deletions | additions       diff --git a/abstract.tex b/abstract.tex  index 0d1b4b3..2ba01bd 100644  --- a/abstract.tex  +++ b/abstract.tex 
...
Widespread applications, from surveillance to search and rescue operations, make Micro Air Vehicles (MAVs) flexible platforms.  However, due to their small size, MAVs have limited processing power and cannot  fall back to standard localization techniques. techniques, such as laser range finders.  To address this issue, this thesis describes an efficient vision-based onboard localization  technique. Using a machine  learning approach, $x,y$-coordinates are estimated within a known, modifiable indoor  environment. The computational power of the approach is scalable to to different platforms, trading off speed and accuracy.  The development, software and hardware  implementation, and results of the localization system are presented. The proposed system uses employs  textons as image features. Textons are small, small  characteristic image patches and patches;  their histograms are used as features for the $k$-Nearest Neighbors ($k$NN) algorithm. The outputs of this regression technique--multiple possible  $x, y$-coordinates are used in a particle filter to neatly aggregate  and smooth the estimates. The estimates of the computer vision-based  system are compared to the ground truth in five on-ground and two  in-flight experiments. Promising %Promising  results were obtained for all tasks: waypoint %waypoint  navigation, accurate landing, and stable hovering in the indoor %indoor  environment. To predict the performance of the proposed algorithm in different environments, an evaluation  technique is developed that compares actual histogram similarities to  ideal histogram similarities based on the  distance between the underlying positions. The technique assigns a loss value to a given set of  images, allowingto compare different environments. Therefore, it  allows  for spotting difficult or ambiguous locations or safe landing  spots. Additionally, a comparing different environments and positions withing the environment. A software  toolis presented that  creates synthetic images that could be taken during an actual flight. The Using 46 high-resolution images, the  synthetic flight  images are used to compare 46 possible maps---images with a high resolution. The  best their potential as ``map''.  %The  %best  one, the worst one and the one with median loss were printed out to %to  compare their performance in the real-world. In fact, the results could %could  be replicated on these maps. The approach is based on three pillars: (i) a shift of processing power to an pre-flight phase to pre-compute  computationally complex steps, (ii) lightweight and adaptable algorithms to ensure real-time performance and portability to different platforms,  (iii) modifiable environments that can be tailored to the proposed algorithm. The estimates of the proposed system are compared to the ground truth in five on-ground and two  in-flight experiments with promising results.