É. Rignot, J.-M Friedt, L. Moreau. 5 mars 2008

Transcription

1 About the tide oscillations and icebergs motion recorded at Jakobshavn isbrae during summer 27 using high, Icefjord, West. Discussion about acceleration of glacier flow in, lubrication at the ice-rock interface, vertical circulation of melt water (moulins) and lubrication recorded in situ under Argentire glacier (Mt-blanc). É. Rignot, J.-M Friedt, L. 5 mars 28 1 / 15

2 Automated digital image capture using a custom circuit for triggering a commercial digital low power consumption of control circuitry (< 2 µa) for an autonomy of several years consumption limits the global autonomy to theoretically 1 frames date & time stored in EXIF header for quantitative processing (problem of powering the s internal clock) Two testbeds : a 1 month long continuous sequence shot in, a 6 month long webcam archive from Chamonix, France. 2 / 15

3 Jakobshavn Selected analysis area blue red green magenta black 3 / 15

4 Measurement horizon (theory) Fixed window : correlation technique only works as long as reference and anlyzed frame look similar. Short term difference of successive frames provides poor accuracy (motion > 1 pixel between two frames) Shift reference window and connect drift curves Reference frame (fixed) + illumination artifact : apparent Y motion with 24 hour period due to sun motion 4 / 15

5 Measurement horizon (theory) Fixed window : correlation technique only works as long as reference and anlyzed frame look similar. Short term difference of successive frames provides poor accuracy (motion > 1 pixel between two frames) Shift reference window and connect drift curves Frame 1 Reference frame (fixed) + illumination artifact : apparent Y motion with 24 hour period due to sun motion 5 / 15

6 Measurement horizon (theory) Fixed window : correlation technique only works as long as reference and anlyzed frame look similar. Short term difference of successive frames provides poor accuracy (motion > 1 pixel between two frames) Shift reference window and connect drift curves Frame 1 Frame 2 Reference frame (fixed) + illumination artifact : apparent Y motion with 24 hour period due to sun motion 6 / 15

7 Measurement horizon (theory) Fixed window : correlation technique only works as long as reference and anlyzed frame look similar. Short term difference of successive frames provides poor accuracy (motion > 1 pixel between two frames) Shift reference window and connect drift curves Frame 1 Frame 2 Frame 3 Reference frame (fixed) + illumination artifact : apparent Y motion with 24 hour period due to sun motion 7 / 15

8 Motion detection (application) Basic but robust technique : intercorrelation looks for translation vector for best match of a reference picture and a measurement picture (xcorr2() with Matlab) We will use an Eulerian descritpion of fluid motion (fixed window, monitor the mass entering and leaving this frame) image(t) image(t+24) image(t+24) xcorr / 15

9 pixel/min reference.215 pixel/min frame h Measurement horizon (results) pixel/min X.193 pixel/min reference frame Y 24 h h pixel/min reference frame 7.23 pixel/min / 15

10 Influence of tide Visually, obvious influence of tide on vertical motion of ice Y motion X motion reference frame (13) high tide low tide Y motion X motion.55 pixels/min.33 pixel/min 27/7 frame 7 13/8 21/7 frame 26 22/8 frame 1 frame x reference frame 23 reference frame (33) high tide x 1 4 Y motion X motion tide amplitude (m) Y motion 5 X motion /8 frame 13 1 low tide.4 29/7.24 frame 1 15 pixel/min date (minutes since 1/1/27) x 1 5 x 1 4 Two tide amplitude maxima over 1 month record : visible signal synchronous with predicted tide max/min. No obvious signal synchronous with tide during amplitude minimum (bot. right) 1 / 15

11 Influence of wind No visible influence of wind 1 : the local influence on a given iceberg is not tracked : would require a Lagrangian description of fluid motion (object tracking) displacement (pixels) magenta green yellow red 4 wind tide amplitude x5, +3 red (ref.) magenta green yellow 6 blue 1 month 8 high tide x1, 1 blue 7 1 low tide x1, time (min.) x x 1 4 Long term vertical motions appears related to tide 2 amplitude (right) 1 Hourly METAR data collected from 2 Web interface to xtide at othernorth.html 11 / 15

12 Influence of picture quality We have demonstrated the use of high resolution (1 Mpixels), low compression images on a fast moving object. What about poor quality images of a slow glacier? red blue green magenta black X motion (pixel) time (frame number) Requires 1 interpolation of the original image to smooth JPEG compression boundaries and improve resolution 2 histogram equalization 12 / 15

13 Strong JPEG compression artifacts Influence of picture quality Matlab s imresize() bilinear interpolation (weighted average of pixels in the nearest 2-by-2 neighborhood) 13 / 15

14 Results and improvements Some frames 3 with poor weather induce high noise level : would require pre-processing to avoid noise Strong effect of shadow histogram equalization reduce influence of noise when connecting successive windows by using linear fit of glacier motion / 15

15 Results and improvements The motion is an average over a given part of the picture : the larger the picture, the longer the horizon, but the worse the average Connection of analysis periods is not always accurate Conversion from pixel to meters requires terrain model + characteristics Validation using calibrated instruments (GPS) Yet, digital image processing provides a means of continuous monitoring of ice flow in area where instuments cannot be positioned (icebergs, strong slopes) 15 / 15