Senthan Mathavan

Pavement crack detection using the Gabor filter

Crack is a common form of pavement distress and it carries significant information on the condition of roads. The detection of cracks is essential to perform pavement maintenance and rehabilitation. Many of the highways agencies, in different countries, are still employing conventional, costly and very time consuming techniques which involve direct human intervention and assessment. Although automated recognition has been successfully performed for many pavement distresses, crack detection remains, to this date, a topic where reservations exist. A novel approach to automatically distinguish cracks in digital pavement images is proposed in this paper. The Gabor filter is proven to be a highly potential technique for multidirectional crack detection that was not done previously using the Gabor filter. Image analysis using the Gabor function is directly related to the mammalian visual perception, hence the choice of this method for crack detection. Results reported in this paper concentrate on pavement images with high levels of surface texture that makes crack detection difficult. An initial detection precision of up to 95% has been reported in this paper showing a good promise in the proposed method.

Metrology and visualization of potholes using the microsoft kinect sensor

Pavement distress and wear detection is of prime importance in transportation engineering. Due to degradation, potholes and different types of cracks are formed and they have to be detected and repaired in due course. Estimating the amount of filler material that is needed to fill a pothole is of great interest to prevent any shortage or excess, thereby wastage, of filler material that usually has to be transported from a different location. Metrological and visualization properties of a pothole play an important role in this regard. Using a low-cost Kinect sensor, the pavement depth images are collected from concrete and asphalt roads. Meshes are generated for better visualization of potholes. Area of pothole is analyzed with respect to depth. The approximate volume of pothole is calculated using trapezoidal rule on area-depth curves through pavement image analysis. In addition pothole area, length, and width are estimated. The paper also proposes a methodology to characterize potholes.

Smart irrigation using low-cost moisture sensors and XBee-based communication

Deficiency in fresh water resources globally has raised serious alarms in the last decade. Efficient management of water resources play an important role in the agriculture sector. Unfortunately, this is not given prime importance in the third world countries because of adhering to traditional practices. This paper presents a smart system that uses a bespoke, low cost soil moisture sensor to control water supply in water deficient areas. The sensor, which works on the principle of moisture dependent resistance change between two points in the soil, is fabricated using affordable materials and methods. Moisture data acquired from a sensor node is sent through XBEE wireless communication modules to a centralized server that controls water supply. A user-friendly interface is developed to visualize the daily moisture data. The low-cost and wireless nature of the sensing hardware presents the possibility to monitor the moisture levels of large agricultural fields. Moreover, the proposed moisture sensing method has the ability to be incorporated into an automated drip-irrigation scheme and perform automated, precision agriculture in conjunction with de-centralized water control.

A Review of Three-Dimensional Imaging Technologies for Pavement Distress Detection and Measurements

With the ever-increasing emphasis on maintaining road assets to a high standard, the need for fast accurate inspection for road distresses is becoming extremely important. Surface distresses on roads are essentially three dimensional (3-D) in nature. Automated visual surveys are the best option available. However, the imaging conditions, in terms of lighting, etc., are very random. For example, the challenge of measuring the volume of the pothole requires a large field of view with a reasonable spatial resolution, whereas microtexture evaluation requires very accurate imaging. Within the two extremes, there is a range of situations that require 3-D imaging. Three-dimensional imaging consists of a number of techniques such as interferometry and depth from focus. Out of these, laser imagers are mainly used for road surface distress inspection. Many other techniques are relatively unknown among the transportation community, and industrial products are rare. The main impetus for this paper is derived from the rarity of 3-D industrial imagers that employ alternative techniques for use in transportation. In addition, the need for this work is also highlighted by a lack of literature that evaluates the relative merits/demerits of various imaging methods for different distress measurement situations in relation to pavements. This overview will create awareness of available 3-D imaging methods in order to help make a fast initial technology selection and deployment. The review is expected to be helpful for researchers, practicing engineers, and decision makers in transportation engineering.

Use of a Self-Organizing Map for Crack Detection in Highly Textured Pavement Images

AbstractA study on using an unsupervised learning technique, called a self-organizing map (SOM) or Kohonen map, for the detection of road cracks from pavement images is described in this paper. The main focus is on highly textured road images that make the crack detection very difficult. Road images are split into smaller rectangular cells, and a representative data set is generated for each cell by analyzing image texture and color properties. Texture and color properties are combined with a Kohonen map to distinguish crack areas from the background. Using this technique, cracks are detected to a precision of 77%. The algorithm also resulted in a recall of 73% despite the background having very strong visual texture. The technique applied here shows a great deal of promise despite the images being captured in an uncontrolled environment devoid of state-of-the-art image-acquisition setups. The results are also benchmarked against an advanced algorithm reported in a recent research paper. The benchmarking ...

Application of Texture Analysis and Kohonen Map for Region Segmentation of Pavement Images for Crack Detection

The first phase of a research study on detecting cracks in pavements is described. For reliable crack detection, various regions in a road image have to be segmented accurately. A procedure based on the texture and color properties of different regions of images is used in conjunction with the Kohonen map, also known as the self-organizing map. Accuracy of 89.7% was obtained with classification based on the Kohonen map of images taken with a regular digital camera and simple lighting setup. Furthermore, a complementary algorithm is described to remove spurious classifications caused by inaccuracies in the trained Kohonen map. With the help of this algorithm, an overall segmentation accuracy of 97.7% is reported. This research is expected to affect other problems in transportation engineering, such as road boundary detection and road marking inspection. The detection of cracks from the segmented regions will be addressed in the future.

Pavement Raveling Detection and Measurement from Synchronized Intensity and Range Images

Raveling on asphalt surfaces is a loss of fine and coarse aggregates from the asphalt matrix. The severity of raveling can be an indicator of the state of pavements, as excessive raveling not only reduces the ride quality but eventually leads to pothole formation or cracking. Hence, raveling must be detected and quantified. In this study and for the first time, raveling was quantified from a combination of two- and three-dimensional images. First, a texture descriptor method called Laws’ texture energy measure was used in conjunction with Gabor filters and other morphological operations to distinguish road areas. Then, digital signal processing techniques were used to detect and to quantify raveling. Hundreds of images captured by an automated pavement surveying system were used to test and to show the promise of the proposed algorithm.

Fast segmentation of industrial quality pavement images using Laws texture energy measures and k-means clustering

Abstract. Thousands of pavement images are collected by road authorities daily for condition monitoring surveys. These images typically have intensity variations and texture nonuniformities that make their segmentation challenging. The automated segmentation of such pavement images is crucial for accurate, thorough, and expedited health monitoring of roads. In the pavement monitoring area, well-known texture descriptors, such as gray-level co-occurrence matrices and local binary patterns, are often used for surface segmentation and identification. These, despite being the established methods for texture discrimination, are inherently slow. This work evaluates Laws texture energy measures as a viable alternative for pavement images for the first time. k-means clustering is used to partition the feature space, limiting the human subjectivity in the process. Data classification, hence image segmentation, is performed by the k-nearest neighbor method. Laws texture energy masks are shown to perform well with resulting accuracy and precision values of more than 80%. The implementations of the algorithm, in both MATLAB® and OpenCV/C++, are extensively compared against the state of the art for execution speed, clearly showing the advantages of the proposed method. Furthermore, the OpenCV-based segmentation shows a 100% increase in processing speed when compared to the fastest algorithm available in literature.

Energy prediction of a combined cycle power plant using a particle swarm optimization trained feedforward neural network

Combined cycle power plants are frequently used for power production. Predicting the power plant output based on operational parameters is in major focus nowadays. The paper presents a novel approach using a particle swarm optimization trained feedforward neural network to predict power plant output. It takes ambient temperature, atmospheric pressure, relative humidity, and vacuum as input parameters to a feedforward neural network to predict average hourly output of the power plant. PSO is used as a learning algorithm. The MSE for training data is found to be 1.019e-04 and 0.005 for testing data. The proposed technique shows promising results to predict power plant output using a PSO trained neural network.

Tiled fuzzy Hough transform for crack detection

Surface cracks can be the bellwether of the failure of any component under loading as it indicates the component’s fracture due to stresses and usage. For this reason, crack detection is indispensable for the condition monitoring and quality control of road surfaces. Pavement images have high levels of intensity variation and texture content, hence the crack detection is difficult. Moreover, shallow cracks result in very low contrast image pixels making their detection difficult. For these reasons, studies on pavement crack detection is active even after years of research. In this paper, the fuzzy Hough transform is employed, for the first time to detect cracks on any surface. The contribution of texture pixels to the accumulator array is reduced by using the tiled version of the Hough transform. Precision values of 78% and a recall of 72% are obtaining for an image set obtained from an industrial imaging system containing very low contrast cracking. When only high contrast crack segments are considered the values move to mid to high 90%.