Precise results of matching are required to obtain the precise locations of objects using stereo images. Accordingly, sub-pixel estimation of the corresponding position in area-based matching is often conducted. Several methods to obtain sub-pixel estimation in area-based matching have been proposed. However, there are few reports on accuracy comparison of those methods. Therefore, we performed an experiment in order to investigate accuracy of sub-pixel estimation methods in area-based matching by using 54 diverse images. Our study focused on two popular methods in digital photogrammetry: cross correlation method with fitting a second order surface by using normalized cross correlation coefficients (CMF), and least squares matching (LSM) . The experiment results demonstrate that the accuracy of sub-pixel estimation methods in area-based matching varies considerably with the properties of the image. Furthermore, the experiment results show that LSM can estimate the corresponding position more accurately than CMF. The estimation error of LSM would be average half of that of CMF.
In general, apparent reflectance data obtained by airborne hyperspectral sensor are affected by Bidirectional Reflectance Distribution Function (BRDF) of surface medium. A complete preprocessing of airborne hyperspectral sensor data should include the correction of a possible brightness gradient for measuring the feature of surface medium. For a line scanner like airborne hyperspectral sensor, the brightness gradients mainly occur in across-track direction and depend on the sensor's view angle. However, when the observations are done with the small planes, hyperspectral sensor attitude changes in the 3 axes directions, such as a rolling, pitching, and yawing. Therefore, in order to correct brightness gradient with sufficient accuracy, it is necessary to take into consideration not only the across-track direction but the alongt-rack direction also. A Highly precise simulation of BRDF using hyperspectral sensor attitude information measured by GPS/IMU is indispensable. This research examines the technique of conducting the simulation of the BRDF with the empirical approach using the lapped images of two or more hyperspectral data. The proposed technique enables the estimation of the nadir reflectance value without any additional field measurements and the estimated nadir reflectance value can be used for any subsequent qualitative and quantitative analysis.
High quality rice production requires periodically collecting rice growth data to control the growth of rice. The height of plant, the number of stem and the color of leaf are well known parameters to indicate rice growth. Rice growth diagnosis method based on these parameters is used operationally, although collecting these parameters by field survey needs a lot of labor and time. A laborsaving method for rice growth diagnosis was proposed which was based on vegetation cover rate of rice. Vegetation cover rate of rice is calculated based on discriminating rice plant areas in a digital camera image that is photographed in nadir direction. In order to estimate nitrogen content of rice plants from vegetation cover rate of rice obtained by a simple image measurement, the relation between both was analyzed in this paper. As results of the analysis, the correlation coefficient of both was 0.93 and RMS was 0.69 g/m2.
A two dimension laser scanner is vertically installed in ground to the measurement chassis. The point data is acquired while moving the measurement chassis. The three dimension map point data is made from the point data by using the locus data of the measurement chassis. In this thesis, we propose the method of making the two dimension map vector data from the three dimension map point data under indoor environment.