Research


Image-based Retro-reflectivity Measurement of Traffic Signs in a Daytime 
  • A vision method that remotely measures traffic sign retro-reflectivity in daytime 
  • The method simulates nighttime visibility from images taken during daytime 
  • The impact of time of day and distance on measurements are studied 
  • The method with accuracy of 95.24% is cheaper, faster & safer than current practice 
  • The method satisfies FHWA measurement requirements on accuracy and granularity
Selected Publications
(1) Balali, V., Sadeghi, M.A., and Golparvar-Fard, M. (2015). "Image-based Retro-Reflectivity Measurement of Traffic Signs in Day Time."
Provisional Patent, US Patent and Trademark Office.
(2) Balali, V., Sadeghi, M.A., and Golparvar-Fard, M. (2015). "Image-based Retro-Reflectivity Measurement of Traffic Signs in Day Time.Journal of Advanced Engineering Informatics, 29(4), 1028-1040.


Detection, Classification, and Mapping of Traffic Signs using Google Street View Images 
  • Creating comprehensive traffic sign inventories using Google Street View images. 
  • Average accuracy of 100% and 83.93% for warning and regulatory sign classification. 
  • Spatial visualization of traffic signs data on Google map/Earth/Street View. 
  • Automated and interactive traffic sign inventory queries. 
  • Visualizing the most probable location of traffic signs using a dynamic heat map.
Selected Publications
(1) Balali, V., Ashouri Rad, A., and Golparvar-Fard, M. (2015). "Detection, Classification, and Mapping of U.S. Traffic Signs Using Google Street View Images for Roadway Inventory Management.Springer Journal of Visualization in Engineering, 3(15), 1-18.
(2) Balali, V., Depwe, E., and Golparvar-Fard, M. (2015). “Multi-Class Traffic Sign Detection and Classification Using Google Street View Images. Transportation Research Board 94th Annual Meeting (TRB), Washington, D.C., USA.




Evaluation of Multi-Class Traffic Sign Detection and Classification 
  • Focusing on application of the video streams that are already collected from cameras mounted on DOT inspection vehicles in the US.
  • Presenting and evaluating the performance of three methods for multiple-class traffic sign detection, localization, and classification.
    • Extracting 2D candidate windows from existing video streams that potentially contain traffic signs – without making any prior assumption about their locations
    • Detecting the presence of signs in these 2D candidate windows
    • Classifying them into several categories of traffic signs based on their shape and color.
  • Introducing a new comprehensive benchmark dataset for US traffic signs containing more than 11,000 signs together with their manually produced ground truth. 
Selected Publications
(1) Balali, V., and  Golparvar-Fard, M. (2015). "Evaluation of Multi-Class Traffic Sign Detection and Classification Methods for U.S. Roadway Asset Inventory Management.ASCE Journal of Computing in Civil Engineering, 04015022.
(2) Balali, V., and Golparvar-Fard, M. (2015). "Recognition and Localization of Traffic Signs via 3D Image-based Point Cloud Models.Proc., ASCE International Workshop on Computing in Civil Engineering (IWCCE), 206-214, Austin, TX, USA.
(3) Balali, V., and Golparvar-Fard, M. (2014). “Video-based Detection and Classification of US Traffic Signs and Mile Markers using Color Candidate Extraction and Feature-based Recognition. Computing in Civil and Building Engineering (ICCCBE), 858-866, Orlando, FL, USA.
(4) Balali, V., Golparvar-Fard, M., and de la Garza, J.M. (2013). “Video-Based Highway Asset Recognition and 3D Localization.” Proc., ASCE International Workshop on Computing in Civil Engineering (IWCCE), 379-386, Los Angeles, CA, USA.



Scalable Non-Parametric Parsing for Segmentation and Recognition of Roadway Assets from Car Mounted Camera Video Streams 

  • A new scalable method for segmentation and recognition of roadway assets from videos
  • Average accuracy of 88.24% anf 82.02% for recognition and segmentations are reported
  • A new dataset from Interstate highway I-57 together with ground-truth are presented
  • Comparison with state-of-the-art Semantic Texton Forest segmentation method is made
  • We show leveraging motion cues and temporal consistency improves the performance
Selected Publications
(1) Balali, V., and  Golparvar-Fard, M. (2015). "Segmentation and Recognition of Roadway Assets from Car-Mounted Video Streams using Scalable Non-Parametric Image Parsing.Journal of Automation in Construction, 49, 27-39.
(2) Balali, V., and Golparvar-Fard, M. (2014). “Scalable Non-Parametric Parsing and 3D Reconstruction for Segmentation and Recognition of Low-Cost Highway Assets from Car-Mounted Video Streams.” Construction Research Congress (CRC), 120-129,  Atlanta, GA, USA.



Semantic Texton Forests Segmentation and 3D Reconstruction of Roadway Assets 
  • Taking the captured frames and using a pipeline of Structure from Motion and Multi View Stereo reconstructs a 3D point cloud model of the highway and surrounding assets
  • Using a Semantic Texton Forest classifier, each geo-registered 2D video frame at the pixel-level is segmented based on shape, texture, and color of the highway assets
  • Based on the results of the 2D segmentation and a new voting scheme, each reconstructed 3D point in the cloud is also categorized for one type of asset and is color coded accordingly.
Selected Publications
(1) Golparvar-Fard, M., Balali, V., and de la Garza, J.M. (2012). “Segmentation and Recognition of Highway assets using Image-based 3D Point Clouds and Semantic Texton Forests.” ASCE Journal of Computing in Civil Engineering, 04014023.

Dataset [Link]




Multi-Criteria Decision Making Method 
  • Integration of ELECTRE III and PROMETHEE II Multiple Criteria Decision Making (MCDM) methods with interval approach
  • Comparison of different MCDM methods in A/E/C industry
  • Selection of appropriate structural systems in building construction projects using Group Decision Support System (GDSS)
  • Selection of appropriate material, construction technique and structural system of bridges by application of MCDM methods
Selected Publications
(1) Balali, V., Zahraie, B., and Roozbahani, A. (2014). “A Comparison of AHP with PROMETHEE Family Decision Making Methods for Selection of Building Structural System.American Journal of Civil Engineering and Architecture, 2 (5), 149-159.
(2) Balali, V., Mottaghi, A., Shoghli, O., and Golabchi, M. (2014). “Selection of Appropriate Material, Construction Technique, and Structural System of Bridges by Use of Multi-Criteria Decision-Making Method.Transportation Research Record: Journal of the Transportation Research Board (TRR), No. 2431, 79-87.
(3) Balali, V., Zahraie, B., and Roozbahani, A. (2012). "Integration of ELECTRE III and PROMETHEE II Decision Making Methods with Interval Approach: Application in Selection of Appropriate Structural Systems."ASCE Journal of Computing in Civil Engineering, 28(2), 297-314.
(4) Balali, V., Zahraie, B., Hosseini, A., and Roozbahani, A. (2010). “Selecting the Appropriate Structural system: Application of PROMETHEE Decision Making Method.” Proc., 2nd International Conference on Engineering Systems and its Applications (ICESMA), American University of Sharjah, Sharjah, UAE.