Visual Landscape Assessment with the Use of Cloud Vision API

Antalya Case

  • Ahmet BENLİAY Akdeniz Üniversity
  • Arzu ALTUNTAŞ Siirt University
Keywords: Antalya, Machine Learning, Visual Landscape Analysis, Visual Perception.


Quantifying and mapping visual character of an area is complex because of their intangibility. But landscape visual character can be captured using indicators derived from several theory-based concepts related to landscape perception.  One of the main methods to capture is to take photographs of a region and evaluate it by an expert group. However, manual content analysis and classification of large numbers of photographs can be time consuming. Therefore, an efficient tools such as Google Vision API image classification service can be used to fasten the process. The API can detect and classify multiple objects including the location of each object within the image and assign labels to images and quickly classify them into millions of predefined categories. This study compares previous landscape visual character analysis techniques and develops a method for automating content analysis of photographs with the use of an online machine learning algorithm. For this purpose, 60 photographs regarding to Antalya City evaluated by Google Vision API image classification service. At this stage, the method of evaluating the photos and capabilities for perceived landscape visual features of the API has been examined. According to the results, the API has been identified 3 main titles in landscape visual character assessment process as naturality, perceptibility and recreation value. As a result. a number of procedural recommendations for further evaluation were developed, and the advantages and disadvantages of this new evaluation technique have been discussed.