HEIGHT ESTIMATION USING STEREO SIGNAL PROCESSING
| dc.contributor.advisor | Salt, J. E. | |
| dc.contributor.advisor | Wood, H. C. | |
| dc.creator | Randhawa, Tejinder Singh | |
| dc.date.accessioned | 2024-07-17T20:24:33Z | |
| dc.date.available | 2024-07-17T20:24:33Z | |
| dc.date.issued | 1989-07-27 | |
| dc.date.submitted | July 1989 | |
| dc.description.abstract | A sensor system, which estimates the mean height of the plane of the sensors above the ground surface, has been proposed and analysed in this thesis. The sensor system consists of two spatially separated optical sensors and a signal processor. The signal processor consists of four subsystems : a crosscorrelator, a filter, an averager and an amplifier. Although this sensor system has a wide range of applications, its applicability in estimating the height of agricultural machines above the ground surface is analysed in par-ticular in this thesis. A mathematical model of the system has been developed. The output signals of the two sensors are modelled as a common signal, time-shifted in the case of one of the sensors, and corrupted by additive noise signals. The shift in the output signals is a function of the height variations of the ground surface. For the purposes of the analysis, the common signal is assumed to be the sample function of a correlation ergodic process and the noise signals are from mean ergodic processes with zero mean. The time varying shift is assumed to be the outcome of a distribution ergodic process. The common signal, the noise 'signals and the shift signal are assumed to be mutually independent and thus uncorrelated. With these two signals as the inputs, the output of the signal processor is the mean height. The output of the crosscorrelator has been proved theoretically to be the convolution of the autocorrelation function of the common signal with the probability density function of the shift. The filter recovers the density function from the crosscorrelator output. The expression of the filter response has also been derived in this thesis. The averager calculates the mean delay, which mul-tiplies with the gain of the amplifier to give the mean height. The model is analysed using computer simulations. Different types of shift signals are used in simulations to verify the working theory of the sig-nal processor. To test the feasibility of the system in estimating the mean height in real time, some preliminary data have been collected using a prototype system. The data have been processed and the results are presented. Some general conclusions are drawn about the efficiency as well as the limitations of this system in estimating the mean height under prac-tical situations. | |
| dc.identifier.uri | https://hdl.handle.net/10388/15839 | |
| dc.title | HEIGHT ESTIMATION USING STEREO SIGNAL PROCESSING | |
| dc.type.genre | Thesis | |
| thesis.degree.department | Electrical Engineering | |
| thesis.degree.grantor | University of Saskatchewan | en_US |
| thesis.degree.level | Masters | |
| thesis.degree.name | Master of Science (M.Sc.) |