributing sky reflected (55) and sea emitted (53) fluxes to this camera measurement.

[0058] The error in brightness temperature induced by a 10 degree misestimation in incidence angles induces about a 0.04 C error in calculation in the reflected downward flux estimation, which when combined with a 0.1 C infrared camera error, results in a total error of 0.108 C, a small contribution. FIG. 6 presents the error induced by errors in determination of incidence angles of the downwelling sky flux and camera observation angle (X-axis is the incidence angle and the Y-axis is the reflection coefficient 1/R dr/d(theta) wherein 59 n=1.2 and 61n=1.3).

[0059] It is required that the near-zenith downwelling flux, the blackbody temperature in the 8 to 14 micron window, be determined or estimated. Typical near-zenith temperatures might be as cold as 50 Kelvins, 223 degrees Celsius, well below the range of most all infrared cameras and noncontact thermometers. This temperature could be calculated from radiosonde or numeric weather model profiles of temperature and water vapor, and the flux calculated utilizing radiative transfer software.

[0060] Additionally, the sky flux is a small contribution to the infrared camera measured combined sky and sea surface temperature fluxes. Thus, some uncertainty is allowed in the angular measurements without significantly affecting the sea surface temperature determination. As seen in TABLE 2, the error in sea surface temperature determination is dominated by the uncertainty in the infrared camera measurement, about 0.3 degrees Celsius for the ICI 8640 camera of the preferred embodiment.

[0061] As seen in FIG. 7, the index of refraction of sea water varies across the 8 to 14 micron (wave number 12.5 to 7 cm.sup.1) band from n=1.13 to =1.31. This variability must be mapped against the response of the infrared camera system.

[0062] FIG. 8 shows the mapping of the band response (X-axis in microns) of an ICI 8640 infrared camera (63) viewing through an ICI germanium window (window transmission at 65) against the variable index of refraction (Y-axis 8640 mapped n left and window transmission n right). The n-seawater is at 66, mapped n is at 67 and window transmission is at 68, The effective seawater index of refraction is calculated to be n=1.186. Misestimation in this effective index of refraction results in errors shown by FIG. 9 wherein the X-axis is incidence angle theta and the Y-axis is index of refraction 1/R (d R/d n) per 0.01 change in n thereby illustrating fractional dependence of amplitude of reflection coefficient upon a 0.01 change in index of refraction as a function of incidence angle.

[0063] FIG. 10 depicts use of infrared camera system 31 of the apparatus of this invention in a method for determining the downwelling sky flux that is required to correct out the reflected flux viewed by infrared camera 33. This method consists of measuring the infrared sky temperature at an elevation angle where the sky temperature is in measurement range of the infrared camera used (an IRT (infrared thermometer) could optionally be utilized), and then mapping this sky temperature to a zenith or desired near zenith value. Sensitivity to possible sources of error are explored, as care must be taken in the observations and calculations to avoid inducing error. The infrared camera system is pointed to an elevated angle above the horizon such that the cold sky temperature is on the lower end of the measurement scale of the infrared camera. This then gives a method of calculating the cold sky temperature at high elevation angles.

[0064] A stratified atmosphere is assumed. The opacity at the higher elevated angle is calculated from the brightness measurement at a lower elevation angle, mapped to the desired high elevation angle, and the high angle temperature calculated from the high angle opacity. The derivation below demonstrates how the downwelling emission can be mapped from one elevation angle to another, and assesses the errors that might be incurred.

[0065] The radiation intensity or Brightness Temperature observed at an IRT or infrared camera (IRC) from atmospheric radiation can be expressed using a modified form of the integral form of Chandrasekhar's Radiative Transfer Equation (RTE) for radiative propagation through an interactive medium:

[00002]$\left(\mathrm{Chandrasekha}.Math.r\right)$$T_B={}_{\mathrm{IRC}}^{}.Math.T\left(s\right).Math.\left(s\right).Math.e^{\left(-{}_{\mathrm{IRC}}^s.Math.\left(s^{}\right).Math.{\mathrm{ds}}^{}\right)}.Math.\mathrm{ds}+T_{\mathrm{cosmic}}.Math.e^{\left(-{}_{\mathrm{IRC}}^{}.Math.\left(s\right).Math.\mathrm{ds}\right)}$

[0066] where

[0067] T.sub.B is the Brightness Temperature, the Blackbody Temperature,

[0068] IRC is the infrared camera position,

[0069] T.sub.cosmic is the origin of emission from the Big Bang,

[0070] T(s) is the atmospheric temperature at distance s from the IRC and

[0071] (s) is the atmospheric absorption at distance s from the IRC.

[0072] Optical depth or opacity tau is defined by the integral in the exponents in the above expression:

=.sub.IR Camera.sup.(s)ds(optical depth or opacity)

[0073] We can linearize the above Fredholm equation by defining the Mean Radiating Temperature T.sub.MR of the intervening atmosphere from the Mean Value Theorem of Calculus and continuous or stepwise numerical integration. The Mean Radiating Temperature T.sub.MR is defined as:

[00003]$\left(\mathrm{Mean}.Math..Math.\mathrm{Radiating}.Math..Math.\mathrm{Temperature}\right)$$T_{\mathrm{mr}}=\frac{{}_0^{}.Math.\left(s\right).Math.T.Math..Math.s}{{}_0^{}.Math.\left(s\right).Math..Math.s}=\frac{{}_{i=0}^{}.Math.{}_i.Math.T_i.Math..Math.s_i}{{}_{i=0}^{}.Math.{}_i.Math..Math.s_i}$

[0074] Thus:

T.sub.B=(1e.sup.)T.sub.mr+T.sub.cbe.sup.(Linearized RTE)

[0075] where T.sub.B is the temperature observed by the IRC. Opacity is then:

[00004]$\left(\mathrm{opacity}\right)$$=\ln .Math..Math.\left(\frac{T_{m.Math.r}-T_{\mathrm{cosmic}}}{T_{m.Math.r}-T_B}\right)$

[0076] Opacity scales with air masses, that is, .sub.zenith=.sub.angle cos where theta is the observation incidence angle measured from zenith. Thus:

T.sub.B(zenith)=(1e.sup..sup..sup.cos )T.sub.mr+T.sub.cbe.sup..sup..sup.cos (zenith Brightness)

[0077] As can be seen from the below differentiation of the above equation, T.sub.MR(Mean Radiating Temperature) plays an important role, especially when brightness temperature T.sub.B is close to T.sub.MR. On the other hand, the reflected sky flux is small, and some level of uncertainty can be tolerated.

[00005]$d.Math.T_B=\left(-T_{c.Math.o.Math.s.Math.m.Math.i.Math.c}.Math.e^{-r}+e^{-r}.Math.T_{M.Math.R}\right).Math.\frac{\left(T_{\mathrm{cosmic}}-T_B\right).Math.T_{M.Math.R}}{\left(T_{\mathrm{mr}}-T_B\right).Math.\left(T_{m.Math.r}-T_{\mathrm{cosmic}}\right)}.Math.d.Math.T_{M.Math.R}$

[0078] This Mean Radiating Temperature can be estimated on a monthly or seasonal basis for a given observation site, and is about 95% of the mean surface temperature.

[0079] Counterintuitively, the value of the Mean Radiating Temperature does not significantly change with elevation angle. Calculations of T.sub.MR utilizing NASA JPL's Radiative Transfer Code to process radiosonde soundings at widely disparate angles and several atmospheric opacities tau demonstrates that T.sub.MR differs by less than several degrees. This variation with elevation angle usually results in less than 0.01 C in the value reflected from the sea surface.

[0080] Another means of measuring the zenith blackbody temperature is to insert an infrared window 71 in the optical path of observation of downwelling flux 72 for neutral density filter 73 or blackbody mask 75 of known temperature and losses and reflections as shown by FIG. 11. Herein is illustrated infrared camera system 31 with camera 33 pointed to a near zenith, viewing through neutral density filter 73 or, alternatively, a blackbody perforated mask 75 at window 71. This gives a method of directly measuring the cold sky temperature at high elevation angles whereby the transmitted or grid emitted flux 77 of the downwelling sky flux 79 is determinable. Reflected sky flux and upwelling flux 81 and filter absorbed flux 83 and emitted flux 85 are thus sorted. This increases the value of the combined flux of the cold sky and the masks, thereby bringing the flux temperature within the range of the camera, and can allow determination of blackbody sky temperatures well below the lower range of the camera. FIG. 12 is a transmission plot of a germanium neutral density filter 73 appropriate to the tasks of this invention, while FIG. 13 depicts a blackbody mask 75 of 35% transmission.

[0081] To identify the pixel in the infrared image that is least contaminated by the sky flux and is dominated by sea surface emission, and therefore from which it is relatively easy to determine the downwelling flux at the desired angle, the blackbody temperatures of the pixels in the image are sorted from warmest to coldest. Quality control algorithms are applied to eliminate pixels containing a reflection of the sun and containing sea foam or biological or other contaminants. Sun glint can occur in some pixels in the IR images of the sea under certain sun angles, camera FOV, ephemerides, and sea slopes. This glint will produce anonymously high temperatures in those IR camera pixels. The Sun subtends about angular degree, and the glint can therefore occur over several adjacent 0.03 degree pixels. The Sun's photosphere is a 5800K blackbody in the visible region, and about 5100K at 10 microns. These brightnesses are greatly decreased by atmospheric absorption. The value in the vicinity of 10 microns is greatly reduced by the amount of total integrated atmospheric water vapor.

[0082] FIG. 14 is an LWIR image shot from the work deck of an ocean going vessel taken at near nadir showing sun glint off of the sea surface where the incidence angle between the sun and the pixel in the infrared image is bisected by the normal to the wave surface (using a FLIR T400 camera, with Sun and camera at approximately the same incidence angle; the bright spots, roughly 4 C warmer than the rest of the sea surface image, presumably being reflections of the Sun). The 1% reflected brightness could be many Kelvins, depending upon atmospheric attenuation of the Sun's sea level LWIR flux, as compared with several K for the clear sky. A quality control (QC) algorithm that searches these anomalies, determines the time-transient nature in anomalous pixels in the time series of the image, and compares temperatures in adjacent pixels will be used to filter out such pixels as discussed below. A histogram may also identify the solar reflections, as the reflected solar images are several degrees above the bulk of the pixel temperatures. The solar ephemeris and angular relations of the FOV and azimuthal pointing of the LWIR camera might be utilized in an algorithm to aid in QC to remove or avoid solar specularity. An on-board inertial measurement system (IMU) of the profiling system can furnish compass heading, latitude and longitude, GMT, and camera point vector. The remaining warmest pixel is least contaminated under skies colder than the sea surface temperature.

[0083] FIGS. 15 through 18 are flowcharts indicative of operational steps embodying the methods of this invention discussed above, including programming of the onboard processor and controllers discussed hereinbelow, necessary for accurate sea surface temperature retrieval. FIG. 15 describes a preferred embodiment of flow of processes utilized for the infrared sea surface temperature measurement system of this invention. The FOV of the infrared camera is captured as thermal values within the pixels in digital form. An infrared camera frame rate up to 9 Hz can be selected via onboard processor.

[0084] Information on the pitch and roll of the ship from an external source, for instance from the ship's data bus, or on-board from a self-contained IMU allows the positioner to be dynamically controlled to maintain a constant view angle with the ocean's surface. Due to the roll, pitch, and heave of the ship, the IMU is not in an inertial frame of reference system, and to determine the true gravitational vector and geometric horizon, the accelerations of the ship need to be subtracted out through long term averaging of the sinusoidal motions of the ship, fitting sinusoids to the accelerations, Fourier analysis of the accelerations, filtering such as Kalman, and other methods.

[0085] More particularly, operational retrieval of true sea surface temperature begins by pointing the camera positioner utilized to the desired azimuth and elevation angle to sea surface as commanded by the onboard processor from the procedure.dat control file of the embedded computer. Infrared image(s) of the oceanic surface are captured, the images ingested into (received by) the processor. GPS/GNSS, surface meteorological and positioner (point and azimuth) data from on board instruments are collected, and pitch roll and location of geometric horizon data (from FIG. 16) are calculated/received at the processor. Sorting of blackbody thermal values of pixels, from warmest to coolest, is applied and the sort extraction algorithm is utilized at the processor to ensure no contamination, including excision of sun glint and sea foam/biomass contaminated pixels. Coordinates of warmest pixel(s) in the image relative to IR Camera and nadir angle thereto are thus located, and incidence and reflected angles determined from the selected warmest pixel(s). Downwelling IR flux in the waveband of the IR camera and at the requisite incidence angle (from FIG. 18) is determined and true sea surface temperature then calculated. A time series of calculated true sea surface temperatures is repetitively compared.

[0086] The flowchart of FIG. 16 shows processes for determination the geometric horizon at the processor. As shown the steps in this process are applied for determining the inertial frame of reference. It also allows observation of the known vicinity of the oceanic horizon, thus enabling determination of refractivity, temperature, and water vapor profiles of the atmosphere if desired for other operations. Determination of geometric horizon proceeds by capture of a long series of IMU pitch and roll angles. The sinusoids are fit to the time series of pitch and roll angles and mean angles then determined. Fourier transform is applied to the time series of pitch-roll, and the dc component of the Fourier transform is inspected. Sinusoid fit is compared to Fourier results and assessed. The geometric horizon is determined from mean angles. A Front-weighted Kalman filter is applied the geometric horizon determination and used as an input to ongoing IMU pitch and roll angle series capture. Alternatively, long-term average of IMU pitch/roll could be used to find the true gravitational vector, or a combination of methods, combined via Optimal Estimation methods, for example.

[0087] The angular relationships between the incidence angles of the brightest pixel in the FOV and the infrared camera point to this brightest pixel are determined at the processor using the steps in FIG. 17. Sky blackbody temperature is mapped to a high elevation angle. A method of measuring sky blackbody temperature at elevated angle is selected from the either the method of pointing the IR camera to the sky at an elevation where the IRC temperature range is on scale, or the method of extracting sky blackbody temperature from an infrared image and determining the elevation angle above horizon. The angle of a selected pixel relative to the geometric horizon (from FIG. 16) is determined. The Mean Radiating Temperature of the atmosphere is determine or estimated. The atmospheric opacity from the Fredholm equation is then calculated and opacity to the desired high angle is mapped using a trigonometric function. Opacity and brightness temperature equations are then used to determine brightness at desired the elevation angle.

[0088] The process for determining the desired high angle downwelling flux from lower elevation angle infrared camera observations is shown in FIG. 18. To determine the warmest pixel, the IRC is pointed seaward to a close by off-nadir angle to preclude reflections of the image of the platform (ship for example) structure. The true camera off nadir angle is received (via FIG. 16). The IRC captures infrared image and sorts for the brightest pixel. Quality control (QC) algorithms are applied for removal of sun glint, sea foam, bio material and other contaminants and rejected pixels are removed from the ensemble of pixels. The brightest uncontaminated pixel in the image is located, and angular relation to the camera and the downwelling sky flux incidence angle is determined. Downwelling flux is then determined (see FIG. 17).

[0089] A preferred embodiment of system 31 of the apparatus of this invention is illustrated in FIGS. 19 to 23. The preferred embodiment utilizes the 8 to 14 micron band of infrared camera 33 (for example, an ICI model 8640 camera with a 25 degree field of view). A visible image camera 91 is also provided (for example, a DAHUA IPC-HFW1420SP). The cameras are mounted in hermetic enclosure 93 such as a ROSE BOPLA ALS 01210 to protect from the harsh marine environment. Enclosure 93 includes main case 95 and lid 97 sealably securable to the main case using bolts 99. Infrared camera 33 views through window 101 such as a germanium window (for example, a KNIGHT OPTICAL WGE6004-C 60 mm diameter hard carbon and antireflective coated 7-14 um germanium window).

[0090] The enclosure contains embedded processor 103 (such as a RUGGED SYSTEMS ECS-4500 computer) that performs system housekeeping functions, controls the operational process steps of the sea surface temperature measurement system hereinabove, captures data from the system subsystems, determines the geometric horizon, obtains data regarding camera position and responsive thereto controls positioner 105 (connected by bolts 106 to enclosure 93) operations, processes the infrared images to determine true sea surface temperature, and communicates the data via Ethernet cable, fiber optics, Wi-Fi or other serial format via interface panel 107 to a selected central site. Panel 107 preferably uses hermetic connectors mounted through a panel of the enclosure (for example GLENAIR hermetic stainless steel circular connectors 109) for connections to Ethernet, elevation-over-azimuth PoE positioner, 12-32 vdc or 90-260 vac, USB, and video. Power supply 111 is mounted on board.

[0091] The internal subsystems include an IMU 113 such as a XSENS MTI-G-710 series Inertial Measurement Unit to determine the pointing attitude of the infrared camera, capture global navigation system satellite (GNSS) position and velocity via antenna 115, and determine heading of the motion of the ship. Surface temperature and relative humidity are measured by sensor 117 (for example, a ROTRONIC HC2A sensor probe) mounted in a separate attached and ventilated enclosure 119 (including main case 121 and lid 123 held by bolts 125). The sensor probe enclosure is ventilated with a small fan 127 drawing in uncontaminated air through a filtered intake 129 by exhausting rather than inputting sample air thereby avoiding heating from the fan operation. Various additional features which may be employed include the GNSS antenna 115, a breathing dissicator mount 131 (AGM, for example, with indicator and pressure differential cracking valve), and a Schrader valve port 133 for pressure testing. Handles 135 are provided for handling and storage operations, and mounting plate 137 accommodates mounting to the positioner. Signal (communication, control and data) and power routing are all accomplished using selected conventional I/O cabling and devices. The apparatus of this invention is designed to operate on 22 to 32 vdc, a common ship board voltage that is not subject to the safety requirements of voltages above 48 volts, or 90 to 260 vac.

[0092] The pan-tilt PoE positioner 105 responsive to the processor/subsystems, such as a J-SYSTEMS JPTH-35PoE articulated elevation-over-azimuth positioner, is preferably used and is a self-contained positioner without external controller. It preferably has a high IP rating, incorporates serial command interfacing, has an elevation range for 90 to +60 angular elevation degrees, is corrosion resistant, and has sufficient payload torque capacity and MilSpec environmental specification suitable to this task. The positioner enables viewing at chosen downward or upward looking angles.

[0093] The infrared measurement system can therefore be directed to various azimuth angles and incidence angles and accommodates selected azimuth point. The preferred installation location shipboard is in the bow area of a ship as illustrated by FIG. 1 such that undisturbed oceanic surface can be observed free of the water movement and mixing caused by the ship passage. Alternatively, the system can be mounted on the ship superstructure or any location where a clear view of the sea surface ahead of the bow wave is available. Thus, an undisturbed image of the reflected downwelling thermal sky radiation in the field of view of the infrared camera system as reflected by the sea surface and thermal radiation emitted by the sea surface are received by the infrared camera system. These two combined flux sources are mathematically separated by calculations of the processor to extract the true sea surface temperature.

[0094] Error sources associated with the hardware and methods of this preferred embodiment are shown in below. The expected error upon the sea surface temperature determination in utilizing the neutral density filter or blackbody screen for the Infrared Cameras Incorporated Model 8640 and preferred hardware and methods is presented in TABLE 1.

TABLE-US-00001 TABLE 1 Estimated total Square Root of Sum of Squares error in sea surface temperature considering only the neutral density filter measurement method and blackbody screen method errors. Note that the infrared camera is the dominating error. Induced sea surface Error Source Magnitude of temperature 35% transmission Maximum Uncertainty Error ND filter 3% 0.028 C. transmission/loss in band Temperature of ND filter 5 C. 0.08 Temperature of blackbody screen 5 C. 0.08 C. Reflected FOV 4 C. 0.032 C. 8640 infrared camera NEDT, delta T 0.30 C. Total Estimated Error, SRSS 0.32 C. SRSS if infrared camera could be calibrated to 0.2 C. 0.22 C.

[0095] The total system expected errors associated with the hardware and methods of this preferred embodiment utilizing the Infrared Cameras Incorporated Model 8640 and preferred hardware and methods is presented in TABLE 2.

TABLE-US-00002 TABLE 2 Estimated total error in sea surface temperature with zenith- mapping measurement method. 200 C. sea surface temperature- sky difference assumed. The camera error dominates. Induced Parameter and sea surface Maximum Uncertainty temperature Error Source Magnitude Error Mean Radiating T.sub.MR 20 degrees 0.1 C. Temperature Incidence/Reflection Angles Theta 10 degrees 0.045 C. Index of Refraction n +/ 0.05 n-units 0.03 C. 8640 infrared camera NEDT, delta T 0.30 C. Misestimation of zenith- Reflected downwelling 0.025 C. mapped downwelling flux flux Total Estimated Error, Square Root Sum of Squares (SRSS) 0.32 C. SRSS if infrared camera could be calibrated to 0.2 C. 0.23 C.

[0096] As may be appreciated from the foregoing, high accuracy apparatus and methods are disclosed for sea surface temperature measurement (absolute accuracy of the order of magnitude of 0.3 C and perhaps less has been obtained in testing). The infrared camera apparatus is also capable of determining vertical profiles of tropospheric temperature, water vapor, and refractivity from the surface to 10 km of altitude, as described in U.S. patent application Ser. No. 16/350,267. Thus the system can perform multiple functions.