| Popis: |
This thesis addresses the problem of cubesat limitations on transmission power and onboard memory storage. The number of small satellites is continuing to increase. The reduced amount of time and budget required for the development of these satellites has considerable advantages. The short time leads to data becoming available faster than from a larger satellite. Consequently, the communication system is very important to ensure that all data from the cubesat can be transmitted to the receiving station. In this thesis the link budget of a cubesat has been studied to identify the constraints on power and data transmission. As cubesat satellites become more complex, additional constraints and requirements are placed on system components. For more complex missions, greater flexibly of the onboard computer architecture is required to support the mission adaptation or changing specifications of onboard devices. Alternative onboard computer architecture for the next generation of cubesats is presented in this thesis and hybrid onboard computer architecture is proposed. There are many cubesats which have provided remote sensing imagery. An issue is how to store the data onboard and how to transmit these data with limited power. A solution is to reduce the size of the original image by pre-image processing. The potential for using image compression and defining the region of interest to decrease the original satellite image size has been examined in this research. Three approaches are studied and described in the context of the region of interest technique. There is image segmentation based on edge, histogram and texture detection. The presented evaluation is focused on the detection of the land part of the image that contains dynamic information and rejecting the ocean where there is less interest. The technique, however, is equally applicable for any region of interest that can be characterised and this is illustrated by considering some examples. The proposed adaptive image compression system is made up of two parts. The first part consists of the identification of the region of interest and the second part the image compression of this region of interest. The accuracy of the proposed system has been examined by comparing the number of different pixels between the proposed automatic region of interest system and the manual detection of the region of interest. Morphological methods are the main technique that has been used in the system. The morphology structure element has different shapes and size and it is necessary to understand how the shape and size of the structure elements affects the proposed system. A study of structure element has been conducted. In the real implementation of the proposed system on a cubesat, additional power would be required. To quantify this increase, a particular proposed system based on edge segmentation for region of interest automatic detection has been studied. The potential for using the proposed image compression to detect the region of interest and image compression was examined using a standard microcontroller. The result shows that the proposed system could be used on a cubesat satellite with reasonable additional power and mass. |
