Top-hat Transform Research Articles

New Martian valley network volume estimate consistent with ancient ocean and warm and wet climate

The volume of Martian valley network (VN) cavity and the amount of water needed to create the cavity by erosion are of significant importance for understanding the early Martian climate, the style and rate of hydrologic cycling, and the possibility of an ancient ocean. However, previous attempts at estimating these two quantities were based on selected valleys or at local sites using crude estimates of VN length, width and depth. Here we employed an innovative progressive black top hat transformation method to estimate them on a global scale based on the depth of each valley pixel. The conservative estimate of the minimum global VN volume is 1.74 × 1014 m3 and minimum cumulative volume of water required is 6.86 × 1017 m3 (or ∼5 km of global equivalent layer, GEL). Both are much larger than previous estimates and are consistent with an early warm and wet climate with active hydrologic cycling involving an ocean.

Plate Recognition Using Backpropagation Neural Network and Genetic Algorithm

Plate recognizer system is an important system. It can be used for automatic parking gate or automatic ticketing system. The purpose of this study is to determine the effectiveness of Genetic Algorithms (GA) in optimizing the number of hidden neurons, learning rate and momentum rate on Backpropagation Neural Network (BPNN) that is applied to the Automatic Plate Number Recognizer (APNR). Research done by building a GA optimized BPNN (GABPNN) and APNR system using image processing methods, including grayscale conversion, top-hat transformation, binary morphological, Otsu threshold and binary image projection. The tests conducted with backpropagation training and recognition test. The result shows that GA optimized backpropagation neural network requires 2230 epochs in the training process to be convergent, which is 36.83% faster than non-optimal backpropagation neural network, while the accuracy is 1,35% better than non-optimized backpropagation neural network.

An infrared small target detection based on multi-feature difference decision coupling Top-Hat transform

In order to improve the detection accuracy of dim target in infrared image with complex background, an infrared small target detection algorithm based on multi-feature tightness difference decision coupling improved Top-Hat transform. Firstly, multi-scale structure elements was obtained by segmenting the single structural element, and the gray change map was defined according to the gray difference between the small target and its surrounding background, and the target decision factor was constructed by calculating the mean and variance of this map, so the new Top-Hat transformation was formed by embedding it and multi-scale structure elements into classical Top-Hat transform. Then the multi-feature tightness difference model was established to extract candidate regions that contain the real weak and small targets were extracted by combining the gray intensity, contrast and structure information. Finally, the pipeline filtering pattern was introduced to eliminate the suspicious objects in the candidate region and keep the real dim target. The experimental data show that this algorithm had higher detection precision to completely check out the dim target with better ROC curve under the complex background.

An Automated Approach for Localizing Retinal Blood Vessels inConfocal Scanning Laser Ophthalmoscopy Fundus Images.

In this work, we present a rules-based method for localizing retinal blood vessels in confocal scanning laser ophthalmoscopy (cSLO) images and evaluate its feasibility. A total of 31 healthy participants (17 female; mean age: 64.0 ± 8.2 years) were studied using manual and automatic segmentation. High-resolution peripapillary scan acquisition cSLO images were acquired. The automated segmentation method consisted of image pre-processing for gray-level homogenization and blood vessel enhancement (morphological opening operation, Gaussian filter, morphological Top-Hat transformation), binary thresholding (entropy-based thresholding operation), and removal of falsely detected isolated vessel pixels. The proposed algorithm was first tested on the publically available dataset DRIVE, which contains color fundus photographs, and compared to performance results from the literature. Good results were obtained. Monochromatic cSLO images segmented using the proposed method were compared to those manually segmented by two independent observers. For the algorithm, a sensitivity of 0.7542, specificity of 0.8607, and accuracy of 0.8508 were obtained. For the two independent observers, a sensitivity of 0.6579, specificity of 0.9699, and accuracy of 0.9401 were obtained. The results demonstrate that it is possible to localize vessels in monochromatic cSLO images of the retina using a rules-based approach. The performance results are inferior to those obtained using fundus photography, which could be due to the nature of the technology.

Classification of Broken Rice Kernels using 12D Features

Integrating the technological aspect for assessment of rice quality is very much needed for the Asian markets where rice is one of the major exports. Methods based on image analysis has been proposed for automated quality assessment by taking into account some of the textural features. These features are good at classifying when rice grains are scanned in controlled environment but it is not suitable for practical implementation. Rice grains are placed randomly on the scanner which neither maintains the uniformity in intensity regions nor the placement strategy is kept ideal thus resulting in false classification of grains. The aim of this research is to propose a method for extracting set of features which can overcome the said issues. This paper uses morphological features along-with gray level and Hough transform based features to overcome the false classification in the existing methods. RBF (Radial Basis function) is used as a classification mechanism to classify between complete grains and broken grains. Furthermore the broken grains are classified into two classes’ i.e. acceptable grains and non-acceptable grains. This research also uses image enhancement technique prior to the feature extraction and classification process based on top-hat transformation. The proposed method has been simulated in MATLAB to visually analyze and validate the results.

Vehicle License Plate Recognition Based on Extremal Regions and Restricted Boltzmann Machines

This paper presents a vehicle license plate recognition method based on character-specific extremal regions (ERs) and hybrid discriminative restricted Boltzmann machines (HDRBMs). First, coarse license plate detection (LPD) is performed by top-hat transformation, vertical edge detection, morphological operations, and various validations. Then, character-specific ERs are extracted as character regions in license plate candidates. Followed by suitable selection of ERs, the segmentation of characters and coarse-to-fine LPD are achieved simultaneously. Finally, an offline trained pattern classifier of HDRBM is applied to recognize the characters. The proposed method is robust to illumination changes and weather conditions during 24 h or one day. Experimental results on thorough data sets are reported to demonstrate the effectiveness of the proposed approach in complex traffic environments.

SEGMENTATION OF RETINAL BLOOD VESSELS BY TOP-HAT MULTI-SCALE DETECTION FOR OPTIC DISC REMOVAL

Nowadays, an automatic retinal vessels segmentation is important component in computer assisted system to detect numerous eye abnormalities. There are various sizes of the retinal blood vessels captured from fundus image modality, which can be detected by using multi-scale approach. However, the main limitation of the current multi-scale approaches is the inability to remove the optic disc from the detected blood vessels. In this paper, a hybrid of multi-scale detection with pre-processing approach is proposed so that clearer vessel segmentation can be obtained. The proposed method embedded with a pre-processing phase that includes four series of processes that include Top-hat transformation as the main part. This technique will reduce the influence of the structure of optic disc and enhance the contrast of the vessel from the background. Then, the result from the pre-processing phase will be fed to the multi-scale detection to perform the segmentation. The proposed method is evaluated on two publicly available online databases: HRF and DRIVE. On HRF database, the best obtained precision and specificity values are 0.9689 and 0.9989, respectively. Meanwhile, for DRIVE database, the system performs well in all performance measures: precision, specificity, accuracy and error with the best values of 0.7541, 0.9739, 0.9510 and 0.0490, respectively. In conclusion, the proposed method is able to filter the unwanted optical disc from the fundus image effectively. Thus, retinal blood vessel image can be used for further analysis process and beneficial for pre-screening system development.

Recognizing Furniture Supported by Legs Based on Vision Sensor

A method of object recognition for objects with special topological structure is developed based on top-hat transformation for service robot. The objects with four legs are easily detected and recognized by service robot with laser ranger array by this method. First, the top-hat transformation in one dimension is reviewed, then the recognition strategy of self-adapting threshold for objects with special topological structure is proposed, and the general data process for object recognition and position is proposed and analyzed. Experimental results show that the process of object recognition based on top-hat transformation proposed in this article is an effective and accurate application.

Automatic Identification of Human Erythrocytes in Microscopic Fecal Specimens.

Traditional fecal erythrocyte detection is performed via a manual operation that is unsuitable because it depends significantly on the expertise of individual inspectors. To recognize human erythrocytes automatically and precisely, automatic segmentation is very important for extraction of characteristics. In addition, multiple recognition algorithms are also essential. This paper proposes an algorithm based on morphological segmentation and a fuzzy neural network. The morphological segmentation process comprises three operational steps: top-hat transformation, Otsu's method, and image binarization. Following initial screening by area and circularity, fuzzy c-means clustering and the neural network algorithms are used for secondary screening. Subsequently, the erythrocytes are screened by combining the results of five images obtained at different focal lengths. Experimental results show that even when the illumination, noise pollution, and position of the erythrocytes are different, they are all segmented and labeled accurately by the proposed method. Thus, the proposed method is robust even in images with significant amounts of noise.

Multi-sensor image enhanced fusion algorithm based on NSST and top-hat transformation

In this paper, a novel method based on nonsubsampled shearlet transform (NSST) and multi-scale top-hat transform (MTHT) is proposed to fuse the SAR image, infrared image and visible light image, which will produce a fused image with higher contrast and better edges. First, three original images are decomposed into low frequency subband coefficients and the bandpass direction subband coefficients by NSST. Second, the effective bright and dark image features from low-frequency coefficients of three original images are extracted by MTHT, then by using the pixel-wise maximum operation, the salient fused bright and dark features are obtained and joined into the combined low frequency coefficients. To reduce the influence of noises of SAR and infrared image on the fused image, the significant factor of local directional information entropy, which can distinguish meaningful image characteristics and noises, is employed as the measurement to select the bandpass direction subband coefficients. Finally, the fused image is obtained by the inverse NSST on the combined coefficients. The experiment results demonstrate the superiority of our method compared with conventional methods in terms of visual inspection and objective measures.

Intoxicated person discrimination using infrared signature of facial blood vessels

In this work, facial thermal infrared images are employed for intoxicated person discrimination. The activity of the facial blood vessels of sober and intoxicated people comes into sight when nonlinear anisotropic diffusion and top-hat transformation are applied to enhance and isolate the vessels from the rest of the information on the face. For an intoxicated person, vessels around the nose and eyes as well as on the forehead become more active, whereas for a person who is sober the vessels’ activity is smoother all over the face’s thermal image. Accordingly, drunkenness can be ascertained by only using the thermal infrared image of the intoxicated person. The proposed method can be incorporated into a non-invasive inspection system used by the police as a first step for intoxicated person detection. The Student’s t-test is employed to assess the degree of confidence in separating the thermal images corresponding to sober and intoxicated people. The 40 participants in the experiment have contributed to the creation of the unique sober–drunk database available on the web.

Detection of chemical fabric defects on the basis of morphological processing

Complex computation and the poor adaptability of ambient light are common problems of the algorithm for fabric defect detection. In this paper, we propose a new fabric defect detection approach based on a morphological filter. First, the structural nodes are analyzed and structural features are extracted for selection of the structural elements. Second, the grayscale morphological operations and top-hat transformation are carried out to highlight the defective areas. At the same time, the phenomenon of uneven brightness distribution is greatly reduced. Power transformation and thresholding are then combined to segment the defective areas. This scheme can detect the defects precisely while reducing both the complexity of the algorithm and the interference of the ambient light. The results of the experiments show that the proposed approach is effective and feasible.

Enhanced Detection of Bearing Faults Based on Adaptive Multi-scale Self-complementary Top-Hat Transformation

Aiming at the difficulty of extracting impulsive fault feature of rolling element bearings in practical engineering,a novel method named adaptive multi-scale self-complementary Top-Hat(AMSTP) transformation is proposed to enhance detection of bearing faults. It can enhance the impulsiveness of the bearing fault vibration signal and depress strong background noise,and constructing multi-scale is better to depress noise and retain detail of signal. The most optimal structure element(SE) scale is selected by using a novel method of feature amplitude energy radio(FAER),and it is applied in detecting fault feature of impulsive signal successfully. The performance of the proposed method is validated by both simulated signal and vibration signals of defective rolling element bearings with ball and inner faults. In addition the method could achieve better effect on feature extraction and have more operation efficiency than open-closing and close-opening combined morphological method based on signal noise ratio(SNR) criterion.

An Improved Top-Hat Filter with Sloped Brim for Extracting Ground Points from Airborne Lidar Point Clouds

Airborne light detection and ranging (lidar) has become a powerful support for acquiring geospatial data in numerous geospatial applications and analyses. However, the process of extracting ground points accurately and effectively from raw point clouds remains a big challenge. This study presents an improved top-hat filter with a sloped brim to enhance the robustness of ground point extraction for complex objects and terrains. The top-hat transformation is executed and the elevation change intensity of the transitions between the obtained top-hats and outer brims is inspected to suppress the omission error caused by protruding terrain features. Finally, the nonground objects of complex structures, such as multilayer buildings, are identified by the brim filter that is extended outward. The performance of the proposed filter in various environments is evaluated using diverse datasets with difficult cases. The comparison of the proposed filter with the commercial software Terrasolid TerraScan and other popular filtering algorithms demonstrates the applicability and effectiveness of this filter. Experimental results show that the proposed filter has great promise in terms of its application in various types of landscapes. Abrupt terrain features with dramatic elevation changes are well preserved, and diverse objects with complicated shapes are effectively removed. This filter has minimal omission and commission error oscillation for different test areas and thus demonstrates a stable and reliable performance in diverse landscapes. In addition, the proposed algorithm has high computational efficiency because of its simple and efficient data structure and implementation.

River Delineation from Remotely Sensed Imagery Using a Multi-Scale Classification Approach

River delineation is an initial yet critical step in river studies. Although the analysis of satellite images shows great potential in river delineation, only a few approaches have been developed. These methods usually focus on rivers at mono-scale and may ignore the large variations in river size. In particular, they may fail to capture the small rivers in the imagery. This paper presents a novel automated multi-scale procedure for delineating complete river networks. This method classifies the large and small rivers separately and combines the two classified results to generate the final delineated river networks. First, a modified normalized difference water index (MNDWI) is adapted to enhance the spectral contrast between open water and land surfaces. Second, a simple OTSU classification is used to delineate the large rivers. Next, a top-hat transformation and multi-scale matched filters are used to enhance the small linear rivers. Then, the OTSU classification is conducted again to delineate the small linear rivers, in concert with a multi-points fast marching method to rejoin the resulting river segments. Finally, the complete river networks are delineated by combining the small and large rivers. A comparison of this procedure with manual digitization when applied to two Landsat-5 TM images demonstrates the former procedure’s value in delineating rivers. It achieves accurate results and outperforms the other three alternative approaches (large river classification, maximum likelihood classifier, and support vector machine classifier) in accuracy, true positive rate, and Kappa coefficient, while also yielding a comparable false positive rate.

Sonar Image Segmentation Based on an Improved Selection of Initial Contour of Active Contour Model

The correct sonar image segmentation is an important foundation for underwater target recognition. Because the contour convergence of the active contour model depends on the selection of initial position, the active contour model is applied in sonar image segmentation. This paper proposed a selection method based on local standard deviation of image as the outline of initial contour. Due to the disturbance of noise, sonar image is usually affected in resolution and contrast. Firstly, sonar image is enhanced by top-hat and bottom-hat transformation in image morphology. Then after image enhancement, a suitable threshold value is chose for rough binarization and the standard deviation of target areas to calculate the local image. According to the size of standard deviation of different regions to determine the scope of the initial contour, sonar image segmentation is achieved by active contour algorithm.

A progressive black top hat transformation algorithm for estimating valley volumes on Mars

The depth of valley incision and valley volume are important parameters in understanding the geologic history of early Mars, because they are related to the amount sediments eroded and the quantity of water needed to create the valley networks (VNs). With readily available digital elevation model (DEM) data, the Black Top Hat (BTH) transformation, an image processing technique for extracting dark features on a variable background, has been applied to DEM data to extract valley depth and estimate valley volume. Previous studies typically use a single window size for extracting the valley features and a single threshold value for removing noise, resulting in finer features such as tributaries not being extracted and underestimation of valley volume. Inspired by similar algorithms used in LiDAR data analysis to remove above-ground features to obtain bare-earth topography, here we propose a progressive BTH (PBTH) transformation algorithm, where the window size is progressively increased to extract valleys of different orders. In addition, a slope factor is introduced so that the noise threshold can be automatically adjusted for windows with different sizes. Independently derived VN lines were used to select mask polygons that spatially overlap the VN lines. Volume is calculated as the sum of valley depth within the selected mask multiplied by cell area. Application of the PBTH to a simulated landform (for which the amount of erosion is known) achieved an overall relative accuracy of 96%, in comparison with only 78% for BTH. Application of PBTH to Ma’adim Vallies on Mars not only produced total volume estimates consistent with previous studies, but also revealed the detailed spatial distribution of valley depth. The highly automated PBTH algorithm shows great promise for estimating the volume of VN on Mars on global scale, which is important for understanding its early hydrologic cycle.

Computer aided quantification for retinal lesions in patients with moderate and severe non-proliferative diabetic retinopathy: a retrospective cohort study.

BackgroundDetection of retinal lesions like micro-aneurysms and exudates are important for the clinical diagnosis of diabetes retinopathy. The traditional subjective judgments by clinicians are dependent on their experience and can be subject to lack of consistency and therefore a quantification method is worthwhile.MethodsIn this study, 10 moderate non-proliferative diabetes retinopathy (NPDR) patients and 10 severe NPDR ones were retrospectively selected as a cohort. Mathematical morphological methods were used for automatic segmentation of lesions. For exudates detection, images were pre-processed with adaptive histogram equalization to enhance contrast, then binary images for area calculation were obtained by threshold classification. For micro-aneurysms detection, the images were pre-processed by top-hat and bottom-hat transformation, then Otsu method and Hough transform were used to classify micro-aneurysms. Post-processing morphological methods were used to preclude the false positive noise.ResultsAfter segmentation, the area of exuduates divided by optic disk area (exudates/disk ratio) and counts of microaneurysms were quantified and compared between the moderate and severe non-proliferative diabetic retinopathy groups, which had significant difference(P < 0.05).ConclusionsIn conclusion, morphological features of lesion might be an image marker for NPDR grading and computer aided quantification of retinal lesion could be a practical way for clinicians to better investigates diabetic retinopathy.

Infrared Image Preprocessing Based on Top-Hat Transformation and Self-Adaption Threshold Segmentation

Infrared targets will render as high gray-scale points in single infrared image because of small size and long distance between targets and space-based infrared sensor,which is similar to noisy points. In order to distinguish targets and noisy points, this paper presents a method of self-adaption threshold segmentation after Top-hat transformation with double structure elements, which could effectively restrain background and noisy points.

Sonar Image Segmentation Using the Level Set Method without Re-Initialization

To solve the problem that many image segmentation methods cannot be applied to forward looking sonar image accurately, an improved level set segmentation method was proposed in this paper. Firstly, the level set evolution without re-initialization was introduced. Secondly the different characteristics of forward looking sonar image from the optical image were analyzed, and we got the factors affecting segmentation. Then, to overcome these negative effects, this paper did preprocessing by morphological top-hat and bottom-hat transformation, and carried on level set method without re-initialization to construct an improved level set sonar image segmentation system. Finally, our method was compared with the traditional level set method in computer experiments. Simulation results show that it is more adapted to forward looking sonar image segmentation.