BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//talks.staging.osgeo.org//foss4g-it-2023//talk//GB9YXT
BEGIN:VTIMEZONE
TZID:GMT
BEGIN:STANDARD
DTSTART:20001029T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10
TZNAME:GMT
TZOFFSETFROM:+0100
TZOFFSETTO:+0000
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000326T020000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3
TZNAME:BST
TZOFFSETFROM:+0000
TZOFFSETTO:+0100
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-foss4g-it-2023-GB9YXT@talks.staging.osgeo.org
DTSTART;TZID=GMT:20230614T151500
DTEND;TZID=GMT:20230614T153000
DESCRIPTION:This research aimed to identify important urban features for su
 stainable development in the urban landscape of Turin\, Italy\, using mach
 ine learning techniques. Specifically\, the study sought to identify physi
 cal and social elements such as buildings\, roads\, vegetation\, and open 
 land. The goal was to contribute to more sustainable urban environments. \
 nThe study employed the open-source platform QGIS and Orfeo Toolbox (OTB)\
 , a software library for processing images from Earth observation satellit
 es. OTB offers various algorithms\, including filtering\, feature extracti
 on\, segmentation\, and classification. The primary dataset used for class
 ification consisted of orthophotos with 3 RGB bands at a resolution of 25 
 cm.\nThe challenge was encountered when classifying pavement and flat roof
 s\, prevalent features in modern urban areas exhibiting similar radiometri
 c contents in the spectral domain. Flat roofs play a significant role with
 in sustainable urban environments\, as they can be utilized to install gre
 en roofs improving energy efficiency and reducing the urban heat island ef
 fect. Additionally\, in Italy\, where most old roofs are typically made of
  “terracotta” tiles\, flat roofs result being a relatively new feature
  in the urban landscape. Identifying flat roofs can\, therefore\, help mon
 itor changes in urban morphology and land use over time. \nTo address this
  challenge\, a 4th band was added as DEM (digital elevation model) exhibit
 ing a Ground Sampling Resolution of 50 cm/pix. Its main application was to
  create an integrated data set providing information on the elevation of t
 he terrain. This helped in distinguishing pavement and flat roofs based on
  their height difference. Adding the 4th band as DEM increased the dimensi
 onality and complexity of the data\, as a single pixel is now classified a
 s four inputs\, RGB and DEM. The random forest algorithm in OTB was applie
 d using pixel-based classification\, a machine-learning algorithm that com
 bines multiple decision trees to create a robust classifier.\nFive classes
  were generated for analysis using the unsupervised learning k-means algor
 ithm from OTB: buildings\, flat roofs\, roads\, vegetation\, and open land
 . These classes represent the most common urban features of the study area
 \, a linear concentration of urban settlements along major transportation 
 routes. The random forest algorithm was then trained on these classes usin
 g a subset of the integrated dataset as training data. The trained model w
 as used to classify the rest of the dataset\, resulting into the final cla
 ssification map.\nApplying the random forest algorithm on the integrated d
 ataset significantly improved accuracy\, increasing the overall classifica
 tion accuracy from 0.83 to 0.90. Notably\, the accuracy for the road class
  rose from 0.796 to 0.944\, while that for the flat roof class improved fr
 om 0.598 to 0.773. These results provide strong evidence for the effective
 ness of using open-source platforms and tools like OTB to identify urban f
 eatures sustainably. Furthermore\, adding more bands\, such as the DEM\, c
 an enhance the potential of these methods for creating more accurate and d
 etailed maps of urban environments.\nThis study departs from traditional l
 and cover and land use classification methods that rely on pixel-based cla
 ssification using only spectral information. Pixel-based classification as
 signs a single class to each pixel based on its spectral signature\, which
  may not fully capture urban features' spatial variability and heterogenei
 ty. Additionally\, discriminating between similar characteristics like pav
 ement and flat roofs requires more than just spectral information. \nIt is
  worth noting that this study focused solely on identifying urban features
 \, including buildings\, flat roofs\, roads\, vegetation\, and open land. 
 However\, suppose the goal is to identify a specific feature\, such as onl
 y roofs or roads. In that case\, the inclusion of irrelevant features in t
 he dataset may result in redundant data and decrease the overall accuracy 
 of the classification. Therefore\, future studies may need to explore more
  advanced algorithms\, such as convolutional neural networks\, to improve 
 the accuracy and efficiency of identifying specific urban features.
DTSTAMP:20260527T002401Z
LOCATION:Sala Videoconferenza @ PoliBa
SUMMARY:The use of open-source machine learning techniques for urban featur
 es extraction - Paolo Dabove
URL:https://talks.staging.osgeo.org/foss4g-it-2023/talk/GB9YXT/
END:VEVENT
END:VCALENDAR