Mathematical Morphology Processing Research Articles

Automatic Extraction Method of Landslide Based on Digital Elevation Model and Digital Orthophoto Map Data Combined in Complex Terrain

This study aims to accurately determine the distribution of landslides in the complex terrain of Jiangdingya, Nanyu Township, Zhouqu County, Gansu Province. The digital orthophoto map (DOM) and digital elevation model (DEM) are used to accurately identify landslide areas and analyze associated data. Based on image-based supervised classification, the influence factor constraint analysis is used to further identify and delineate the landslide area. Three mathematical morphology operations—erosion, dilation, and opening—are then applied to automatically identify and extract landslides. Experimental results demonstrate that achieving an accuracy, precision, and recall of 98.02%, 85.24%, and 84.78% shows that it is possible to better avoid interference caused by complex terrain with rich features. High-resolution DEM and DOM data contain rich spectral and texture information. These data can accurately depict geomorphic features of complex terrain and aid in identifying landslide-prone areas when combined with mathematical morphology processing. This contribution is important for identifying landslides in complex terrain and emergency disaster management.

A Noncontact Fall Detection Method for Bedside Application With a MEMS Infrared Sensor and a Radar Sensor

With the rapid development of economy, science and technology, the aging issues become more and more serious. People aged above 65 have a risk of 28%-35% to fall. Among them, bedside falls happen most frequently. Therefore, the capability to detect fall events of the elderly is very important. In this paper, a novel non-contact fall detector based on a MEMS low-resolution infrared sensor and a low-cost radar sensor is developed to detect bedside fall. Besides, IR image processing algorithms based on adaptive filter, successive approximation, double boundary scans and mathematical morphology processing are proposed detailedly. Partition processing algorithm is used to suppress the influence of residual or existed heat sources on the bed or ground. Then, the statistical features of the center, area, temperature and duration, as well as stable flag and fall action flag, are extracted for fall recognition. Finally, a three-layer RBF neural network is applied to distinguish the fall events from the non-fall events. Considering the influence factors of ambient temperature, brightness, gender, dressing, fall posture, fall location and scenario, a total of 640 tests are conducted and 5-fold cross validation is used to evaluate the classification performance. Experimental results indicate that the averages of the recall, precision, F1-Score and detection accuracy are measured to be 91.25%, 94.76%, 92.97% and 93.13%, respectively, which demonstrates that the proposed fall detection method is effective. Besides, the detection accuracy decreases from 96.88% to 85.94% as the ambient temperature rises. Hence, this non-contact fall detector can be widely applied for bedside fall detection at home, which is low-cost, non-wearable, unobtrusive, non-invasive and privacy-preserved.

Paper Defect Detection Algorithm Based on the Mathematical Morphology and Computer Vision

A method for detecting paper’s surface defects is proposed, which is based on mathematical morphology algorithm and computer vision technique. Firstly, the mean filtering and wavelet denoising in image pre-processing are used to eliminate the image noise, and wavelet analysis can achieve image enhancement and increase the contrast between the target area and the background. And then, an improved threshold Canny operator is used to segment the image to extract the complete scratch line and other defect edges. Finally, the target region features are coloured by mathematical morphology processing. The method’s stimulating result is verified in the Matlab software, from which we can conclude that the method’s detecting effect is very effective.

Pseudo Overlap Functions, Fuzzy Implications and Pseudo Grouping Functions with Applications

Overlap and grouping functions are important aggregation operators, especially in information fusion, classification and decision-making problems. However, when we do more in-depth application research (for example, non-commutative fuzzy reasoning, complex multi-attribute decision making and image processing), we find overlap functions as well as grouping functions are required to be commutative (or symmetric), which limit their wide applications. For the above reasons, this paper expands the original notions of overlap functions and grouping functions, and the new concepts of pseudo overlap functions and pseudo grouping functions are proposed on the basis of removing the commutativity of the original functions. Some examples and construction methods of pseudo overlap functions and pseudo grouping functions are presented, and the residuated implication (co-implication) operators derived from them are investigated. Not only that, some applications of pseudo overlap (grouping) functions in multi-attribute (group) decision-making, fuzzy mathematical morphology and image processing are discussed. Experimental results show that, in many application fields, pseudo overlap functions and pseudo grouping functions have greater flexibility and practicability.

MEA-Net: A Lightweight SAR Ship Detection Model for Imbalanced Datasets

The existing synthetic aperture radar (SAR) ship datasets have an imbalanced number of inshore and offshore ship targets, and the number of small, medium and large ship targets differs greatly. At the same time, the existing SAR ship detection models in the application have a huge structure and require high computing resources. To solve these problems, we propose a SAR ship detection model named mask efficient adaptive network (MEA-Net), which is lightweight and high-accuracy for imbalanced datasets. Specifically, we propose the following three innovative modules. Firstly, we propose a mask data balance augmentation (MDBA) method, which solves the imbalance of sample data between inshore and offshore ship targets by combining mathematical morphological processing and ship label data to greatly improve the ability of the model to detect inshore ship targets. Secondly, we propose an efficient attention mechanism (EAM), which effectively integrates channel features and spatial features through one-dimensional convolution and two-dimensional convolution, to improve the feature extraction ability of the model for SAR ship targets. Thirdly, we propose an adaptive receptive field block (ARFB), which can achieve more effective multi-scale detection by establishing the mapping relationship between the size of the convolution kernel and the channel of feature map, to improve the detection ability of the model for ship targets of different sizes. Finally, MEA-Net is deployed on the Jeston Nano edge computing device of the 2 GB version. We conducted experimental validation on the SSDD and HRSID datasets. Compared with the baseline, the AP of MEA-Net increased by 2.18% on the SSDD dataset and 3.64% on the HRSID dataset. The FLOPs and model parameters of MEA-Net were only 2.80 G and 0.96 M, respectively. In addition, the FPS reached 6.31 on the Jeston Nano, which has broad application prospects.

Comparative analysis of deep learning classifiers for diabetic retinopathy identification and detection

ABSTRACT Diabetic retinopathy (DR) is a micro vascular problem caused by diabetes that can lead to loss of sight. The early detection of diabetic retinopathy is important to avoid the severity of sightlessness. In this manuscript, a comparative analysis of several deep learning methods for DR identification is proposed. The input fundus images are taken from a standard dataset pre-processed by the Mathematical Morphology process. Moreover, the images are segregated using a Multilevel segmentation of the Region of interest (ROI) based on the split and merge algorithm. After that, an original deep learning architecture is utilized to categorize the segregated fundus images. Deep learning methods, such as Convolution neural network (CNN), Recurrent Neural Network (RNN), Support Vector Machine (SVM), Fuzzy K-means cluster (FKM) and Discriminant Analysis (DA) are proposed to classify the DR. The proposed DR identification and detection with CNN provides 65.54% SP, 100% SE, 78.54% SV and 96.95% ACC. Finally, CNN shows better performance than other classifiers.

Grayscale mathematical morphological processing for clutter mitigation in airborne radars

Grayscale mathematical morphological processing for clutter mitigation in airborne radars

Study on Detection Image Processing Method of Offshore Cage

The paper aims to test the integrity of offshore cages. In this paper, offshore robots are used to collect video images. Through preprocessing methods such as color space conversion and Retinex algorithm of fusion-oriented filtering, the color distortion of the image is improved and the edge information of the image is maintained, so that the collected cage image is better recognized and calculated by the computer. And the paper uses the Canny algorithm to accurately locate the edge of the image, so that the boundary information can be reflected in more detail. In this paper, the adaptive threshold segmentation algorithm combined with mathematical morphology processing method is used to segment the offshore cage image, and extract the skeleton of the decomposed image to realize the detection of the integrity of the offshore net. Finally, a ROV equipped with a camera was used to verify the proposed method in a wave-flow water tank, which confirmed the effectiveness of the method and optimized the integrity recognition and image processing of the cage.

Digital Topological Properties of an Alignment of Fixed Point Sets

The present paper investigates digital topological properties of an alignment of fixed point sets which can play an important role in fixed point theory from the viewpoints of computational or digital topology. In digital topology-based fixed point theory, for a digital image ( X , k ) , let F ( X ) be the set of cardinalities of the fixed point sets of all k-continuous self-maps of ( X , k ) (see Definition 4). In this paper we call it an alignment of fixed point sets of ( X , k ) . Then we have the following unsolved problem. How many components are there in F ( X ) up to 2-connectedness? In particular, let C k n , l be a simple closed k-curve with l elements in Z n and X : = C k n , l 1 ∨ C k n , l 2 be a digital wedge of C k n , l 1 and C k n , l 2 in Z n . Then we need to explore both the number of components of F ( X ) up to digital 2-connectivity (see Definition 4) and perfectness of F ( X ) (see Definition 5). The present paper addresses these issues and, furthermore, solves several problems related to the main issues. Indeed, it turns out that the three models C 2 n n , 4 , C 3 n − 1 n , 4 , and C k n , 6 play important roles in studying these topics because the digital fundamental groups of them have strong relationships with alignments of fixed point sets of them. Moreover, we correct some errors stated by Boxer et al. in their recent work and improve them (see Remark 3). This approach can facilitate the studies of pure and applied topologies, digital geometry, mathematical morphology, and image processing and image classification in computer science. The present paper only deals with k-connected spaces in DTC. Moreover, we will mainly deal with a set X such that X ♯ ≥ 2 .

MR Image Analysis of Neonatal Brain using Adaptive Mathematical Morphological Processing and Region based Segmentation

In the study of evaluation of infant brain development, the segmentation of obtained MR images is an important step. When compared with the MR images of adult brain it is difficult to identify the different regions in infant brains with that of the methods used for the analysis of adult brains. This is due to the size difference of the brain and the differences in the properties of the brain tissues. So, for analyzing these MR images it requires manual interaction with the images resulting in the bias of the results. For this problem we propose another approach for the segmentation of neonatal brain MR images. This method doesn’t require any manual interaction and produces unbiased results. Our algorithm segments the different layers (right hemisphere, left hemisphere, cerebellum, brain stem) and the different tissues like sub cortical gray matter, Militated & un mylinated gray matter and cerebrospinal fluid, resulting in the better understanding of the development of different parts of the brain. Our algorithm can be used for the analysis of MR images of infant brains of age as less as 3to6 months.

Identification of ghost artifact using texture analysis in pediatric spinal cord diffusion tensor images

PurposeGhost artifacts are a major contributor to degradation of spinal cord diffusion tensor images. A multi-stage post-processing pipeline was designed, implemented and validated to automatically remove ghost artifacts arising from reduced field of view diffusion tensor imaging (DTI) of the pediatric spinal cord. MethodA total of 12 pediatric subjects including 7 healthy subjects (mean age=11.34years) with no evidence of spinal cord injury or pathology and 5 patients (mean age=10.96years) with cervical spinal cord injury were studied. Ghost/true cords, labeled as region of interests (ROIs), in non-diffusion weighted b0 images were segmented automatically using mathematical morphological processing. Initially, 21 texture features were extracted from each segmented ROI including 5 first-order features based on the histogram of the image (mean, variance, skewness, kurtosis and entropy) and 16s-order feature vector elements, incorporating four statistical measures (contrast, correlation, homogeneity and energy) calculated from co-occurrence matrices in directions of 0°, 45°, 90° and 135°. Next, ten features with a high value of mutual information (MI) relative to the pre-defined target class and within the features were selected as final features which were input to a trained classifier (adaptive neuro-fuzzy interface system) to separate the true cord from the ghost cord. ResultsThe implemented pipeline was successfully able to separate the ghost artifacts from true cord structures. The results obtained from the classifier showed a sensitivity of 91%, specificity of 79%, and accuracy of 84% in separating the true cord from ghost artifacts. ConclusionThe results show that the proposed method is promising for the automatic detection of ghost cords present in DTI images of the spinal cord. This step is crucial towards development of accurate, automatic DTI spinal cord post processing pipelines.

MODEL-BASED BUILDING DETECTION FROM LOW-COST OPTICAL SENSORS ONBOARD UNMANNED AERIAL VEHICLES

Abstract. The automated and cost-effective building detection in ultra high spatial resolution is of major importance for various engineering and smart city applications. To this end, in this paper, a model-based building detection technique has been developed able to extract and reconstruct buildings from UAV aerial imagery and low-cost imaging sensors. In particular, the developed approach through advanced structure from motion, bundle adjustment and dense image matching computes a DSM and a true orthomosaic from the numerous GoPro images which are characterised by important geometric distortions and fish-eye effect. An unsupervised multi-region, graphcut segmentation and a rule-based classification is responsible for delivering the initial multi-class classification map. The DTM is then calculated based on inpaininting and mathematical morphology process. A data fusion process between the detected building from the DSM/DTM and the classification map feeds a grammar-based building reconstruction and scene building are extracted and reconstructed. Preliminary experimental results appear quite promising with the quantitative evaluation indicating detection rates at object level of 88% regarding the correctness and above 75% regarding the detection completeness.

Extraction of Ocean Fronts Based on Empirical Mode Decomposition

Ocean front is a narrow transitional zone that the penetration of sea is obviously different between two or more waters there. It is an important feature of geophysical turbulence which plays an important role in ocean dynamics. Ocean fronts become visible on radar images because they are associated with a variable surface current which modulates the sea surface roughness and thus the backscattered radar power. This paper propose a new integrated method to extract ocean fronts based on two-dimensional Empirical Mode Decomposition (EMD), image edge detection and mathematical morphology processing. Experimental results show that this integrated method can be effective in ocean front feature extraction.

Automated Extraction of the Archaeological Tops of Qanat Shafts from VHR Imagery in Google Earth

Qanats in northern Xinjiang of China provide valuable information for agriculturists and anthropologists who seek fundamental understanding of the distribution of qanat water supply systems with regard to water resource utilization, the development of oasis agriculture, and eventually climate change. Only the tops of qanat shafts (TQSs), indicating the course of the qanats, can be observed from space, and their circular archaeological traces can also be seen in very high resolution imagery in Google Earth. The small size of the TQSs, vast search regions, and degraded features make manually extracting them from remote sensing images difficult and costly. This paper proposes an automated TQS extraction method that adopts mathematical morphological processing methods before an edge detecting module is used in the circular Hough transform approach. The accuracy assessment criteria for the proposed method include: (i) extraction percentage (E) = 95.9%, branch factor (B) = 0 and quality percentage (Q) = 95.9% in Site 1; and (ii) extraction percentage (E) = 83.4%, branch factor (B) = 0.058 and quality percentage (Q) = 79.5% in Site 2. Compared with the standard circular Hough transform, the quality percentages (Q) of our proposed method were improved to 95.9% and 79.5% from 86.3% and 65.8% in test sites 1 and 2, respectively. The results demonstrate that wide-area discovery and mapping can be performed much more effectively based on our proposed method.

A Mathematical Morphological Processing of Spectrograms for the Tone of Chinese Vowels Recognition

As One of Features from other Languages, the Chinese Tone Changes of Chinese are Mainly Decided by its Vowels, so the Vowel Variation of Chinese Tone Becomes Important in Speech Recognition Research. the Normal Tone Recognition Ways are Always Based on Fundamental Frequency of Signal, which can Not Keep Integrity of Tone Signal. we Bring Forward to a Mathematical Morphological Processing of Spectrograms for the Tone of Chinese Vowels. Firstly, we will have Pretreatment to Recording Good Tone Signal by Using Cooledit Pro Software, and Converted into Spectrograms; Secondly, we will do Smooth and the Normalized Pretreatment to Spectrograms by Mathematical Morphological Processing; Finally, we get Whole Direction Angle Statistics of Tone Signal by Skeletonization way. the Neural Networks Stimulation Shows that the Speech Emotion Recognition Rate can Reach 92.50%.

PCB Defect Detection, Classification and Localization using Mathematical Morphology and Image Processing Tools

An automated visual printed circuit board (PCB) inspection is an approach used to counter difficulties occurred in human’s manual inspection that can eliminates subjective aspects and then provides fast, quantitative, and dimensional assessments. Various concentrated work on detection of defects on printed circuit boards (PCBs) have been done, but it is also crucial to classify these defects in order to analyze and identify the root causes of the defects. This project proposes a PCB defect detection and classification system using a morphological image segmentation algorithm and simple the image processing theories. However, besides the need to detect the defects, it is also essential to classify and locate these defects so that the source and location of these defects can be identified. Based on initial studies, some PCB defects can only exist in certain groups. Thus, it is obvious that the image processing algorithm could be improved by applying a segmentation exercise. This project uses template and test images of single layer, bare, grayscale computer generated PCBs.

COLOR SEGMENTATION OF MGG COLORED CYTOLOGICAL IMAGES USING NON LINEAR OPPONENT COLOR SPACES

This paper presents a color image segmentation method for the quantification of viable cells from samples obtained after cytocentrifugation process and May Grunwald Giemsa (MGG) coloration and then observed by optical microscopy. The method is based on color multi-thresholding and mathematical morphology processing using color information on human visual system based models such as CIELAB model, LUX (Logarithmic hUe eXtension) model and CoLIP (Color Logarithmic Image Processing) model, a new human color vision based model also presented in this article. The results show that the CoLIP model, developed following each step of the human visual color perception, is particularly well adapted for this type of images.

Perceptual color hit-or-miss transform: application to dermatological image processing

The hit-or-miss transform is a mathematical morphological processing designed to find objects in image. Its extension to grayscale domain is not unique, but Barat’s method is the most appropriate to find specific objects with bandwidth in space and color evolutions. This tolerance is possible with the use of non-flat structuring elements. In this paper, a color extension of the Barat’s method is presented. For this purpose, a new mathematical morphology method, based on the concept of convergence in the CIELAB space and where the definition of non-flat structuring elements is possible, is used. A comparison with the existing approaches in the literature is done. Results are given and commented on synthetic and real images.

Particle shape characterisation and its application to discrete element modelling

Increasing importance has been placed on particle shape implementation within discrete element modelling (DEM) in order to more accurately reflect the non-spherical behaviour of the bulk material being handled. As computational resources grow, complex particle shapes are increasingly being modelled as the associated simulation times become more realistic to provide timely solutions. The objective of this research is to assess particle shape descriptors through a digital image segmentation technique, and to further implement particle shape parameters into generation of corresponding irregular shaped DEM particles. Separated and lumped particle images were analysed and reconstructed through the development of two distinct methodologies. Subsequently, various particle shape descriptors were obtained using combinations of image segmentation algorithms, including mathematical morphology processing, thresholding, edge detection, region growing, region splitting and region merging. DEM particles were subsequently created using particle shape results obtained above. Shape parameters of DEM particles were then examined and validated against the real particle shape parameters.

Size and Defect Detection of Hami Big Jujubes Based on Computer Vision

In order to realize the rapid nondestructive testing for Hami Big Jujubes’ quality detection, a detecting system based on computer vision was established to detect Hami Big Jujubes’ size and defect. The image grabbing card and CCD camera were consisted of the hardware system, which was used to collect image data. Visual Basic6.0 and image processing toolbox of Mil9.0 constituted the software system. The function of MIL9.0 was called in the Visual Basic6.0 to realize the detection. During image processing, the threshold was all chose (0.1,0.7).Many methods were used to identify the features rapidly and get the H value’s mean and variance, such as colour space transformation, mathematical morphology processing and mask etc. Experimental results showed that the correlation coefficient between the projective areas and weights was 0.945.The correlation between projective areas, transverse diameter and vertical diameter was 0.951.The defects grading models were built by BP neural network .The discriminating rate was as high as 99.16％ in training set，and 91.43％ in prediction set. The average testing time was 80 milliseconds, which can satisfy the detection system’s requirements of time．