In this paper, we present a machine learning based quantitative method for the interpretation of signals gathered from non-destructive-testing (NDT) of steel pipelines via a semi-autonomous in-line-inspection (ILI) robot. The robot has a magnetic-flux-leakage (MFL) sensor that produces three axis data for each point of pipeline with specific intervals. Both the robot and the MFL sensor have been developed in-house. The signals collected via MFL sensor are converted into images to be used as input for the proposed defect detection model. We propose a combination of a defect detection model based on SwinYv5 object detection algorithm and a quantification model based on Cross-Residual Convolutional Neural Network (CR-CNN). The detected defect locations are used to extract the Region of Interest (ROI) images of defects that are used as input for the quantification model. In data collection phase, numerous tests have been conducted via a special test mechanism and a custom data augmentation technique has been deployed in order to increase the amount and variety of training data. According to test results, the proposed method is capable of detecting defects with a precision of 98.9% and quantifying them with maximum errors of 1.30, 1.65 and 0.47 mm for length, width and depth respectively.

This paper presents a study of the feasibility of using in-line inspection (ILI) techniques with long-range communication-capable robotic systems deployed with advanced inspection sensors in natural gas distribution pipelines, which are rare in the literature. The study involved selecting appropriate antennas and determining the appropriate communication frequency for an ILI robot operating on Istanbul 12″ and 16″ steel pipelines. The paper identifies the frequency windows with low losses, presents received signal strength indicator (RSSI) and signal-to-noise ratio (SNR) information for various scenarios, and evaluates the impact of T-junctions, which are known to be the worst components in terms of communication. To utilize the pipeline as a waveguide, low-attenuation-frequency windows were determined, which improved communication by a factor of 500 compared to aerial communication. The results of laboratory tests on a 50 m pipeline and real-world tests on a 2.4 km pipeline indicate that long-distance communication and video transmission are possible at frequencies of around 917 MHz with low-gain antennas. The study also assessed the impact of the early diagnosis of anomalies without incidents on the environment, achievable with ILI robots using long-range wireless communication.

Görüntüler üzerinden hata ve anomalilerin otomatik tespiti, özellikle demiryolu taşımacılığı için kestirimci bakım uygulamalarında gün geçtikçe daha fazla yer bulmaktadır. Geleneksel bakım teknikleri, trenler için tam olarak güvenilemeyecek manuel çözümler içerirken; günümüzde görüntü işleme sayesinde daha sağlam çözümler üretilmektedir. Bu sayede bakım maliyetleri ve bakım süreleri de azalmaktadır. Trenler uzun mesafeler kat ettiğinden, trendeki sistemlerin yol boyunca zarar görmesi çok olasıdır. Hasar görme riski yüksek olan bu sistemlerden bir tanesi de pantograf-katener sistemidir. Pantografın katener ile temas eden bölgesi karbon şerittir ve katener sistemdeki elektriği trene iletmesi sebebiyle trene bağlı bileşenlerden en önemlilerinden bir tanesidir. Bu öneminden dolayı, karbon şeridi ağaç dalları gibi hasar verebilecek dış etkenlerden korumak için pantograf boynuzlarına ihtiyaç vardır. Karbon şerit yerine pantograf boynuzları hasar gördüğü için pantograf boynuzlarının bakıma ihtiyacının olup olmadığının bilinmesi önemlidir. Hasarlı boynuzların tespiti, görüntü işleme ve derin öğrenme tekniklerinden yararlanılarak otomatik olarak yapılabilir. Bu çalışmada, farklı çevresel koşullarda farklı hatalara sahip olan ve olmayan pantograf boynuzlarını içeren 34 adet simülasyon videosu oluşturulmuştur. Bu videolardan elde edilen kareler ayrı ayrı işlenmiştir. Temel görüntü işleme teknikleri sayesinde pantograf dışındaki arka plan nesneleri kaldırılmış, derin öğrenme modeli ile pantograf boynuz bölgeleri tespit edilmiş ve evrişimli sinir ağı sayesinde bir boynuzun hatalı olup olmadığı sınıflandırılmıştır. Oluşturulan metot ile bir pantograf boynuzunun hatalı veya sağlıklı olmasına %95,36 oranında doğru karar verilmektedir. Yüksek doğrulukla çalışan bu yeni yöntem, pantograf boynuz analizi üzerine literatüre önemli bir katkı sağlamaktadır.

In this work, a pedestrian detection method based on adaptive boosting is proposed. The proposed method works on still images. The features utilized in the work are derived from Haar-like templates. An Adaboost classifier is utilized for both feature selection and classification. To show the effectiveness of the proposed algorithm, the system is trained by using Nicta Pedestrian Dataset and tested by using Penn Fudan Pedestrian Dataset. The experimental result shows the proposed method's effectiveness.

In this work, a novel classifier which has the ability of making binary classifications by supervised learning is introduced. The proposed classifier generates a finite state machine which is derived from the dataset used for training. The states of this machine show the likelihood that the visiting samples belong to one of the classes concerned. Learning process is realized by recording the states mostly visited by the samples of a class. And the classification is done by examining the states visited by the sample being tested are mostly visited by the samples of which class during the training. In the realized experiments, it has been seen that the proposed classifier makes as successful classifications as artificial neural networks and binary classification trees. The main advantages of the proposed classifier are that it requires fewer iterations than neural networks and does not require rules such as splitting criterion as in binary decision trees. Its another advantage is that learning can be performed while data is flowing in one direction, without the need for a back-propagation method such as used in neural networks.