Journal of Surveying Engineering

For the past three years, an annual campaign of global positioning system (GPS) measurements has been mounted in the Okavango Delta in Botswana, as part of a multidisciplinary research project. The project examines the hydrology, sedimentology, and plant ecology of the wetland. Critical to the project is the determination of the relative heights of water gauges and other points within this 22,000 km² area. GPS has proven to be a key technology in achieving this aim. Nevertheless, its use does introduce a complication in that the heights derived using GPS are ellipsoidal, whereas orthometric heights are required in the context of modeling water flow. Conversion from one type of height to the other requires an accurate model of the geoid, but this is not easy to produce in such a data-scarce area. In this paper, the writers describe the GPS surveys and the adjustment and connection of the GPS networks to existing survey control. Two geoid models for the region are assessed and their significance evaluated.

A method for estimating the accuracy of data collected from aerial photographs is presented. The emphasis is on its implementation in accuracy estimation when gathering data for digital terrain model databases. The method, based on error propagation of the photogrammetric solution process, facilitates the determination of accuracy for applications based both on direct photogrammetric measurements and on digital terrain model databases in addition to determining the measurements' accuracy. The mechanism and implementation at different levels of complexity is described, and various aspects of the proposed method are demonstrated, especially with applications dealing with the altimetric component of the terrain.

In global positioning system (GPS) data processing, incorrect stochastic models for double-differenced measurements will result in unreliable statistics for ambiguity search and biased positioning results. In the commonly used stochastic model, it is usually assumed that all the raw GPS measurements are independent and that they have the same variance. In fact, these assumptions are not realistic. Measurements obtained from different satellites cannot have the same accuracy due to varying noise levels. In this paper, a new method based on modern statistical theory is proposed to directly estimate the covariance matrix for double-differenced GPS measurements. Three different stochastic models have been tested and analyzed. Test results indicate that by using the proposed stochastic models the volume of ambiguity search space can be reduced and the reliability of the ambiguity resolution is improved. Also, the statistics of the baseline components estimated with the proposed stochastic models are more efficient.