Update image links.

Change-Id: I7068c8ba7a5e699e67a7dbbb3f1345364d5954ff

model_development.ipynb

@@ -41,7 +41,7 @@
     "* The XE121 is connected to a Raspberry Pi, running the Acconeer's exploration server.\n",
     "* The Raspberry Pi is powered by a power bank.\n",
     "\n",
-    "![Setup](doc/setup.png)\n",
+    "<img src=\"doc/setup.png\" style=\"width:30%; height:30%\">\n",
     "\n",
     "The purpose of tilting the angle of the sensor is to get sufficient energy back from both the floor directly under the sensor, but also from areas in front of the sensor.\n",
     "The reason why this is necessary will become apparent when the features of the classification algorithm are discussed in a coming section.\n",
@@ -63,7 +63,7 @@
     "\n",
     "The following picture shows the different surface types\n",
     "\n",
-    "![Surface types](doc\\surface_types.png)\n",
+    "<img src=\"doc/surface_types.png\" style=\"width:50%; height:50%\">\n",
     "\n",
     "The characteristics of the surface is as follows(from left to right):\n",
     "* Large tiles.\n",
@@ -95,7 +95,8 @@
     "As the sensor has a fairly wide HPBW, the majority of the reflected energy will come from the area directly underneath the sensor, yielding a large amplitude at the corresponding distance.\n",
     "Before the location of the amplitude is determined, the signal is smoothened using a second order Butterworth signal to get a more consistent estimate.\n",
     "\n",
-    "![Distance estimation](\\doc\\distance_to_floor_plot.png)\n",
+    "<img src=\"doc/distance_to_floor_plot.png\" style=\"width:30%; height:30%\">\n",
+    "\n",
     "\n",
     "For details, see *estimate_distance* in *utils.py*.\n",
     "\n",
@@ -109,7 +110,7 @@
     "The reason for using the variance as a metric instead of the absolute values is to alleviate the issue with strong static direct leakage when measuring close to the sensor. \n",
     "Direct leakage is the energy traveling directly from the transmitting to receiving antenna, and is present when measuring close to the sensor, resulting in a strong static component in the signal.\n",
     "\n",
-    "![Signal variance](\\doc\\signal_variance.png)\n",
+    "<img src=\"doc/signal_variance.png\" style=\"width:30%; height:30%\">\n",
     "\n",
     "For details, see *calculate_variance_at_fixed_distance* in *utils.py*.\n"
    ]