مقایسه تعداد لایه‌های تجزیه موجک تصویر به منظور کلاس‌بندی تمام‌خودکار بافت خرابی‌های روسازی آسفالتی

نوع مقاله : مقاله کامل پژوهشی

نویسندگان

گروه عمران، دانشکده مهندسی، دانشگاه فردوسی مشهد

چکیده

پیمایش خرابی‌های سطحی راه جزء مراحل اصلی فرآیند ارزیابی عملکردی روسازی محسوب می‌شود. در سال‌های اخیر، تحقیقات زیادی در حوزه توسعه روش‌های هوشمند و خودکار شناسایی خرابی‌های روسازی انجام گرفته است. اغلب این روش‌ها بر پایه فنون بینایی کامپیوتر می‌باشند. از مهم‌‌ترین عناصر تشکیل‌دهنده سیستم‌های بینایی کامپبوتر، استخراج ویژگی تصویر می‌باشد. روش‌های پردازش چندرزولوشن (چنددقته) هم‌چون تجزیه موجک، ابزاری منحصربفرد جهت آنالیز ویژگی‌های بافتی تصویر فراهم آورده است. لبه‌ها، جزئیات ساختاری بافت تصاویر خرابی‌ سطح روسازی را تشکیل داده و تعداد سطوح تجزیه تصویر توسط اعمال تبدیل موجک، نقش موثری در آشکارسازی و تفکیک‌پذیری مکانی بافت گسستگی‌های (لبه) تصویر ایفا کرده و بایستی به طور بهینه انتخاب گردد. در این پژوهش، پس از برداشت تصاویر خرابی روسازی آسفالتی در شرایط روشنایی ثابت، از سه لایه تبدیل موجک دوبعدی هار (Haar) به منظور تجزیه تصاویر و از ماتریس هم‌رخداد و ماتریس طول تکرار سطوح خاکستری، جهت توصیف آماری بافت زیرباندهای فرکانسی حاصل شده، استفاده گردید. نتایج حاصل از طبقه‌بندی تصاویر خرابی بر اساس روش کمینه فاصله ماهالانوبیس (Minimum Mahalanobis Distance)، حاکی از آن است که آماره‌های مستخرج از زیرباندهای موجک تا مرحله دوم تجزیه، در شناسایی بافت انواع خرابی نتایج برتری به دنبال داشته است. میانگین دقت عملکردی کلاس‌بندی تصاویر خرابی بر پایه تجزیه موجک تک‌مرحله‌ای، دو‌مرحله‌ای و سه‌مرحله‌ای به ترتیب برابر با 68 درصد، 79 درصد و 64 درصد می‌باشد. هم‌چنین آمارگان مرتبه دوم حاصل از ماتریس هم‌رخداد سطوح خاکستری (Grey Level Co-occurrence Matrix (GLCM))، با میانگین دقت طبقه‌بندی 81 درصد، عملکرد برتری نسبت به ماتریس طول تکرار در تشخیص و تفکیک خودکار تصاویر خرابی سطح آسفالت دارا می‌باشند.

کلیدواژه‌ها

موضوعات


عنوان مقاله [English]

Comparative Analysis of Wavelet Decomposition Levels in Order to Full-Automated Classification of Asphalt Pavement Distresses Texture

نویسندگان [English]

  • Reza Shahabian Moghaddam
  • Seyad Ali Sahaf
  • Abolfazl Mohammadzadeh Moghaddam
Civil Engineering Dept. ; Engineering Faculty; Ferdowsi University of Mashhad, Iran
چکیده [English]

Evaluation of the pavement performance plays a vital role in Pavement Management System (PMS). Identification of road surface failures (pavement distresses) is of the main elements in the process of pavement evaluation at the network and project level road assessment. The simplest method of pavement distress evaluation is visual inspection and rating of apparent condition of the road surface by experts. This approach of pavement evaluation not only increases the cost and time of inspection operations, but also depends on the personal judgment (subjective) of assessors and brings about non-repeatable results. Furthermore, it exposes the inspectors to unsafe working situations on highways. In the last decade, extensive researches have been conducted for the development of semi-automated and fully-automated methods of pavement condition evaluation to overcome the deficiencies and problems related to manual and visual evaluation of pavement distresses. Although some standards and protocols such as AASHTO R 55-10 have been developed in accordance with automatic distress data collection conditions, limited success has been achieved in recognition and classification of different pavement distresses. This can be attributed to severe disorders and irregularities of distress patterns, especially in asphalt pavements. Additionally, most of the current automatic evaluation systems are associated with heavy computational loads due to employing complex algorithms and they mainly focus on automatic detection and classification of pavement cracks. It should be noted that cracks only represent one important aspect of pavement distresses. Other types of distresses such as potholes, patching, bleeding and raveling also play an important role in the decline of pavement quality index and affect the treatments and maintenance options proposed by PMS. Considering these limitations, the authors employed a method with superior precision and efficiency to identify and process pavement distresses.

کلیدواژه‌ها [English]

  • Automatic pavement evaluation
  • Image textural analysis
  • Wavelet transform
  • Gray level co-occurrence matrix (GLCM)
  • Gray level run length matrix (GLRLM)
Arabi P, Joshi G, Deepa V, “Performance evaluation of GLCM and pixel intensity matrix for skin texture analysis”, Perspectives in Science, 2016, 8, 203-206. https://doi.org/10.1016/j.pisc.2016.03.018
Armi L, Ershad S, “Texture image analysis and texture classification methods-a review”, International Online Journal of Image Processing and Pattern Recognition, 2019, 2 (1), 1-29. https://doi.org/10.48550/arXiv.1904.06554
Bordiuness A, Khabiri, MM, “Field and laboratory study of the effect of type of emulsion bitumen and chip seal aggregate size applied on roller concrete pavement”, Journal of Civil and Environmental Engineering, 2021, 51 (102), 35-46. https://doi.org/10.22034/jcee.2020.10783
Celic T, Tjahjadi T, “Texture classification and retrieval based on complex wavelet subbands”, Computer and Information Sciences, 2011, 62, 259-264. https://doi.org/10.1007/978-90-481-9794-1_50
Chandra Prabha K, “Texture analysis using glcm & glrlm feature extraction methods”, International Journal for Research in Applied Science and Engineering Technology, 2019, 7 (11), 2059-2064. https://doi.org/10.22214/ijraset.2019.5344
Cheng HD, Glazier C, Hu YG, “Novel approach to pavement cracking detection based on fuzzy set theory”, Journal of Computing in Civil Engineering, 1999, 13 (3), 270-280. https://doi.org/10.1061/(ASCE)08873801(1999)13:4(270)
Chua KM, Xu L, “Simple procedure for identifying pavment distresses from video images”, Journal of Traportation Engineering, 1994, 120 (3), 412-431. https://doi.org/10.1061/(ASCE)0733947X(1994)120:3(412)
Lee D, “A robust position invariant artificial neural network for digital Pavement crack analysis”, Technical report, TRB Annual Meeting, 2003, Washington, DC, USA.
Moghadas Nejad F, Zakeri H, “A comparison of multi-resolution methods for detection and isolation of pavement distress”, Expert Systems with Applications, 2011, 38 (3), 2857-2872. https://doi.org/10.1016/j.eswa.2010.08.079
Moghadas Nejad F, Zakeri H, “An expert system based on wavelet transform and radon neural network for pavement distress classification”, Expert Systems with Applications, 2011, 38 (3), 7088-7101. https://doi.org/10.1016/j.eswa.2010.12.060
Moghadas Nejad F, Zakeri H, “An optimum feature extraction method based on Wavelet-Radon Transform and Dynamic Neural Network for pavement distress classification”, Expert Systems with Applications, 2011, 38 (3), 9442-9460. https://doi.org/10.1016/j.eswa.2011.01.089
Nallamothu S, Wang KCP, “Experimenting with recognition accelerator for pavement distress identification”, Transportation Research Record, 1996, 1536, 130-135.
Nourani V, Ranjbar S, Tootoonchi F, “Change detection of hydrological processes using wavelet-entropy complexity measure case study: Urmia Lake”, Journal of Civil and Environmental Engineering, 2015, 45 (80), 75-86.
Ramola A, Shakya A, Pham D, “Study of statistical methods for texture analysis and their modern evolutions”, Engineering Reports, 2020, 2, 203-206. https://doi.org/10.1002/eng2.12149
Reis M, Bauer A, “Using wavelet texture analysis in image-based classification and statistical process control of paper surface quality”, Computer Aided Chemical Engineering, 2009, 27, 1209-1214. https://doi.org/10.1016/S1570-7946(09)70592-9
Rosa P, “Automatic pavement crack detection and classification system”, Transportation Research Board, National Research Council, Washington DC, 2014, 57-65.
Selvaraj A, Ganesan L, “Texture classification using wavelet transform”, Pattern Recognition Letters, 2003, 24 (9), 1513-1521. https://doi.org/10.1016/S01678655(02)00390-2
Shahabian Moghaddam R, Sahaf A, “Automatic recognition and classification of asphalt pavement distress texture based on wavelet transform”, Road, 2021, 29 (107), 175-197. https://doi.org/10.22034/road.2021.64977.
Shahabian Moghaddam R, Sahaf A, Mohamadzadeh Moghaddam A, Pourreza HR, “A comparison of image texture analysis methods for automatic recognition and classification of asphalt pavement distresses”, Journal of Transportation Infrastructure Engineering, 2017, 3 (3), 1-22. https://doi.org/10.22075/jtie.2017.11322.1211.
Wang KCP, “Wavelet-based pavement distress image edge detection with Trous algorithm”, Transportation Research Record: Journal of the Transportation Research Board, 2009, 2024, 73-81. https://doi.org/10.3141/2024-09
Wang KCP, Li Q, J Yang G, Zhan Y, Qiu Y, “Network level pavement evaluation with 1 mm 3D survey system”, Journal of Traffic and Transportation Engineering, 2015, 2 (6), 391-398. https://doi.org/10.1016/j.jtte.2015.10.005
Zakeri H, Moghadas Nejad F, Fahimifar A, “Image based techniques for crack detection, classification and quantification in asphalt pavement: a review”, Archives of Computational Methods in Engineering, 2016, 24 (4), 1-43. https://doi.org/10.1007/s11831-016-9194-z
Zou Q, Cao Y, Li Q, Mao Q, Wang S, “Crack-tree: automatic crack detection from pavement images”, Pattern Recognition Letters, 2012, 33 (3), 227-238. https://doi.org/10.1016/j.patrec.2011.11.004