Ryssänlampi magnetic survey using Radai UAV system and its comparison to airborne and ground magnetic data of GTK - PDF

Ryssänlampi magnetic survey using Radai UAV system and its comparison to airborne and ground magnetic data of GTK Detailed Survey Report Markku Pirttijärvi Radai Oy Contents 1. Introduction...

Please download to get full document.

View again

of 26
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 49 | Pages: 26

Extension: PDF | Download: 0

Ryssänlampi magnetic survey using Radai UAV system and its comparison to airborne and ground magnetic data of GTK Detailed Survey Report Markku Pirttijärvi Radai Oy Contents 1. Introduction Radai's UAV measurement system Ryssänlampi magnetic survey Auxiliary magnetic data Data processing Equivalent layer modelling (ELM) Results Comparison to GTK's airborne and ground magnetic data Discussion and conclusions References Introduction In August 2015 Radai Oy carried out a geophysical magnetic survey at Ryssänlampi site using an unmanned aerial vehicle (UAV). The measurements were committed for Geological Survey of Finland (GTK) as a part of the UAV-MEMO project funded by Tekes (Finnish Funding Agency for Innovation and Technology) and four mining companies operating in Finland (AA Sakatti Mining, Agnico-Eagle, Mawson, Nordkalk). The purpose of the survey was to investigate the quality and reliability of an UAV based magnetic measurement system. This detailed survey report describes Radai's magnetic UAV system, the magnetic survey in Ryssänlampi site, and the results (total magnetic field) based on equivalent layer modelling. It also shows the raw and processed data, gives some details on the data processing, shows comparisons to GTK's airborne and ground magnetic data and discusses the meaning of the results. 2. Radai's UAV measurement system Radai's UAV measurement system is installed in a custom-made UAV (Terrain Scout). The flight is controlled by an autopilot. The autopilot records flight data including GPS time and position (latitude and longitude), the orientation (roll, pitch and yaw), and barometric pressure. The real-time flight is controlled by PC software via a telemetry (radio) link. The nominal accuracy of the GPS position is about ±10 m. The magnetic field is measured using a digital 3-component flux-gate magnetometer. The magnetometer data (X, Y, Z components and total field), are recorded by Radai's own data logger. The GPS time and position are synchronized with the autopilot. A base station located near the mobile telemetry/control station measures the time variation of the total magnetic field using a proton precession magnetometer. Technical information on Radai's UAV system is given in Table 1. The UAV is illustrated in Figure 2.1 2 Table 1: Technical information on Radai's UAV system UAV parameters Electric engine Wingspan Mass Payload Flight speed Flight time UAV magnetometer Noise level Dynamic range Sampling frequency Base station magnetometer Resolution Dynamic range Sampling frequency Weight Value 780 W 2.12 m 3 kg 1 kg m/s 40 min Flux-gate ±0.5 nt ±65 T 10 Hz 0.1 nt T 1 Hz 2.2 kg Fig Radai's Terrain Scout UAV. 3. Ryssänlampi magnetic survey The survey site is located in Northern Finland about 35 km North-East from Rovaniemi. Figure 3.1 shows the survey area defined by GTK. The coordinates of the four corner points of the survey area (the red rectangle) are: (468090, ), (468762, ), (469893, ) and (469221, ). The total area of the survey site is 1.52 km 2, the suggested line spacing is 50 m and the direction of the c. 1 km long lines is along SW-NE. Please, note that ETRS-TM35FIN coordinates are used throughout this report, and the DEM data and the background maps are provided by the National Land Survey of Finland. The magnetic survey was carried out on Aug 7th The UAV survey was divided into four separate flights. Ari Saartenoja was the principal pilot and Ari Takanen, Arto Karinen and Markku Pirttijärvi were co-pilots. Communication between the pilot and the co-pilots was established using walkie-talkies. The waypoints designed for the four flight areas and the actual flight lines are plotted in Figure 3.1 using symbols and lines, respectively. The altitude of the waypoints was designed to follow the topography at the height of 40 m using the digital elevation model (DEM) of the survey area (2 m x 2 m grid). The maximum gap between the flight lines is about 75 m, which is 50% of the nominal line spacing. 3 Fig Waypoints designed for Ryssänlampi survey site (symbols and labels) and the location of the actual flight lines (solid and dotted lines). For clarity every second waypoint is omitted. The red rectangle depicts the borders of the survey area. Background map National Land Survey of Finland Maps of the measured and processed data are shown in Figs Figure 3.2 shows the UAV orientation as an image map. The meaning of the orientation parameters, roll, pitch and yaw, is illustrated in Figure 3.7. Roll becomes visible when the UAV turns from one line to another. Pitch shows up when the plane descends or ascends due to topography. Yaw is mostly 45 or 225 degrees because the light lines are directed towards NE or SW. 4 Fig Roll, pitch and yaw from Ryssänlampi survey. 5 Fig Height (from ground), altitude (above sea level) and DEM topography from Ryssänlampi survey. 6 Fig Raw magnetic X, Y and Z data components from Ryssänlampi survey. 7 Fig Corrected magnetic X, Y and Z data components from Ryssänlampi survey. 8 Fig Raw and corrected (base station adjusted) total magnetic field from Ryssänlampi survey. Fig Definition of roll, pitch and yaw (https://en.wikipedia.org/wiki/flight_dynamics_(fixedwing_aircraft)). 9 Because the raw magnetic data (Fig. 3.4) are defined on the reference frame of the UAV, the raw XYZ components are almost useless before the data are rotated using measured roll, pitch and yaw angles. The corrected data (Fig. 3.5) contain still some effects arising from the direction of the flight lines, indicating that either the flux-gate calibration or the orientation correction has not worked ideally due to heading related error. Therefore, the data interpretation or modelling should be based on the total field data alone. Please, note that the corrected total field data (Fig. 3.6) include base station and heading correction but not tie line levelling. 3.1 Auxiliary magnetic data Auxiliary magnetic data were measured on Aug 20th 2015 (a day after the UAV-MEMO demonstration flight). The auxiliary data consist of 1) tie lines measured perpendicularly to the original flight lines and 2) two surveys measured above the main magnetic anomaly at an intermediate altitude of 70 m and high altitude of 130 m. The location of the auxiliary flight lines (and the original lines) is shown in Figure 3.8. The total magnetic field data are shown in Figure 3.9. Please, note the different axis limits used in the top figure and the decrease in the anomaly amplitude with increasing altitude. Tie line data were measured to level the original data with the auxiliary 70 m and 130 m altitude data for the equivalent layer modelling (ELM) discussed later. Tie line survey was also needed because the base station was situated in a different place than earlier. However, tie line levelling was not performed because the heading correction (clover leaf test) was not made successfully. Instead, the ELM inversion algorithm was modified so that separate value of magnetic field base level could be given for each dataset measured at height levels 40, 70 and 130 m. 10 Fig Location of the auxiliary flight lines measured in Ryssänlampi survey. Tie lines are drawn by solid black lines, data measured at 70 m and 130 m altitude are shown by red and blue lines, respectively. Original flight lines are depicted by thin dotted lines. The red rectangle depicts the limits of the survey area. 11 Fig Corrected total magnetic field along the tie lines (top) and above the main anomaly at 70 m altitude (middle) and at 130 m altitude (bottom). 12 4. Data processing The data processing was done using interactive RadaiPros software by Markku Pirttijärvi. The basic data processing steps that are executed on the raw data are listed below. 1. Removal of dummy/missing data values 2. Removal/averaging of points with identical coordinates 3. Verification of all angle, pressure and time values 4. Computation of barometric height from pressure data 5. Computation of rectangular X and Y coordinates (ETRS-TM35FIN) 6. Computation of profile distance coordinate and azimuth/heading angle 7. Application of flux-gate calibration parameters (derived from a separate calibration measurement) 8. Correction of magnetic X, Y, Z components for UAV orientation (roll, pitch and yaw). 9. Computation of the corrected total magnetic field 10. GPS lag correction (not used because test measurement was not made) 11. Computation of IGRF reference values (B x, B y, B z, B tot, inclination and declination) After initial processing, each dataset was manually edited and unnecessary points (e.g. when the UAV travels to the survey area and away from there) were cut away. After data editing, the total number of data points is 50386, total profile length is 70.9 km and the mean velocity is 12.0 m/s. The overall time to measure the four flight areas was about 3.5 hours. The final data contains about 77 minutes of data. Some statistical information regarding the survey are given in Tables 2 and 3. The estimates of the total magnetic field (B tot ) are given for the corrected data. Table 2. Flight statistics on Ryssänlampi magnetic survey Parameter Area 1 Area 2 Area 3 Area 4 Total Points Length (m) Fly-time (min) Mean speed (m/s) Mean sampling (m) Mean height (m) Min B tot (nt) Max B tot (nt) Mean B tot (nt) Table 3. Flight statistics on the Ryssänlampi auxiliary magnetic data Parameter Tie lines z = 70 m z = 130 m Total Points Length (m) Fly-time (s) Mean speed (m/s) Mean sampling (m) Mean height (m) Mean B tot (nt) Determination of the flight altitude is based on barometric pressure. The reference point is where the UAV system is initialized before the take-off. Topographic DEM data provided by the National Land Survey of Finland is used to compute the actual height from the ground and the altitude from sea-level for each data point. The calibration of the flux-gate sensor is based on the method described in Merayo et al. (2000). The base station data have been utilized to account for the changes in the magnetic field during the survey. Likewise, pressure and temperature data could be used to correct the barometric height, but in this case, the change in pressure was found to be insignificant to correct the heights. Base station data were measured with a proton precession magnetometer at a time interval of 30 seconds. Figure 4.1 shows the base station data and the duration of the four flights on the August 7th. The strongest magnetic disturbance of about 250 nt occurred during the flight for area 2. Before the correction, the median filter was applied to the magnetic field to remove outliers. The effect of base station correction can be observed by comparing the two maps in Figure 3.6. The gap between time interval s happened due to a loss of a 12 V battery. The time used in RadaiPros and Figure 4.1 is the time in seconds from the midnight. Fig Magnetic total field (red dots) recorded at the base station on August 7th The black line represents (7-point) median-filtered magnetic field that is used in the base station correction. The blue lines indicate the duration of the four flights on Aug 7th Inspection of the corrected total magnetic field data reveals level differences between lines measured towards NE and SW. Therefore, a heading correction was applied. Normally, the heading correction is based on level differences obtained from special calibration flight (clover leaf test) high over a magnetically stable area. In this case, however, the heading test measurement failed and offsets (+60 nt for NE an -60 nt for SW) were established on a trial-and-error basis by visual inspection of the data. For a similar reason, GPS lag time correction was not made at all because there was no suitable data for that purpose. 14 5. Equivalent layer modelling (ELM) Equivalent layer modelling (ELM) is used to grid the total field data at a constant elevation level using a physical modelling of the data. The main advantage of ELM is that it removes effects of varying flight altitude and uneven sampling of the data points. ELM is a process that begins with data harvesting, which reduces the number of data by discarding uninformative data points. Harvesting is necessary because the original total amount of data is usually so large that practical inversion would be impossible. In case of Ryssänlampi, data reduction was about 92% (from to 3900 points). The next step in ELM is the generation of a 3D susceptibility model that consists of a single layer of rectangular prisms. The horizontal size of the elements was set 40 m 40 m and the vertical height was 400 m. The depth to the top of the model was 20 m and the terrain topography was not taken into account. Margin elements are added to the sides of the model to move the effect of the model edges away from the bordering data points. The total number of elements was = 3782, but only 1826 of them were free in the inversion. After setting up the data and the model, numerical inversion was used to optimize the susceptibility of the model elements so that the synthetic response of the 3D layer model fits the UAV measured magnetic total field data. The forward computation is based on the solution for dipping magnetized prism by Hjelt (1972). The unconstrained inversion method has been described in Pirttijärvi (2003). The constrained inversion method, which aims to minimize the model roughness together with the data misfit, is based on the Occam inversion method used in Grablox2 software (Pirttijärvi, 2014b). In case of Ryssänlampi, the modelling was executed using the unconstrained (SVD) inversion method. The final RMS-error (computed from the normalized data) is 6.9%. The model obtained from the inversion was then used to compute the magnetic field on a constant altitude on a regular (20 m by 20 m) grid. The intensity and the direction of the inducing magnetic field were derived from IGRF field (B 0 =53292 nt, I=76.9, D=11.2 ) computed using a year value of After inversion, the response (X, Y, Z components and total field) was also computed at the original data locations and saved to processed UAV data file to be used for comparisons. Figures 5.1a 5.1f show the fit between measured and modelled data for each dataset as profile plots. In general, the fit is rather good. The lowest minima and short wavelength anomalies are not fitted as well as maxima, because of the rather large size of the model elements (40 m by 40 m). It should be reminded that the data used in ELM is un-filtered. As a matter of fact, the ELM performs like a low-pass filter that removes high frequency content from the data. The 130 m high altitude dataset has the greatest noise and is not fitted as well as the other datasets. The cause of the noise is probably related to the stronger wind at the higher altitude. Interestingly, the base level of the magnetic field is biggest at the highest altitude (40 m= nt, 70 m = nt, 130 m = nt). One possible reason is the decrease in the amplitudes of the anomaly minima around the main target when the flight altitude increases. 15 Fig. 5.1a. The fit between measured (red) and modelled (blue) magnetic total field for dataset 1. Fig. 5.1b. The fit between measured (red) and modelled (blue) magnetic total field for dataset 2. Fig. 5.1c. The fit between measured (red) and modelled (blue) magnetic total field for dataset 3. 16 Fig. 5.1d. The fit between measured (red) and modelled (blue) total field for dataset 4. Fig. 5.1e. The fit between measured (red) and modelled (blue) total field for the 70 m altitude dataset. Fig. 5.1f. The fit between measured (red) and modelled (blue) total field for the 130 m altitude dataset. 17 Fig Measured (harvested) data (top), computed total field (middle) and the difference between them (bottom) from Ryssänlampi equivalent layer modelling (ELM). The distribution of the 2386 harvested data points (on the 40 m flight level) are shown in the bottom map. 18 Figure 5.2 shows the measured and modelled data and the data misfit as image maps. To make the comparison easier, the color scale is the same in the first two maps. The range of the misfit is rather large (c. ±300 nt), the mean and of the median being -24 nt and -15 nt, respectively. The standard deviation of the data misfit, 70.1 nt, can be considered as a measure of the noise in the data. Despite of the rather good visual fit between the measured and computed data (cf. Fig. 5.1), the statistical analysis gives rather poor estimate for the fit. The biggest errors are located where the auxiliary data was measured. When the inversion was done using the low altitude (40 m) data alone, the standard deviation of the misfit was only 58 nt, and the misfit was distributed more evenly. Considering that the base station recording interval was rather sparse (30 s), the diurnal variation of the magnetic field might have too large (Fig. 4.1) during the second flight. As a consequence, the second dataset should have been measured again. Figure 5.3 shows the magnetic susceptibility model obtained from the equivalent layer modelling. Because the depth to the top of the model is 20 m and the height of the model is rather small (400 m), the interpreted values of magnetic susceptibility are overestimated. Full 3D inversion would be needed to estimate the true susceptibility distribution. Fig Map of the magnetic susceptilibility obtained from equivalent layer modelling (ELM). The dots depict the location of harvested data points used in ELM. 19 6. Results The results of Ryssänlampi magnetic survey are presented in Figures The results are provided in digital format as: a) A column formatted (UAV) text file containing the UAV data at the original irregular data locations, b) Three ESRI/ARC Ascii (ASC) grid files containing total magnetic field computed on an even grid (20 m x 20 m) on a constant height level of 0, 20 and 40 m above surface, and c) A column formatted (XYZ) text file containing the magnetic field computed at the height of 0, 20 and 40 m at the same locations as the GTK measured ground data. Fig Comparison between the corrected (left) and the computed (right) magnetic total field at original data locations for the Ryssänlampi survey site. Fig Magnetic total field for Ryssänlampi survey area computed at the constant height of 40 m using ELM. 20 Fig Magnetic total field for Ryssänlampi survey area computed at the height of 20 m using ELM. The small dots depict the location of the harvested data points. Fig Individual X, Y, and Z components of the magnetic field and total magnetic field computed at original data locations using ELM. 21 The processed data is saved in a special column formatted UAV text file format of RadaiPros. An example of the file header and the first two rows are given below (note that the data lines are so long line that they continue on the next line). Ryssänlampi UAV data ! NPRO, NTOT, VERS ! YEAR, LON, LAT 1 0 0! ALTI, BSTAT, TIELINE ! NPTS_i, i, BASE_i, XB_i, YB_i Time Pres Roll Pitch Yaw Temp Curr QC Lon Lat Alt1 RawX RawY RawZ RawT X_TM25 Y_TM35 Alt2 Topo Dist Azim MagB CorX CorY CorZ CorT ComX ComY ComZ ComT The first line is a header line which defines the data. The second line defines the number of profiles (NPRO), total number of data points (NTOT), and file version number (VERS). The third line defines the decimal year of the measurements (YEAR) and the mean longitude (LON) and latitude (LAT) of the survey site. This information is used to compute the IGRF reference field. The fourth line contains on= 1/off= 0 flags that define whether or not flight height has been corrected using DEM model (ALTI), or base station correction (BSTAT), or tie-line levelling (TIELINE) has been made. The next line is defined separately for each dataset (survey flights). It defines the number of points (NPTS_i), and the base pressure (BASE_i), and x and y coordinates (XB_i, YB_i) of the reference point used to define the heights from the groun
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!