Confirmation
The analysis program in development has been continually tested with sample data to find some preliminary results and ensure it functions correctly. The following is a description of the steps taken to make sure the program works correctly.
One of the first challenges that arose was to extract information from the coordinate file to convert all x and y positions from pixels to some user defined unit, define the center and a diameter line for line count and velocity processing, and to scale x y position graph. To test this function, a sample set of coordinate were tested in the program and graphed through MS Excel. This function was confirmed working when the Excel plot correctly showed the coordinates and line crossings.
To confirm the accuracy of the radial count and speed, sample data was processed by the program and Excel was again used to confirm the same results. Furthermore, histograms of these categories were also graphed to show further evidence that the program would yield the correct values. While no FFTs were done through excel, since the FFTs are built in functions in Labview that take in the same data as the histogram function, they can be assumed to be correct when the raw analysis and histogram are found to be correct.
Line Count uses vector operations to find the when the fly crosses the line. The cross product of a vector lying along the dish diameter and the vector from a common point to the fly position will be positive or negative depending on which side of the first vector the second vector lays. Thereby, the finding the sign of the determinant and counting how many times it changes will determine the number of line crossings. A set of data was tested for the determinant sign function and compared with its excel graph to confirm that the function correctly determined which side of the line the fly was. Further LabView tests of the final result of a data set with known number of crossings determined this function to be correct.
To Test radial and angular velocities, the center of the dishes had to be correctly identified. This confirmation is described above in the graph of the sample coordinates. The radial velocity was then confirmed by Excel and LabView processing of the same sample data. Angular velocity had to first be confirmed by testing the find angle function of the program by inputing data generated from Excel with known angles. When this was confirmed accurate, it was also tested in LabView and Excel with the same sample data to confirm the accuracy of the data.
It should be noted that in all tests, there were some slight expected discrepancies between the program and Excel because the program can only process a certain set of data points at a time causing it to lose one data point for each set processed. Since the tracker takes data in tenths of a second, this loss was determined to be negligible.