Discrimination Method between Prolate and Oblate Shapes of Leaves Based on Polarization Characteristics Measured with Polarization Film Attached Cameras

Method for discrimination between prolate and oblate shapes of leaves based on polarization characteristics is proposed Method for investigation of polarization characteristics of leaves by means of Monte Carlo Ray Tracing: MCRT is also proposed. Validity of the proposed discrimination method is confirmed with MCRT simulations. Also field experiments are conducted. Through field experiments at the tea estates situated in Saga prefecture, a validity of the proposed method is confirmed. Also discrimination between prolate and oblate shapes of leaves is attempted. The results show that the proposed method is valid and discrimination can be performed.


I. INTRODUCTION
Plant identification is very complex and required an appropriate background with significant experience.Also the number of researchers who can identify medicinal plants is very limited.Generally, identification process of medicinal plants has been done manually by a herbarium taxonomist using guidebook of taxonomy/dendrology.The proposed method is designed to help taxonomist to identify leaf medicinal plant automatically using a computer aided system.
In particular, it would be better to identify leaves by using airborne based radiometer data Although spectral reflectance can be estimated with the radiometer data, it is difficult to identify the leaves in concern by using only the spectral reflectance.The method proposed here is to use spectral polarization camera or radiometer.It is obvious that polarization characteristics are different between needle leaves and broad leaves [1]- [10].Therefore, it may be possible to identify the leaves in concern based on the polarization characteristics.
There is another strong demand to discriminate between vertically standing leaves and horizontally laying leaves.For instance, new borne and young leaves are vertically standing leaves while elder leaves are horizontally laying leaves.In order to determine the most appropriate harvesting timing for tealeaves, such the discrimination is needed.Spectral characteristics between both young and elderly tealeaves are not so easy to discriminate.Obviously, although Bi-Directional Reflectance Distribution Function: BRDF are different between young and elderly tealeaves [5], it does costly for measurement.Therefore, the proposed discrimination method with polarization characteristics measurements is useful.It is effective and relatively easy as well as comparatively cheap.Only instrument required for this is cheap camera with polarization film.
As the first step, the method which allows discriminate needle leaves and broad leaves is proposed.The proposed method is also possible to distinguish between prolate and oblate shapes of leaves even if the size and scale of the leaves in concern are same.
The following section describes the proposed method and system including basic concept of the proposed method, scientific background together with Monte Carlo Ray Tracing: MCRT simulation method for validation of the proposed method [2], [7], [9].Then simulation results and field experiment results are described followed by conclusion with some discussions.

A. Proposed System
Only thing required for the proposed system is polarization film attached camera.Major specification of the camera Cannon S-100 used for the experiment is shown in Table 1.At the optical entrance of the camera, polarization film has to be attached.The camera has to be set-up at the relatively high pole or tower.Most of farm areas have lighting poles or fan attached tower for air convections for avoidance from the frosting damage.In particular, tea farm areas have fan attached poles with electricity supply.Therefore, the camera www.ijarai.thesai.orgcan be amounted on at a high position.The recommendable observation angle is 55 degrees.Tealeaves surface is used to be covered with oily materials and water.In order to acquire their polarization characteristics, Brewstar angle between air and water or oily materials (55 degrees) is recommendable.Camera acquired data can be transferred through WiFi or wireless LAN communication links.Figure 1 shows the proposed system.

B. Process Flow
The proposed method in this paper uses polarization radiometer data of leaves in concern which is acquired with airborne based polarization radiometer.p and s polarization of reflectance of the leaves in concern is observed.Then Degree of Polarization: DP which is represented as equation ( 1) is calculated.
where Rp and Rs denote surface reflectance for p polarization and that for s polarization, respectively.DP depends on the shape of leaves in concern obviously.Therefore, discrimination between prolate and oblate shapes of leaves as well as distinguishes between needle leaf and broad leaf becomes available results in identification of medicinal leaves.In order to validate the proposed method, simulation study with Monte Carlo Ray Tracing model is conducted together with field experiments.

C. Monte CarloRay Tracing Simulation
In order to validate the proposed method, MCRT simulation study and field experimental study is conducted.MCRT allows simulation of polarization characteristics of sea surface with designated parameters of the atmospheric conditions and sea surface and sea water conditions.Illustrative view of MCRT is shown in Figure 2. Photon from the sun is input from the top of the atmosphere (the top of the simulation cell).Travel length of the photon is calculated with optical depth of the atmospheric molecule and that of aerosol.There are two components in the atmosphere; molecule and aerosol particles while three are also two components, water and particles; suspended solid and phytoplankton in the ocean.When the photon meets molecule or aerosol (the meeting probability with molecule and aerosol depends on their optical depth), then the photon scattered in accordance with scattering properties of molecule and aerosol.The scattering property is called as phase function 1 .In the visible to near infrared wavelength region, the scattering by molecule is followed by Rayleigh scattering law [11] while that by aerosol is followed by Mie scattering law [11].Example of phase function of Mie scattering is shown in Figure 3 (a) while that of Rayleigh scattering is shown in Figure 3   In the atmosphere, there are absorption due to water vapor, ozone and aerosols together with scattering due to the atmospheric molecules, aerosols.Atmospheric Optical Depth: AOD (optical thickness) in total, Optical Depth: OD due to water vapor (H2O), ozone (O3), molecules (MOL), aerosols (AER), and real observed OD (OBS) are plotted in Figure 4 as an example.For simplifying the calculations of the atmospheric influences, it is assumed that the atmosphere containing only molecules and aerosols.As shown in Figure 4, this assumption is not so bad.Thus the travel length of the photon at once, L is expressed with equation (2).
where Zmax, τ, RND(i) are maximum length, altitude of the atmosphere, optical depth, and i-th random number, respectively.In this equation, τ is optical depth of molecule or aerosol.The photon meets molecule when the random number is greater than τ.Meanwhile, if the random number is less than τ, then the photon meats aerosol.The photon is scattered at the molecule or aerosol to the direction which is determined with the aforementioned phase function and with the rest of the travel length of the photon.

D. Ground Surface with Slopes
When the photon reaches on the ground, the photon reflects at the ground surface to the direction which is determined by random number.Lambertian surface [11] is assumed.Therefore, reflectance is constant for all the directions (from all the aspects).The reflected photon travels with the rest of travel length.Flat Lambertian surfaces are assumed.

E. Top of the Atmosphere: TOA Radiance Calculation
If the photon reaches on the wall of the simulation cell, the photon disappears at the wall and it appears from the corresponding position on the opposite side wall.Then it travels with the rest of travel length.Eventually, the photons which are reached at the top of the atmosphere are gathered with the Instantaneous Field of View: IFOV of the Visible to Near Infrared Radiometer: VNIR onboard satellite.At sensor radiance, I+ with direction and IFOV of μ, μ0 can be calculated with equation ( 4) where N+ is the number of photons which are gathered by VNIR, Ntotal denotes the number of photons input to the simulation cell.Also I denotes extraterrestrial irradiance at the top of the atmosphere.

F. Leaf Surface Model
Prolate and oblate shapes of leaf surface models are created as shown in Figure 5.These leaves are aligned with d of interval.New borne and young tealeaves are vertically standing leaves while elder tealeaves are horizontally laying leaves.Prolate leaves are assumed to be new borne and relatively young tealeaves while oblate leaves are also assumed to be comparatively well grown elderly tealeaves.The surfaces of the prolate and oblate leaves are modeled with Lambertian surface.The reflectance of the leaves is determined by empirical models.In order to determine the most appropriate harvesting timing for tealeaves, such the discrimination is needed.www.ijarai.thesai.org

A. Simulation Results with MCRT
MCRT based simulation is conducted for p and s polarizations individually.The sizes of prolate and oblate shapes of tealeaves are ranged from 5 cm to 20 cm with 5 cm step.10 million of photons are input to the atmosphere with optical depth of 0.15 for aerosols and 0.3 for atmospheric molecules at the wavelength of 500 nm as well as 0.2 for atmospheric molecules and 0.02 for aerosols at the 730 nm.The reflectance of the tealeaves are set at 0.215 for p polarization and 0.46 for s polarization in the case of oblate shape while 0.22 for p polarization and 0.21 for s polarization in the case of prolate shape), respectively.The reflectance is provided by Grant, L. (Polarized and non-polarized components of leaf reflectance, Doctoral Dissertation, Purdue University, 1985).
Figure 6 shows the number of photons reached at the top of the atmosphere at the wavelength of 730 nm.Although it is not clear the differences, DP which is shown in Figure 7 shows clear difference between prolate and oblate shapes of tealeaves.Fig. 7. DP calculated for prolate and oblate shapes of tealeaves at 730 nm Meanwhile, the reflectance of the tealeaves is 0.1 for p polarization and 0.5 for s polarization for prolate shape while 0.08 for p polarization and 0.21 for s polarization for oblate shape at the wavelength of 550 nm, respectively.MCRT results for 550 nm which is shown in Figure 8 shows almost twice much DP in comparison to those for 730 nm.

B. Experimental Results at Saga Prefectural Tea Institute
Field experiments are conducted at the Saga Prefectural Tea Institute: SPTI situated at 33:07:03.9N, 129:59:47.0E on June 1 2008.The first harvesting is finished in the begging of May.After the harvesting of tealeaves, the top tea trees are used to be cut out.Then new tealeaves appear after that.June 1 is middle moment between the first and the second harvests.There are four tea farm areas which are situated North, East, South, and West sides of the SPTI main building as shown in Figure 9.  Figure 11 shows top view of the tealeaves at the four different tea farm areas together with the grass situated at just beside of east tea farm area.Although it is hard to distinguish among these visually, reflectance in near infrared wavelength region of tealeaves is different each other as shown in Figure 12. Figure 12 also shows not only spectral reflectance but also calculated spectral DP.In this case, spectral DP is measured with spectral radiometer of MS-720 manufactured by Eiko Co. Ltd, Japan.Major specification of MS-720 is shown in Table 2.  Figure 12 Measured spectral reflectance and DP at tea farm areas which are situated at north, east, south and west tea farm areas of SPTI.Much appropriate wavelengths for discrimination of prolate and oblate shapes of tealeaves are www.ijarai.thesai.org500 nm, 675 nm and 800 nm as shown in Table 3.As shown in Table 3, DP of the grass is always small in comparison to the others.On the other hand, DP of the tealeaves at north tea farm area and that at east tea farm area shows specific features.Namely, DP of the tealeaves at east tea farm area is quite large in comparison to the others while DP of the tealeaves at north tea farm area shows also remarkably large comparing to the others.Sizes and shapes as well as grow up rates of tealeaves are different from each other species, obviously.Both species of east and north tea farm areas are same, Yabukita.In the begging of July, just two weeks later of field experiments, the second tealeaves harvest is conducted.Best harvesting time periods are different by species by species, farm area by farm area.Measured DP would be useful to determine best harvesting time period.

IV. CONCLUSION
Method for discrimination between prolate and oblate shapes of leaves based on polarization characteristics is proposed Method for investigation of polarization characteristics of leaves by means of Monte Carlo Ray Tracing: MCRT is also proposed.Validity of the proposed discrimination method is confirmed with MCRT simulations.Also field experiments are conducted.Through field experiments at the tea estates situated in Saga prefecture, a validity of the proposed method is confirmed.Also discrimination between prolate and oblate shapes of leaves is attempted.The results show that the proposed method is valid and discrimination can be performed.As the results from simulation study and field experiments, it is found that the proposed method is useful for tealeaves harvesting time period.It also can be used validated.The proposed tea farm area monitoring system does works well.Tea farmers may take a look at their farm area with their mobile phone together with meteorological data.

Fig. 1 .
Fig. 1.Proposed system with polarization film attached camera mounted on the relatively high position which allows look down the farm area with 55 degree of Brewstar angle.Proposed system consists of the polarization film attached camera mounted on the relatively high position which allows look down the farm area with 55 degree of Brewstar angle and meteorological station.Camera data and meteorological data are gathered through WiFi or wireless LAN communication links at farmers' houses and the farm area monitoring center through Internet

Fig. 2 .
Fig. 2. Illustrative view of MCRT for the atmosphere and sea water (b).

Fig. 4 .
Fig. 4. Example of observed atmospheric optical depth in total and the best fit curves of optical depth due to water vapor, ozone, molecules, and aerosols calculated with MODTRAN of atmospheric radiative transfer software code..

Fig. 6 .
Fig. 6.Example of MCRT simulation results of the number of photons reached at the top of the atmosphere at the wavelength of 730 nm.

Fig. 8 .
Fig. 8. DP calculated for prolate and oblate shapes of tealeaves at 550 nm It is clearly found that the calculated DPs for prolate and oblate shapes of tealeaves are quite different.Approximately twice much DP is shown for prolate shape of tealeaves in comparison to the DP for oblate shape of tealeaves.Therefore, it is possible to discriminate between prolate and oblate shapes of tealeaves.

Fig. 10 .
Fig. 10.Examples of photos of the tea farm area of SPTI and meteorological data which area situated at just beside the tea farm areas Fig. 11.Top views of tealeaves of the tea farm areas of SPTI.Bracket indicates species of the tealeaves Only the reflectance of the leaves at the wavelength of 550 and 730 nm are given by Grant.Although the reflectance at are similar.

Fig. 12 .
Fig. 12. Measured spectral reflectance and DP at tea farm areas which are situated at north, east, south and west tea farm areas of SPTI

TABLE III .
DP CALCULATED WITH MEASURED SPECTRAL P AND S POLARIZED REFLECTANCE AT EAST,WEST, SOUTH AND NORTH TEA FARM AREAS AS WELL ASGGRASS SITUATED JUST BESIDE THE EAST TEA FARM AREA ON JUNE 1 2008