2. SYSTEM DESCRIPTION

Transcription

1 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will Miniaturized hyperspectral imager calibration and UAV flight campaigns Heikki aari a, Ilkka Pölönen b, Heikki alo b, Eija Honkavaara c, Teemu Hakala c, Christer Holmlund a, Jussi Mäkynen a, ami Mannila a, Tapani Antila a, Altti Akujärvi a a VTT Technical esearch Centre of Finland, P.O. Box 000, FI-0044 VTT, Finland; b Jyväskylä University, Jyväskylä, Finland; c Finnish Geodetic Institute, Kirkkonummi, Finland ABTACT VTT Technical esearch Centre of Finland has developed Tunable Fabry-Perot Interferometer (FPI) based miniaturized hyperspectral imager which can be operated from light weight Unmanned Aerial Vehicles (UAV). The concept of the hyperspectral imager has been published in the PIE Proc. 7474, 874 and This instrument requires dedicated laboratory and on-board calibration procedures which are described. During summer 0 extensive UAV Hyperspectral imaging campaigns in the wavelength range nm at resolution range 0 40 FWHM were performed to study forest inventory, crop biomass and nitrogen distributions and environmental status of natural water applications. The instrument includes spectral band limiting filters which can be used for the on-board wavelength scale calibration by scanning the FPI pass band center wavelength through the low and high edge of the operational wavelength band. The procedure and results of the calibration tests will be presented. A short summary of the performed extensive UAV imaging campaign during summer 0 will be presented. Keywords: hyperspectral sensors (missions, designs, performance, technologies, airborne sensors, etc.) Fabry-Perot Interferometer, Piezo actuators, imaging spectrometer, UAV, airborne. INTODUCTION Lightweight UAV platform based aerial imaging is a cost effective and versatile method for gathering information from small and medium sized areas. The light weight UAV imaging is typically performed with color, CI (Color Infrared) or multispectral cameras. ome trials have been made also with small, push-broom hyperspectral imagers with light weight UAV platforms. VTT has developed hyperspectral imaging technology for small UAV platforms,3,4 and it has been used during the 0 and 0 flight campaign in the project UAI- Unmanned Aerial ystem Innovations 5. The first prototype was used during summer 0 to gather spectral data for forest and agriculture applications,3. With the acquired experience and results from the first flight campaign a second prototype was developed for the summer 0 flight campaigns 4. With a combination of a spectral imager and a CI camera it is possible to generate georeferenced spectral mosaics, digital surface models and CI mosaics from the data acquired during a single flight 6. For the needs of precision agriculture this opens up possibilities for crop monitoring, plant disease detection and improved fertilization planning. In the forest applications the main target is to improve the biomass estimation and tree species detection accuracy.. YTEM DECIPTION The UAI imaging system is modular enabling to use the CI camera and the spectral imager separately if the payload capacity of the UAV platform is not adequate to carry both imagers at the same time. The block diagram of the main components of the imaging system is shown in Figure. The imaging system has its own dedicated GP module which is used to provide an accurate time correlation between captured images and UAV flight data. The imaging system is also equipped with downwelling and upwelling irradiance sensors to monitor illumination conditions during the flight. For this purpose reference targets or ground based measurements can also be used. The image triggering can be controlled by the autopilot or the system can be set to take images at predefined time intervals V. 4 (p. of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

2 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will Figure Block diagram of the UAV operated Hyperspectral and Color Infrared (CI) imaging system. 3. HYPEPECTAL IMAGE 3. Operation principle of the Fabry-Perot Interferometer hyperspectral imager The operation principle of the UAI Hyperspectral imager is based on multiple orders of the Fabry-Perot Interferometer (FPI) that are used matched to the different sensitivities of the image sensor channels. For example in a Bayer pattern GB sensor or in a three CCD video camera based on a wavelength separation prism there are different pixels for three wavelength channels,7. If the air range of FPI is chosen right, there will be only one to three transmission peaks at any time. Now when a transmission spectrum with one to three peaks is recorded with a normal GB color image sensor, the spectral information can be retrieved, as red, green and blue pixels have a different spectral response throughout the wavelength range (see Figure ). This makes it possible to measure three spectral channels at once if necessary. In the lower part of Figure the combined quantum efficiencies of the detector and the FPI are shown for clarification for the gases of two and three transmitted FPI orders. 3. UAI Hyperspectral imager Detailed descriptions of the hyperspectral imagers used in the summer 0 and 0 flight campaigns are given in references and 4. The UAI 0 prototype uses a CMV Meixel CMO image sensor from CMOI8. The sensor has 5.5 µm x 5.5 µm pixels with global electronic shutter and bit ADC. The wavelength selective component in the hyperspectral camera is a Fabry-Perot interferometer. As explained in previous chapter the basic principle of the sensor is to provide different spectral layers by changing the FPI air. The FPI spectral camera can operate in the wavelength range of nm, with a full width at half maximum (FWHM) of 0-40 nm; the number bands and their characteristics can be selected flexibly according to the requirements of the application. The weight of the latest prototype is only about 600 g and it is therefore suitable even for very light-weight UAVs. Components of the imaging system in the airborne use include a hyperspectral imager, a 3 Gbyte compact flash memory card, irradiance sensor(s), a GP receiver, and a LiPo battery (Figure 3). The image size in the typical configuration is 04 x 648 pixels V. 4 (p. of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

3 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will and the pixel size is μm. Characteristic to the FPI imaging principle is that different layers of the spectral data cube are collected with small time delays, so the layers are not perfectly overlapping 6 (Honkavaara et al.,) Transmission Quantum efficiency Air = 060 nm Air = 00 nm Air = 40 nm Wavelength/[nm] Blue/B pixels Green/G pixles ed/ pixels Wavelength/[nm] imulated transmission of an FPI at three different air s. Measured quantum efficiency of the CMOI CMV4000 GB color sensor Quantum efficiency 0.05 Quantum efficiency Wavelength/[nm] B pixels at the Air = 680 nm G pixels at the Air = 680 nm pixels at the Air = 680 nm Wavelength/[nm] B pixels at the Air = 00 nm G pixels at the Air = 00 nm pixels at the Air = 00 nm Figure imulated FPI spectral transmissions, measured quantum efficiency of CMOI CMV4000 GB image sensor and combined quantum efficiency of the FPI and the detector for two different air values. The major specifications of the instruments are presented in Table V. 4 (p.3 of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

4 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will Figure 3 The Fabry-Perot hyperspectral imaging system including the camera, a 3 Gbyte compact flash memory card, irradiance sensor, a GP receiver and a LiPo battery. Table pecifications of the UAI 0 hyperspectral imager prototype. PAAMETE PECIFIED VALUE EMAK Horizontal FOV Vertical FOV pectral range > 50 > nm pectral resolution 0 40 nm, FWHM pectral step < nm F-number.7 Image sensor CMV4000 CMOI CMO image sensor with 5.5µm x 5.5µm pixels. ensor size 048 x 048 pixels Image sensor pixel clock frequency 80 Meixels/s eadout of the whole 4 Meixel image takes 50 ms Default spectral image dimensions 04 x 648 pixels x binning Max spectral image dimensions 048 x 40 pixels Power consumption < 5W Weight < 600g Main dimensions 80 mm x 97 mm x 59 mm pectral range can be selected from range nm with long and short pass filters The spectral resolution depends on the transmission order of the FPI and on the selected wavelength. For a high spectral resolution the spectral range needs to be limited V. 4 (p.4 of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

5 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will 4. Mathematical description of the concept 4. CALIBATION In the following the concept for a utilization of multiple orders of the FPI together with multispectral image sensor will be presented,7. Number of generated signal electrons of an GB image sensor Blue (B), Green (G) and ed () pixels are ( d ) = λ max η T ( λ, d ) T Φ dλ B B FPI sys ph () λmin ( d ) = λ max η T ( λ, d ) T Φ dλ G G FPI sys ph () λmin ( d ) = λ max η T ( λ, d ) T Φ dλ FPI sys ph (3) λmin In Eqs. 3 the T FPI (λ,d ) is the transmission of the FPI at the wavelength λ and at the FPI air value d, the η B (λ), η G (λ) ja η (λ) are the quantum efficiences of the B-, G- ja -pixels at the wavelength λ and the Φ ph (λ) is the spectral photon flux entering the system. The combined spectral transmission of the optical system without the FPI is included in the term T sys (λ). electing a d in such a way that there are at most three trasmittted wavelengths at three different orders of the FPI in the range λ min λ λ max the signals of the B-, G- and -pixels are originating mostly from the light at the three narrow wavelength bands of the transmitted three FPI orders. The center wavelengths of the three transmitted FPI pass bands at the FPI orders n, n+, n+ can be approximated by Where d is the FPI air width. d λ n = (4) n For the retrieval of the spectral signal at the three selected wavelength bands a calibration is required for the responses of the B-, G- and -pixels around the selected three FPI pass bands. Bn ( d, n) = d d ( + ) n n ( n ) η B TFPI ( λ, d ) Tsys dλ (5) n ( ) n n ( n+ ) V. 4 (p.5 of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

6 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will Gn ( d, n) = d d ( + ) n n ( n ) η G( λ) TFPI ( λ, d) Tsys( λ) dλ (6) ( ) n n ( n+ ) n ( d, n) = d d ( + ) n n ( n ) η ( λ) TFPI ( λ, d) Tsys( λ) dλ (7) ( ) n n ( n+ ) Where d is the FPI air width, n is the FPI order, the η B (λ), η G (λ) ja η (λ) are the quantum efficiences of the B-, G- ja -pixels at the wavelength λ, T FPI (λ, d ) is the spectral transmission of the FPI and T sys (λ) is the combined spectral transmission of the optical system without the FPI. The calibration data of the hyperspectral imager consists of the FPI air values and of the responses Bn,,, Gn,,, n, n+ ja n+ of of the B-, G- and -pixels at these air values. For a selected air value the signals of the B-, G- and -pixels Bm, Gm and m are given by Eqs. 3. The measured signals at the three narrow wavelength bands corresponding the three transmitted orders of the FPI can be calculated using the responses Bn,,, Gn,,, n, n+ and n+ (Eqs. 5-7) derived in the calibration. Bm Gm m = n+ n+ Bn Gn n n n + + n (8) Where n+, n+ and n are the unkown spectral irradiances at the pass bands of the FPI orders n+, n+ and n. The irradiances n+, n+ and n can now be solved by n n + + n = n+ n+ Bn Gn n Bm Gm m (9) The coefficients that must be applied to signals at B-, G- and -pixels to get the signal at the selected air value are given by matrix below Bn Gn n+ n+ n = n+ n+ Bn Gn n (0) V. 4 (p.6 of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

7 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will 4. Calibration measurements with monochromator The calibration was performed with a setup shown in Figure 4 for the UAI 0 and 0 prototypes. The first step in the calibration is the measurement of the spectral photon flux focused from the output slit of the monochromator to the system to be calibrated. This was performed using a calibrated absolute radiometer (UDT QED-00). The signals of the absolute radiometer and the reference detector were recorded for wavelength ranges nm at nm intervals and at a resolution of FWHM. When we know the spectral photon flux for each wavelength it is possible to determine the calibration coefficients Bn,,, Gn,,, n, n+ ja n+ as they are defined in Eq. 0. The test was carried out for 0 FPI air values from 45 nm to 653 nm. Monochromator Bentham TMc300 Integrating phere Halogen Lamp Beam splitter plate Focussing lens Absolute photon flux calibration detector Three photodiode light trap photodetector QED (Calibrated absolute photometer for quantum efficiency measurements) eference detector Piezo actuated Fabry- Perot Interferometer Module Long and hort pass filters defing the spectral range GB image sensor QED-00 Absolute radiometer Input spectral photon flux measurement arrangement Figure 4 Measurement setup for the calibration of FPI Hyperspectral imager and the input spectral photon flux. 4.3 Calibration measurements using the low and high pass filter edges A hyperspectral Imager based on a Fabry-Perot interferometer and an GB type image sensor contains the FPI order sorting filters that limits the operational spectral range of the device. Typically the order sorting consists of a long pass and a short pass filter. The order sorting filters are typically interference filters whose temperature dependence is low typically (~0.04 nm/ºc) for the edge wavelengths of the pass band. The change of the edge wavelengths as a function of the temperature can be measured and therefore the knowledge of the edge wavelengths is accurate if the ambient temperature of the filters is known. It is possible to perform the wavelength calibration of the instrument utilizing the knowledge on the spectral transmission of the FPI order sorting filters. This calibration was tried using the setup shown in Figure 5. The light from the halogen lamp is focused to the integrating sphere and the hyperspectral imager is directed to the integrating sphere which enables to record the spatially uniform spectral data cube of the lamp. A typical measured raw signals and st derivate of the -, G- and B-pixels and their sum signal as a function of the FPI air at air are presented in Figure 6 for spectral range limited by 500 nm long pass and 900 nm short pass filters. It can be seen that the sum signal of the -, G-, and B-pixels has a clear slope at the FPI air ranges nm, nm, nm and nm. The st derivate has the first clear minimum at the FPI air value close to 305 nm. The reason for this is that the FPI st order transmission band center wavelength passes the edge of the 900 nm low pass filter. The center wavelength moves from 880 nm to 90 nm when the FPI air changes from 90 to 30 nm V. 4 (p.7 of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

8 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will Integrating phere Halogen lamp or other wide band light source Beam splitter plate Piezo actuated Fabry-Perot Interferometer Module FPI order sorting filter or Long and hort pass filters defining the spectral range GB image sensor Figure 5 Test setup used in the wavelength calibration based on the short and long pass filters defining the operational spectral range of the spectrograph. The next derivate minimum is at the FPI air value 76 nm. This is caused by the nd FPI order transmission band center wavelength moving over the 900 nm low pass filter edge. imilarly the derivate minimums at the FPI air values 76 nm and 60 nm originate from the 3 rd and 4 th FPI order transmission band center wavelength moving over the 900 nm low pass filter edge. The wavelength calibration based on the short and long pass filter edges was also tested using chott BG0 and BG36 colored glass filters in front of the tested Hyperspectral imager. The spectral transmissions of these filters have many steep slopes which can be seen as st derivate minima in the raw signal in the calibration. However it was possible to separate the minima caused by the edge filters from the BG0 and BG36 caused minima. In an UAV application the wavelength calibration can be performed before and after the flight with a spectrally grey target. The absorption features of the atmosphere and the sun spectra characteristics are known a priori and can therefore be taken into account in the wavelength calibration. The wavelength calibration can be used for the correction of the wavelength scale of the UAV flight campaign spectral data cubes V. 4 (p.8 of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

9 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will Pixel raw signal/[adu] aw signal of pixels aw signal of G pixels aw signal of B pixels um raw signal of, G & B pixels FPI air /[nm] st Derivate of the pixel raw signal/[adu/nm] aw signal of pixels aw signal of G pixels aw signal of B pixels um raw signal of, G & B pixels FPI air /[nm] Figure 6 Measured raw signals and st derivate of the -, G- and B-pixels and their sum signal as a function of the FPI air V. 4 (p.9 of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

10 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will 5. UAI FLIGHT CAMPAIGN IN UMME 0 5. Overview of test flights Extensive UAV imaging campaigns were performed in the UAI project during summer 0. The objective of campaigns was to study how light weight UA imaging system can be used in the forest, agriculture and natural water monitoring applications. The imaging was carried out for forest and agricultural areas and around a small lake located in outhern Finland. For boreal forest the preferable time period is from beginning of June to beginning of August. For barley and wheat the imaging was performed just after seeding in May, Mid-June for weed detection and beginning of July for the determination of biomass. To study new application opportunities the potato test site of the Finnish Potato esearch Institute in Ylistaro, Finland were imaged both with hyperspectral and CI cameras in Mid-July and Mid-August. The possibilities to use UAV in natural water monitoring were studied by imaging Lake Petäjärvi in Kirkkonummi, Finland at wavelength ranges nm, nm and nm. The UAI imaging campaigns in 0 were carried out with the following UAVs; C-Astral fixed wing Bramor-Ortho UAV 9, Infotron.A. IT80-5 TH UAV Helicopter 0, Mikrokopter UAV helicopter, Logo UAV Helicopter and Microdrones MD4-000 UAV helicopter 3. The hyperspectral imager and CI camera was tested with four different UAV helicopters during the summer 0. Flight tests were made in collaboration with East Lapland Vocational College, Eastern Post Oy Ltd., and Finnish Geodetic Institute (FGI). The false color cameras were flown with all helicopters and with the fixed wing Bramor UAV. In total 9 test flights were carried out during the time period 0 th May to th eptember 0 (see Table ). The imaging data from the performed test flights is adequate for the assessment of how well the proposed system can be used operationally to provide the forestry, farming and environmental authority end users the information they need in the forest inventory, cut planning and precision farming and environmental status of natural waters etc.. Table UAI system test flights during the time period 0 th May to th eptember, 0. Date Test site in UAV Imagers Goal of test flight emarks Finland Vihti Mikrokopter pectral imager Crop field just after flights and CI Camera seeding flights Evo IT-80 pectral imager and CI Camera Forest studies with simultaneous spectra and CI imaging Vihti MD4-000 Weed identification flight Evo Bramor Olympus E40 Forest studies with 3 flights CI CI imaging.6.0 Vihti Bramor Olympus E40 CI Crop studies for fertilization map flight Vihti and Jokioinen Logo UAV pectral imager and CI Camera determination Crop field studies 9 flights.7. and Ylistaro and Logo UAV pectral imager Crop and potato 4 flights Isokyrö filed studies Lake, Logo UAV pectral imager tudy of the 4 flights Petäjärvi, Kirkkonummi Environmental status of a lake.9.0 Kemijärvi Bramor pectral Imager Forest studies flight V. 4 (p.0 of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

11 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will 5. esults The results of UAI Flight campaigns have been used in several application studies. Honkavaara4 et.al. presented results whose objective was to investigate the processing and use of this new imaging technology in a water quality mapping. Objectives of these investigations were to investigate the use of this technology in water quality mapping. Idea is to develop a fast method to provide high-resolution data from complex environments such as lakes, rivers and harbour areas and places where use of traditional sampling methods is limited. The technique could be potential for the measurement and monitoring of water quality parameters such as type, frequency and intensity of algae blooms, water transparency, turbidity, organic carbon, total phosphorus concentrations, and chlorophyll-a. They carried out imaging campaigns over a small lake (Lake Petäjärvi, see Figure 7) in August and late eptember 0 using a light-weight unmanned airborne vehicle (UAV) and a small manned airborne vehicle (MAV). These campaigns included many new features: they were the first water monitoring campaigns with the UAI FPI hyperspectral imager, the FPI imager was operated for the first time from a manned platform, and many filter combinations that have not been used ever before were used. The first evaluations showed that the technology is very promising4. Figure 7 Hyperspectral imaging of the Lake Petäjärvi on with UAI Hyperspectral imager using the Logo UAV helicopter. Example of a green, red and near-infra-red band composite collected at the lake Petäjärvi Kaivosoja5 et. al. and Pölönen6 et.al. have used the UAI hyperspectral data acquired from a test field at Vihti located in outhern Finland for the determination of the amount of extra fertilization for cereal crop. eeding and fertilization was done in May. Hyperspectral and CI imaging campaigns were executed in mid-summer 0 just before tillering. As Figure 8 indicates two species of wheat and barley were seeded. Different amount seed and nitrogen fertilizer were applied. The conclusion of the studies was that the extra fertilization determined using the UAV Hyperspectral image data is beneficial. However, studies also indicated that there are still many uncertainties within the process chain5,6. Figure 8 eeding and fertilizing planning of the Vihti test field used in the UAV Hyperspectral imaging campaign V. 4 (p. of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes:

12 Please verify that () all pages are present, () all figures are correct, (3) all fonts and special characters are correct, and (4) all text and figures fit within the red margin lines shown on this review document. Complete formatting information is available at eturn to the Manage Active ubmissions page at and approve or disapprove this submission. Your manuscript will 6. CONCLUION In this paper we present dedicated laboratory and on-board calibration procedures for Tunable Fabry-Perot Interferometer (FPI) based miniaturized hyperspectral imager. The instrument includes spectral band limiting filters which can be used for the on-board wavelength scale calibration by scanning the FPI pass band center wavelength through the low and high edge of the operational wavelength band. The test results and analysis of this calibration procedure are included. A short summary of the performed extensive UAV imaging campaign during summer 0 are explained and example results are also shortly reviewed. ACKNOWLEDGMENT The research has been funded by Tekes (Finnish Funding Agency for Technology and Innovation), VTT trategic esearch, Jyväskylä University, MTT Agrifood esearch Finland and Finnish Forest esearch Institute. We would also like to thank Mr. Jari Nykänen, Mr. aimo Iivarinen and Mr. Esa ärkelä from East Lapland Vocational College who operated the IT80 UAV helicopter in test flights and Mr. Patrik aski from Eastern Post Oy Ltd. who operated LOGO UAV helicopter. EFEENCE [] Zarco-Tejada, P.,J., Conzales-Dugo, V., Berni, J.J.,J., Fluorescence, temperature and narrow-band indices acquired from a UAV platform for water stress detection using a micro-hypespectral imager and a thermal camera. emote ensing of Environment 7(), (0). [] aari, H., Aallos, V., Akujärvi, A., Antila, T., Holmlund, C., Kantojärvi, U., Mäkynen, J. and Ollila, J., Novel Miniaturized Hyperspectral ensor for UAV and pace Applications, Proc. PIE 7474 (009). [3] aari, H., Pellikka, I., Pesonen, L., Tuominen,., Heikkilä, J., Holmlund, C., Mäkynen, J., Ojala, K. and Antila, T., Unmanned Aerial Vehicle (UAV) operated spectral camera system for forest and agriculture applications, Proc. PIE vol. 874 (0). [4] Mäkynen, J., Holmlund, C., aari, H., Ojala, K., Antila, T., Multi- and hyperspectral UAV imaging system for forest and agriculture applications, Proc. PIE vol (0). [5] UAI- Unmanned Aerial ystem Innovations project website, (9 August 03). [6] Honkavaara, E., Hakala, T., Markelin, L., osnell, T., aari, H., Mäkynen, J., A Process for adiometric Correction of UAV Image Blocks Photogrammetrie, Fernerkundung, Geoinformation (PFG) In press (0) [7] aari H., pectrometer and interferometric method, U Patent, U 8,30,380 B, Mar 6, 0. [8] CMOI NV., [9] C-Astral d.o.o., [0] AL Infotron, [] Hiystems GmbH, [] Kuvatekniikka Oy Patrik aski, [3] Microdrones GmbH, [4] E. Honkavaara, T. Hakala, J. Kirjasniemi, A. Lindfors, J. Mäkynen, K. Nurminen, P. uokokoski, H. aari, L. Markelin. New light-weight stereosopic spectrometric airborne imaging technology for high-resolution environmental remote sensing Case studies in water quality mapping, IP Volume XL-/W, IP Hannover Workshop 03, 4 May 03, Hannover, Germany., I-7, pp , 03. [5] Kaivosoja, J., Pesonen, L., Kleemola, J., Pölönen, I., alo, H., Honkavaara, E., aari, H., Mäkynen, J., and ajala, A., A case study of a precision fertilizer application task generation for wheat based on classified hyperspectral data from UAV combined with farm history data, to be published in Proc. PIE 8887 (03). [6] Pölönen, I., Honkavaara, E., Kaivosoja, J., Pesonen, L., aari H., Hyperspectral imaging based biomass and nitrogen content estimations from light-weight UAV, to be published in Proc. PIE 8887 (03) V. 4 (p. of ) / Color: No / Format: A4 / Date: 8/6/03 :3:50 AM PIE UE: DB Check, Prod Check, Notes: