Possible image processing to capture objects (segmentation) on aerial images

Possible image processing to capture objects (segmentation) on aerial images

Pixel-based:

• Thresholding
• Otsu's method
• Blob detection
• Cluster analysis e.g. k-means clustering

Pro's

• easy implementation
• quick algorithms

Con's

• no coherent segmentation or area
• results depend solely on one-dimensional parameter

Edge-based:

• Sobel operator, see also Sobel Edge Detector

• LaPlace operator

• Canny Edge Detection - Multi-stage algorithm to detect a wide range of edges in images Java Code Example

• Watershed transformation - Grey-level image interpreted as topographic relief. Grey level of pixels are interpreted as altitude in the relief. A drop of water falling on a topographic relief flows along a path to finally reach a local minimum. Intuitively, the watershed of a relief correspond to the limits of the adjacent catchment basins of the drops of water.

• Felzenszwalb-Huttenlocher algorithm (C++ Code Example included)

• Active contour model also known as Snakes

• Active-Shape-Models

• optimal edge detection & sequential edge detection (extraction), both related to Canny Edge Detection

• Livewire Segmentation Technique

Pro's

• coherent segmentation or area
• easy to find geometric representation

Con's

• edge-tracking-algorithm required that can handle gaps or bifurcations

For more information about edge-detection in general see this wikipedia article

Region-based:

• Region growing - This approach to segmentation examines neighboring pixels of initial “seed points” and determines whether the pixel neighbors should be added to the region. The process is iterated on, in the same manner as general data clustering algorithms. Java Code Example

• Split and Merge also known as Quadtree Segmentation

• Region Merging

• Pyramid Linking

Pro's

• coherent segmentation or area
• finds regions with the properties that we defined

Con's

• may cause holes or oversegmentation

Model-based

• Hough Transformation
• Template-Matching
• Statistical model

Pro's

• better results

Con's

• just suitable for special images

Texture-based

• [Markoff Random Fields](http://www.psi.toronto.edu/~vincent/research/presentations /Markov%20Random%20Fields%20and%20Texture%20Analysis.pdf)
• Coocurrence-Matrix
• Texture-Energy-Measure
• Run-Length-Matrix

Pro's

• could optimize the result in addition with other algorithms

Con's

• hard to get good results

More information

• Colour aerial image segmentation using a Bayesian homogeneity predicate and map knowledge
• Segmentation Based Classification of Aerial Images and its Potential to Support the Update of Existing Land Use Data Bases
• Automatic Segmentation of Aerial Photographs
• Segmentation optimization for aerial images with spatial Constraints
• Data mining and image segmentation approaches for classifying defoliation in aerial forest imagery
• ImprovedMapping and Image Segmentation by Using Semantic Information to Link Aerial Images and Ground-Level Information