Mathematical Morphology Operations Research Articles

The Dynamic Change of Bosten Lake Area in Response to Climate in the Past 30 Years

The widely distributed lakes, as one of the major components of the inland water system, are the primary available freshwater resources on the earth and are sensitive to accelerated climate change and extensive human activities. Lakes play an important role in the terrestrial water cycle and biogeochemical cycle and substantially influence the health of humans living in the surrounding areas. Given the importance of lakes in the ecosystem, long-term monitoring of dynamic changes has important theoretical and practical significance. Here, we extracted water body information and monitored the long-term dynamics of Bosten Lake, which is the largest inland lake in China. We quantified the meteorological factors of the study area from the observation data of meteorological stations between 1988 and 2018. The characteristics of climate change and its correlation with the change of area in the Bosten Lake Basin in the past 30 years were analyzed. The major contributions of this study are as follows: (1) The initial water body was segmented based on the water index model Normalized Difference Water Index (NDWI) and Modified Normalized Difference Water Index (MNDWI) with a pre-assigned threshold value. The results were evaluated with the area extracted through artificial visual interpretation. Then we conducted mathematical morphology operators, opening and closing operations, and median filter to eliminate noise to ensure the accuracy of water body information extraction from the Bosten Lake. A long-term water surface area database of the Bosten Lake was established from high-resolution remote sensing images during 1988–2018. (2) Due to the seasonal difference of snow, ice content, and other objects on images, the areadynamics of Bosten Lake in the recent 30 years were analyzed separately in dry season and rainy season. The water surface area of Bosten Lake showed large inter-annual variations between 1988–2018. (3) Based on the assumption that climatic change has more direct effects on lake than human activities, six meteorological factors were selected to analyze the impacts of climate change on the annual mean lake surface area. The result indicated that in the past 30 years, climate conditions in the Bosten Lake Basin fluctuated greatly. We conducted correlations analysis between the areal dynamics of the Bosten Lake and the meteorological factors. Here, the annual average evaporation had the highest correlation with the areal dynamics of Bosten Lake followed by air temperature, precipitation, sunshine hours, and relative humidity, while the annual average wind speed had the weakest correlation.

Automatic forensic identification using 3D sphenoid sinus segmentation and deep characterization.

Recent clinical research studies in forensic identification have highlighted the interest in sphenoid sinus anatomical characterization. Their pneumatization, well known as extremely variable in degrees and directions, could contribute to the radiologic identification, especially if dental records, fingerPrints, or DNA samples are not available. In this paper, we present a new approach for automatic person identification based on sphenoid sinus features extracted from computed tomography (CT) images of the skull. First, we present a new approach for fully automatic 3D reconstruction of the sphenoid hemisinuses which combines the fuzzy c-means method and mathematical morphology operations to detect and segment the object of interest. Second, deep shape features are extracted from both hemisinuses using a dilated residual version of a stacked convolutional auto-encoder. The obtained binary segmentation masks are thus hierarchically mapped into a compact and low-dimensional space preserving their semantic similarity. We finally employ the ℓ2 distance to recognize the sphenoid sinus and therefore identify the person. This novel sphenoid sinus recognition method obtained 100% of identification accuracy when applied on a dataset composed of 85 CT scans stemming from 72 individuals. Automatic Forensic Identification using 3D Sphenoid Sinus Segmentation and Deep Characterization from Dilated Residual Auto-Encoders.

Identifying the comprehensive pore structure characteristics of a rock from 3D images

Characterization of pore structure in reservoirs plays a significant role in predicting properties of rocks and classifying the reservoirs. The focus is on determining comprehensive geometric and topological parameters of pore structure. X-ray computerized tomography scanning provides a 3D image of the pore structure. However, these images cannot directly generate pore structure parameters without numerical characterization of the images. A new method is developed to determine comprehensive geometric and topological parameters of pore structure from images. These parameters were determined by utilizing common mathematical morphology operations to segment the entire pore space into a series of space blocks of different radii, subsequently identifying these blocks as pores and throats based on the proposed morphological features of pore and throat. The parallel scheme was also studied to obtain the highest efficiency. Thereafter, the characteristics of pore structures and the performance of the method were assessed for various samples. The results were compared to the results calculated by other methods. Results showed that the method produced reliable pore structure characteristics for a wide variety of rocks. Additionally, in order to discuss whether our results can be used as the basic input data for other researches, an example about two-phase flow simulation was made and a comparison of simulation results and experimental data was made. Our results were a competent choice for other studies. Also, after proposing the detection criteria of pore and throat, the subsequent detecting procedure is without artificial adjustable parameters, making the method convenient to use. This method can comprehensively, efficiently, accurately, and conveniently characterize the pore structure based only on images.

Micro-computed tomographic and SEM study of porous bioceramics using an adaptive method based on the mathematical morphological operations

Numerous clinical studies have demonstrated the influence of the size, number and shape of pores into calcium phosphate ceramics on the process of bone regeneration.The main objective of this study is to determine the microstructure, the morphological characteristics and classes of pores of the prepared hydroxyapatite bioceramic using an adaptive method based on the mathematical morphological operations. The study was carried out using X-ray microtomography and Scanning Electron Microscopy images. The conventional method of openings alone presents limitation of calculation and not sufficient to achieve our objective. The proposed method allowed us to extract local characteristics and calculate precisely the morphological parameters while preserving the original volume of pores. The number and classes of pores with their size, surface of contact of the component and the number of connected pores to each pore were calculated. The method is subjectively and quantitatively evaluated using different computed phantoms and its efficiency is clearly demonstrated through the different reports and measurements generated. The proposed method can have interesting applications in the characterization of porous materials used in the medical field or in other sectors.

A real-time approach for automatic defect detection from PCBs based on SURF features and morphological operations

This paper presents an automatic inspection approach for Printed Circuit Boards (PCBs) with accurate determination of the fault location and identification of the fault type. This approach depends on several digital image processing techniques including registration, filtering, foreground segmentation, mathematical morphological operations, subtraction, feature extraction, and component matching. The Speeded Up Robust Feature extraction (SURF) technique is used for two purposes: registration of the PCB to be checked to a reference PCB and detection of feature points of each missing component on the PCB that is localized from the subtraction process from the reference PCB. Operation is performed on the hue component of the color PCB images. A dictionary is first built for all possible components on the available PCBs with SURF feature descriptors, and hence if a missing item is detected on a PCB during the inspection process, the SURF feature descriptors for features extracted from the difference between the tested and reference PCBs at the position of the lost component are matched with those in the built dictionary or database. A distance metric is used in the matching process. The importance of the proposed approach lies in its ability to build a dictionary of feature descriptors for all possible components in a diversity of PCBs and its ability to localize and identify the missing components regardless of the PCB position, rotation, or type. All operations are formulated in a Graphical User Interface (GUI) using MATLAB environment.

A Highly Flexible Architecture for Morphological Gradient Processing Implemented on FPGA

Nowadays, image processing algorithms present essential components of advanced systems in industrial, robotic and medical applications. In most cases, there are high requirements for reduced delays and optimal processing performances. Mathematical morphological operations are one of the most popular and powerful tools used in image processing offering high-level performance operations at low design complexity. However, mathematical morphology is based on repetitive calculations for a wide range of data, resulting in high execution time and memory requirement. Hence, hardware acceleration presents one of the most appropriate solutions to overcome this limitation. In this paper, we propose an efficient hardware architecture aiming to increase performances of morphological gradient computation for grayscale images. The architecture exploits both intra-level and inter-level parallelisms to speed up calculations. In addition, it processes data on stream which decreases memory utilization. The architecture allows extracting the standard edge gradient as well as the external and internal edge gradients at the desired magnitude and thickness level. Unlike most of existing works, the proposed architecture supports reconfigurable shapes and sizes of structuring elements. It is successfully implemented on FPGA. The proposed architecture can process data at a throughput of 356 Mpx/s. Accordingly, a high frame rate for moderate size of structuring elements and high image resolution is achieved.

An automated early diabetic retinopathy detection through improved blood vessel and optic disc segmentation

This paper presents an automated early diabetic retinopathy detection scheme from color fundus images through improved segmentation strategies for optic disc and blood vessels. The red lesions, microaneurysms and hemorrhages are the earliest signs of diabetic retinopathy. This paper essentially proposes improved techniques for microaneurysm as well as hemorrhages detection, which eventually contribute in the overall improvement in the early detection of diabetic retinopathy. The proposed method consists of five stages- pre-processing, detection of blood vessels, segmentation of optic disc, localization of fovea, feature extraction and classification. Mathematical morphology operation is used for pre-processing and blood vessel detection. Watershed transform is used for optic disc segmentation. The main contribution of this model is to propose an improved blood vessel and optic disc segmentation methods. Radial basis function neural network is used for classification of the diseases. The parameters of radial basis function neural network are trained by the features of microaneurysm and hemorrhages. The accuracy of the proposed algorithm is evaluated based on sensitivity and specificity, which are 87% and 93% respectively.

A mathematical morphology based method for hierarchical clustering analysis of spatial points on street networks

Spatial clustering analysis is an important issue that has been widely studied to extract the meaningful subgroups of geo-referenced data. Although many approaches have been developed in the literature, efficiently modeling the network constraint that objects (e.g. urban facility) are observed on or alongside a street network remains a challenging task for spatial clustering. Based on the techniques of mathematical morphology, this paper presents a new spatial clustering approach NMMSC designed for mining the grouping patterns of network-constrained point objects. NMMSC is essentially a hierarchical clustering approach, and it generally consists of two main steps: first, the original vector data is converted to raster data by utilizing basic linear unit of network as the pixel in network space; second, based on the specified 1-dimensional raster structure, an extended mathematical morphology operator (i.e. dilation) is iteratively performed to identify spatial point agglomerations with hierarchical structure snapped on a network. Compared to existing methods of network-constrained hierarchical clustering, our method is more efficient for cluster similarity computation with linear time complexity. The effectiveness and efficiency of our approach are verified through the experiments with real and synthetic data sets.

A hybrid method for accurate iris segmentation on at-a-distance visible-wavelength images

This work describes a new hybrid method for accurate iris segmentation from full-face images independently of the ethnicity of the subject. It is based on a combination of three methods: facial key-point detection, integro-differential operator (IDO) and mathematical morphology. First, facial landmarks are extracted by means of the Chehra algorithm in order to obtain the eye location. Then, the IDO is applied to the extracted sub-image containing only the eye in order to locate the iris. Once the iris is located, a series of mathematical morphological operations is performed in order to accurately segment it. Results are obtained and compared among four different ethnicities (Asian, Black, Latino and White) as well as with two other iris segmentation algorithms. In addition, robustness against rotation, blurring and noise is also assessed. Our method obtains state-of-the-art performance and shows itself robust with small amounts of blur, noise and/or rotation. Furthermore, it is fast, accurate, and its code is publicly available.

Mejora de contraste de imagenes de mamografía mediante procesamiento digital de imagen

This paper shows the obtained results of applying image enhancement techniques in regions of interest extracted of mamographies. The objective is to improve the detection of breast lesions, such as the microcalcifications and thus contribute to the early detection of breast cancer. Usually, the quality of the region of interest images is not better, and they have some deficiencies that difficult the microcalcifications detection. Due to this, the approach in this paper is to increase the quality of the region of interest image by means of edge and contour enhancement, modifying the contrast and noise reduction and its possible suppression of the background of the image. The techniques that were implemented are the modification of the contrast by lineal and non-lineal transformations, the histogram equalization technique, spatial filtering technique and a technique based on mathematical morphology operations. Finally, a contrast improvement index was applied in order to evaluate in a quantitative way each of the contrast enhancement technique implemented.

Hardwood Species Classification with Hyperspectral Microscopic Images

We propose a hardwood species identification method based on wood hyperspectral microscopic images. A SOC710VP hyperspectral stereomicroscope was used to acquire microscopic images of a hardwood cross section. In these microscopic images, each part’s spectral features are discussed. We found that the spectral divisibility of wood vessels’ peripheral and central regions in the hyperspectral microscopic images can be used for hardwood species recognition. Mathematical morphological operation and K-L divergence were used to extract spectral features at the wood vessels’ peripheral regions and central regions, respectively. By comparing wood vessels’ spectral similarity across wood species samples, we found that wood vessels’ peripheral spectral divisibility is larger than its central. Finally, the spectral information from randomly selected regions of interest (i.e., ROI) and that of wood vessels’ peripheral and central regions have value as a classification basis. In our hardwood species classification experiments, three dimensionality reduction algorithms, principal component analysis (PCA), kernel principal component analysis (KPCA), and multidimensional scaling (MDS), and the three classifiers, BP neural network, support vector machine (SVM), and Mahalanobis distance (MD), are combined to perform hardwood species classification work. Experimental results indicate that the best recognition effect can be achieved at the peripheral region of wood vessels using PCA or MDS with the MD algorithm.

MOBILE LASER SCAN DATA FOR ROAD SURFACE DAMAGE DETECTION

Abstract. Road surface anomalies affect driving conditions, such as driving comfort and safety. Examples for such anomalies are potholes, cracks and ravelling. Automatic detection and localisation of these anomalies can be used for targeted road maintenance. Currently road damage is detected by road inspectors who drive slowly on the road to look out for surface anomalies, which can be dangerous. For improving the safety road inspectors can evaluate road images. However, results may be different as this evaluation is subjective. In this research a method is created for detecting road damage by using mobile profile laser scan data. First features are created, based on a sliding window. Then K-means clustering is used to create training data for a Random Forest algorithm. Finally, mathematical morphological operations are used to clean the data and connect the damage points. The result is an objective and detailed damage classification. The method is tested on a 120 meters long road data set that includes different types of damage. Validation is done by comparing the results to a classification of a human road inspector. However, the damage classification of the proposed method contains more details which makes validation difficult. Nevertheless does this method result in 79% overlap with the validation data. Although the results are already promising, developments such as pre-processing the data could lead to improvements.

An improved vessel extraction scheme from retinal fundus images

Vessel extraction from retinal fundus images is essential for the diagnosis of different opthalmologic diseases like glaucoma, diabetic retinopathy and hypertension. It is a challenging task due to presence of several noises embedded with thin vessels. In this article, we have proposed an improved vessel extraction scheme from retinal fundus images. First, mathematical morphological operation is performed on each planes of the RGB image to remove the vessels for obtaining noise in the image. Next, the original RGB and vessel removed RGB image are transformed into negative gray scale image. These negative gray scale images are subtracted and finally binarized (BW1) by leveling the image. It still contains some granular noise which is removed based on the area of connected component. Further, previously detected vessels are replaced in the gray-scale image with mean value of the gray-scale image and then the gray-scale image is enhanced to obtain the thin vessels. Next, the enhanced image is binarized and thin vessels are obtained (BW2). Finally, the thin vessel image (BW2) is merged with the previously obtained binary image (BW1) and finally we obtain the vessel extracted image. To analyze the performance of our proposed method we have experimented on publicly available DRIVE dataset. We have observed that our algorithm have provides satisfactory performance with the sensitivity, specificity and accuracy of 0.7260, 0.9802 and 0.9563 respectively which is better than the most of the recent works.

Brain tumor detection using statistical and machine learning method

Background and ObjectiveBrain tumor occurs because of anomalous development of cells. It is one of the major reasons of death in adults around the globe. Millions of deaths can be prevented through early detection of brain tumor. Earlier brain tumor detection using Magnetic Resonance Imaging (MRI) may increase patient's survival rate. In MRI, tumor is shown more clearly that helps in the process of further treatment. This work aims to detect tumor at an early phase. MethodsIn this manuscript, Weiner filter with different wavelet bands is used to de-noise and enhance the input slices. Subsets of tumor pixels are found with Potential Field (PF) clustering. Furthermore, global threshold and different mathematical morphology operations are used to isolate the tumor region in Fluid Attenuated Inversion Recovery (Flair) and T2 MRI. For accurate classification, Local Binary Pattern (LBP) and Gabor Wavelet Transform (GWT) features are fused. ResultsThe proposed approach is evaluated in terms of peak signal to noise ratio (PSNR), mean squared error (MSE) and structured similarity index (SSIM) yielding results as 76.38, 0.037 and 0.98 on T2 and 76.2, 0.039 and 0.98 on Flair respectively. The segmentation results have been evaluated based on pixels, individual features and fused features. At pixels level, the comparison of proposed approach is done with ground truth slices and also validated in terms of foreground (FG) pixels, background (BG) pixels, error region (ER) and pixel quality (Q). The approach achieved 0.93 FG and 0.98 BG precision and 0.010 ER on a local dataset. On multimodal brain tumor segmentation challenge dataset BRATS 2013, 0.93 FG and 0.99 BG precision and 0.005 ER are acquired. Similarly on BRATS 2015, 0.97 FG and 0.98 BG precision and 0.015 ER are obtained. In terms of quality, the average Q value and deviation are 0.88 and 0.017. At the fused feature based level, specificity, sensitivity, accuracy, area under the curve (AUC) and dice similarity coefficient (DSC) are 1.00, 0.92, 0.93, 0.96 and 0.96 on BRATS 2013, 0.90, 1.00, 0.97, 0.98 and 0.98 on BRATS 2015 and 0.90, 0.91, 0.90, 0.77 and 0.95 on local dataset respectively. ConclusionThe presented approach outperformed as compared to existing approaches.

Image-Based Abnormal Data Detection and Cleaning Algorithm via Wind Power Curve

This paper proposes an image-based algorithm for detecting and cleaning the wind turbine abnormal data based on wind power curve (WPC) images. The abnormal data are categorized into three types, negative points, scattered points, and stacked points. The proposed algorithm includes three steps, data pre-cleaning, normal data extraction, and data marking. The negative abnormal points, whose wind speed is greater than cut-in speed and power is below zero, are first filtered in the data pre-cleaning step. The scatter figure of the rest wind power data forms the WPC image and corresponding binary image. In the normal data extraction step, the principle part of the WPC binary image, representing the normal data, is extracted by the mathematical morphology operation (MMO). The optimal parameter setting of MMO is determined by minimizing the dissimilarity between the extracted principle part and the reference WPC image based on Hu moments. In the data mark step, the pixel points of scattered and stacked abnormal data are successively identified. The mapping relationship between the wind power points and image pixel points is built to mark the wind turbine normal and abnormal data. The proposed image-based algorithm is compared with k -means, local outlier factor, combined algorithm based on change point grouping algorithm and quartile algorithm (CA). Numerous experiments based on 33 wind turbines from two wind farms are conducted to validate the effectiveness, efficiency, and universality of the proposed method.

Dental hard tissue morphological segmentation with sparse representation-based classifier.

In the field of dental image processing and analysis, automatic segmentation results of dental hard tissue can provide a useful reference for the clinical diagnosis and treatment process. However, the segmentation accuracy is greatly affected due to the limitation of imaging conditions in the oral environment, as well as the complexity of dental hard tissue topology. To further improve the precision of dental hard tissue segmentation, a novel algorithm was presented by using the sparse representation-based classifier and mathematical morphology operations. First, the captured dental image was preprocessed to eliminate the impact of imbalance local illumination. Then, the preliminary dental hard tissue areas were calculated as the initial marker regions based on color characteristics analysis, and the sparse representation-based classifier was applied sequentially to optimize the initial marker regions combined with certain morphological operations. Finally, a modified marker-controlled watershed transform was employed to segment dental hard tissue regions on the basis of the optimized marker regions, and the final results were obtained after homogeneous region merging. The experimental results show that our method has better adaptability and robustness than existing state-of-the-art methods. Graphical abstract.

Diagnosis of diabetic retinopathy using multi level set segmentation algorithm with feature extraction using SVM with selective features

Diabetic retinopathy is a major cause of blindness in diabetic patients. It is an eye disease caused by diabetes mellitus which affects the retina. Recognition of the severity of this disease at early stage is a challenging factor for the ophthalmologists. In this article, a novel diagnosis system for identifying the severity of diabetic retinopathy is proposed using a multi level set segmentation algorithm and support vector machine with selective features along with genetic algorithm. The proposed system uses some mathematical morphological operations for clustering. After that the clusters are passed to the multi level set segmentation algorithm and some features are extracted using Local Binary Patterns as a texture descriptor for retinal images, color moments and statistical features such as mean, median etc. to detect the major regions of retina. Then the extracted features are given to the support vector machine classifier to classify the disease severity. This system was evaluated and compared using measures of sensitivity and specificity. We obtain sensitivity of 97.14%, specificity of 100% and accuracy of 99.3% on an average. From the seen results, it is observed that our proposed system is suited for the diagnosis of diabetic retinopathy at the early stage.

Realization of remote sensing image segmentation based on K-means clustering

Segmentation of remote sensing image is the key technology of positioning system. Firstly, we transform the remote sensing image from RGB pace to Lab space. Then, three centres are iterated by using K- means algorithm. Finally, in order to eliminate the influence, the closed operation of mathematical morphology is used to correct the segmented image. The results show that it can segment the road from the remote sensing image in lab mode, by using K-means clustering algorithm. Moreover, Lab mode is more suitable for k-mean than other modes.

심층학습 전처리를 통한 도면 문자 인식 성능 개선

An OCR(Optical Character Recognition) based on deep learning for recognition of characters and numbers on drawings for machine maintenance, is proposed. The proposed pre-processing uses mathematical morphology operation to reduce false detection and recognition due to the leading lines and small parts of the drawings. The LSTM model is used to train and infer drawings of parts book. The results of experiments show that the proposed OCR improves the performance by 6.95 % compared to the classical LSTM In addition, the processing speed is improved compared to the conventional approaches.

A novel clustering algorithm based on mathematical morphology for wind power generation prediction

Wind power has the characteristic of daily similarity. Furthermore, days with wind power variation trends reflect similar meteorological phenomena. Therefore, wind power prediction accuracy can be improved and computational complexity during model simulation reduced by choosing the historical days whose numerical weather prediction information is similar to that of the predicted day as training samples. This paper proposes a new prediction model based on a novel dilation and erosion (DE) clustering algorithm for wind power generation. In the proposed model, the days with similar numerical weather prediction (NWP) information to the predicted day are selected via the proposed DE clustering algorithm, which is based on the basic operations in mathematical morphology. And the proposed DE clustering algorithm can cluster automatically without supervision. Case study conducted using data from Yilan wind farm in northeast China indicate that the performance of the new generalized regression neural network (GRNN) prediction model based on the proposed DE clustering algorithm (DE clustering-GRNN) is better than that of the DPK-medoids clustering-GRNN, the K-means clustering-GRNN, and the AM-GRNN in terms of day-ahead wind power prediction. Further, the proposed DE clustering-GRNN model is adaptive.