QuickBird Satellite Images Research Articles

Large paleoearthquakes and Holocene faulting in the Southeastern Gorny Altai: implications for ongoing crustal shortening in Central Asia

ABSTRACT Shrinking of intermontane basins and expansion of their flanking ranges by reverse faulting and backthrusting in two counter-dipping systems is a typical mechanism of crustal shortening and mountain building in Central Asia. This mechanism is realized along the Kurai Fault Zone (southeastern Gorny Altai). Motions on two reverse fault systems maintained thrusting of the Kurai Range and the Kubadru Uplift on the Kurai Basin sediments and caused the growth of a foreberg before the mountain front. Forberg separates narrow Aktash Basin from the Kurai Basin. The paleoearthquakes were generated by reverse faults that delineate the foreberg. Analysis of the QuickBird satellite images, drone imagery, trenching, archaeoseismological research, radiocarbon dating, dendrochronology, and previous results show that eleven large (М W = 6.5–7.6) paleoearthquakes left traces as surface ruptures along the Kurai Fault Zone: twice before 7.5 ka BP, three events between 7.5 and 5.9 ka BP (7.0, 6.3, and 5.9–5.8 ka BP), one from 5.8 to 4.6 ka BP, four more at 4.6, 3.2, 1.5, and 1.3–1.2 ka BP, and the ultimate earthquake no older than 1450–1650 AD. The time difference between large earthquakes was from 200 to 1700 years. Surface faulting occurred mainly along the northern border of the foreberg where fault scarps are progressively younger northward and the Cenozoic sediments of the Aktash Basin thus become involved into uplift. GPR data to a depth of 12 m confirm the complex structure and slip geometry of the observed surface ruptures. The fault scarps are located < 1 km from the planned route of the gas pipeline from Russia to China, and the potential seismic hazard has to be taken into account in its design and construction.

Classification of a quickbird satellite image by Machine learning techniques: Mapping an urban Environement by decision tree method

Classification is a crucial stage in the processing of satellite images that influence considerably the quality of the result. A variety of methods is proposed in the literature for the purposes of image classification. They present many differences in their basic principles, thus in the quality of the results obtained. Therefore, a study of different classification methods seems to be essential. The classification of satellite images with conventional methods can be done in several ways using different algorithms. These algorithms can be divided into two main categories: supervised and non-supervised. Decision tree on the contrary is a machine learning tool. It is a plain model characterized by the simplicity of understanding and interpretation. This work aims firstly, to classify a high resolution Quickbird satellite image of an urban area by the decision tree method and compare it with the conventional classification algorithms in order to evaluate its efficiency. The methodology consists of two main stages: classification and evaluation of results. The second is based on the calculation of a number of statistical indices derived from the confusion matrix: the statistical parameter “kappa’ and the overall coefficient of precision.

A Framework for Improving Urban Land Cover Using Object and Pixel-Based Techniques via Remotely Sensed Data

Recently, the advancement of remote sensing technology played a key role in urban land/cover mapping, planning, tourism, and environmental management. Images with a high spatial resolution for urban classification are widely used. Despite the high spectral resolution of the image, spectral confusion happens among different land covers. Furthermore, the shadow problem also causes poor results in the classification based on traditional per-pixel spectral approaches. This study looks at ways of improving the classification of urban land cover using QuickBird images. Maximum likelihood (ML) pixel-based supervised as well as Rule-based object-based approaches were examined on high-resolution QuickBird satellite images in Karbala City, Iraq. This study indicates that the use of textural attributes during the rule-based classification procedure can significantly improve land-use classification performance. Furthermore, the results show that rule-based results are highly effective in improving classification accuracy than pixel-based. The results of this study provide further clarity and insight into the implementation of using the object-based approach with various classifiers for the extended study. In addition, the finding demonstrated the integration of high-resolution QuickBird data and a set of attributes derived from the visible bands and geometric rule set resulted in superior class separability, thus higher classification accuracies in mapping complex urban environments.

OBJECT-BASED APPROACHES FOR LAND USE-LAND COVER CLASSIFICATION USING HIGH RESOLUTION QUICK BIRD SATELLITE IMAGERY (A CASE STUDY: KERBELA, IRAQ)

Land Use / Land Cover (LULC) classification is considered one of the basic tasks that decision makers and map makers rely on to evaluate the infrastructure, using different types of satellite data, despite the large spectral difference or overlap in the spectra in the same land cover in addition to the problem of aberration and the degree of inclination of the images that may be negatively affect rating performance. The main objective of this study is to develop a working method for classifying the land cover using high-resolution satellite images using object based method. Maximum likelihood pixel based supervised as well as object approaches were examined on QuickBird satellite image in Karbala, Iraq. This study illustrated that use of textural data during the object image classification approach can considerably enhance land use classification performance. Moreover, the results showed higher overall accuracy (86.02%) in the o object based method than pixel based (79.06%) in urban extractions. The object based performed much more capabilities than pixel based.

Optimum Feature and Classifier Selection for Accurate Urban Land Use/Cover Mapping from Very High Resolution Satellite Imagery

Feature selection to reduce redundancies for efficient classification is necessary but usually time consuming and challenging. This paper proposed a comprehensive analysis for optimum feature selection and the most efficient classifier for accurate urban area mapping. To this end, 136 multiscale textural features alongside a panchromatic band were initially extracted from WorldView-2, GeoEye-3, and QuickBird satellite images. The wrapper-based and filter-based feature selection were implemented to optimally select the best ten percent of the primary features from the initial feature set. Then, machine leaning algorithms such as artificial neural network (ANN), support vector machine (SVM), and random forest (RF) classifiers were utilized to evaluate the efficiency of these selected features and select the most efficient classifier. The achieved optimum feature set was validated using two other images of WorldView-3 and Pleiades. The experiments revealed that RF, particle swarm optimization (PSO), and neighborhood component analysis (NCA) resulted in the most efficient classifier and wrapper-based and filter-based methods, respectively. While ANN and SVM’s process time depended on the number of input features, RF was significantly resistant to the criterion. Dissimilarity, contrast, and correlation features played the greatest contributing role in the classification performance among the textural features used in this study. These trials showed that the feature number could be reduced optimally to 14 from 137; these optimally selected features, alongside the RF classifier, can produce an F1-measure of about 0.90 for different images from five very high resolution satellite sensors for various urban geographical landscapes. These results successfully achieve our goal of assisting users by eliminating the task of optimal feature selection and classifier, thereby increasing the efficiency of urban land use/cover classification from very high resolution images. This optimal feature selection can also significantly reduce the high computational load of the feature-engineering phase in the machine and deep learning approaches.

Картографирование защитной лесистости территорий и оценка ее оптимальности

Protective forest plantations are an integral part of agroforestry landscapes that serve as an ecological frame and contribute to the expanded reproduction of soil fertility. The problem of assessing the actual protective forest cover of territories is relevant. The solution of this issue allows you to identify the need, sequence and volume of forest reclamation work. The problem of mapping and assessing the protective forest cover of territories is closely related to the tasks of identifying the optimal forest cover of territories and substantiating approaches to the allocation of spatial-territorial complexes. The boundaries of the identified contours directly affect the values of specific indicators, which are the basis for the gradation of the need for forest reclamation work. The conducted studies demonstrate the possibilities of using the QGIS software package for the purpose of assessing the protective forest cover of the territory. The interpretation and creation of a cartographic model of protective forest plantations of the study area was carried out using QuickBird satellite images. The assessment was made on the basis of the calculation of specific indicators within the boundaries of landscapes for the territory of the Ilovlinsky District of the Volgograd region. As a standard for comparing the studied indicators, indicators of the protective forest cover of the “Kachalino” pilot farm were used. The application of the landscape-typological approach to the identification of territorial complexes made it possible to approach the issue of assessing the protective forest cover of territories in a more differentiated way than when using administrative boundaries. As a result of the study, landscape areas were identified, the protective forest cover of which is close to the normative, as well as those areas that need agroforestry measures. Cartographic work carried out in the QGIS software package confirms its effectiveness and allows us to recommend its use in agroforestry research related to the parallel application of various approaches to the allocation of territorial complexes.

Mineral Exploration in Mawat Region, Kurdistan-Iraq, Based on Satellite Data and Terrain Prospection

The main goal of the presented exploration was to estimate potential for mineralization in the Mawat ophiolitic massif in Kurdistan, Iraq. The aim of the study was to explore existing copper mineralization and assessor elements gold, platinoids and chromium. Geological exploration detected two types of Cu occurrence a) secondary Cu carbonates (malachite) and b) Cu sulfides (chalcopyrite-pyrite). The Mawat region is mostly built of ultrabasic and basic rocks: peridotites, gabbros, serpentinites and basalts which are heavily deformed, with faults mostly oriented NNW-SSE, and NE-SW. The first phase of exploration comprised digital processing of ASTER and QuickBird satellite images, with appropriate geometrical and radiometric corrections and transformation into coordinate system. Color composite images were produced in different scales. They served to define lithological composition, tectonic settings, location of the points of interest etc. The field work was designed to check satellite data in situ, with focus on perspective rock formations, which might host copper mineralization, and other elements. The host rocks of the ore occurrences are primarily gabbros and metagabbros intersected by diabase dykes, epidote and quartz veins. Secondary mineralization is the product of surficial weathering and it is represented by malachite and limonite. The geophysical survey was very useful in the detection of area with elevated induced polarization and low resistivity. Three perspective areas have been selected for detailed explorations: Waraz, Mirava-Chenara and Konjirin-Kuradawi. The concentration of copper varies highly in very wide ranges; the maximum measured concentration of Cu was determined in Waraz area 6.7%. Some rock samples also show concentration of gold from 0.36 to 2.59 ppm Au. Keywords: Mawat ophiolitic massif, geologic-geophysical explorations, copper mineralization, Kurdistan-Iraq

Improvement of Region-Merging Image Segmentation Accuracy Using Multiple Merging Criteria

Image segmentation plays a significant role in remote sensing image processing. Among numerous segmentation algorithms, the region-merging segmentation algorithm is widely used due to its well-organized structure and outstanding results. Many merging criteria (MC) were designed to improve the accuracy of region-merging segmentation, but each MC has its own shortcomings, which can cause segmentation errors. Segmentation accuracy can be improved by referring to the segmentation results. To achieve this, an approach for detecting and correcting region-merging image segmentation errors is proposed, and then an iterative optimization model is established. The main contributions of this paper are as follows: (1) The conflict types of matching segment pairs are divided into scale-expression conflict (SEC) and region-ownership conflict (ROC), and ROC is more suitable for optimization. (2) An equal-scale local evaluation method was designed to quantify the optimization potential of ROC. (3) A regional anchoring strategy is proposed to preserve the results of the previous iteration optimization. Three QuickBird satellite images of different land-cover types were used for validating the proposed approach. Both unsupervised and supervised evaluation results prove that the proposed approach can effectively improve segmentation accuracy. All explicit and implicit optimization modes are concluded, which further illustrate the stability of the proposed approach.

Forest Aboveground Biomass Estimation and Mapping through High-Resolution Optical Satellite Imagery—A Literature Review

This paper provides a comprehensive literature review on forest aboveground biomass (AGB) estimation and mapping through high-resolution optical satellite imagery (≤5 m spatial resolution). Based on the literature review, 44 peer-reviewed journal articles were published in 15 years (2004–2019). Twenty-one studies were conducted in Asia, eight in North America and Africa, five in South America, and four in Europe. This review article gives a glance at the published methodologies for AGB prediction modeling and validation. The literature review suggested that, along with the integration of other sensors, QuickBird, WorldView-2, and IKONOS satellite images were most widely used for AGB estimations, with higher estimation accuracies. All studies were grouped into six satellite-derived independent variables, including tree crown, image textures, tree shadow fraction, canopy height, vegetation indices, and multiple variables. Using these satellite-derived independent variables, most of the studies used linear regression (41%), while 30% used linear multiple regression and 18% used non-linear (machine learning) regression, while very few (11%) studies used non-linear (multiple and exponential) regression for estimating AGB. In the context of global forest AGB estimations and monitoring, the advantages, strengths, and limitations were discussed to achieve better accuracy and transparency towards the performance-based payment mechanism of the REDD+ program. Apart from technical limitations, we realized that very few studies talked about real-time monitoring of AGB or quantifying AGB change, a dimension that needs exploration.

Building change detection in very high-resolution remote sensing image based on pseudo-orthorectification

ABSTRACT When using very high-resolution (VHR) remote sensing images acquired at different times to detect building changes, the building positional inconsistencies caused by different satellite imaging angles are an outstanding issue. To tackle this problem, a novel building change detection method based on pseudo-orthorectification (PO) is proposed. First, to determine the building displacement value, a fast line detection method is used to accurately extract the building vertical facade contour lines under the constraint of the Object Space Positioning Consistency (OSPC). Second, the building roof sample selection is automatically conducted under the constraint of building facade contour lines, and the Grab-Cut algorithm is used to extract the roofs combining with corresponding geometric rules. Then, the roof of each building is shifted along the elevation line to its real location. Finally, subtraction is applied to generate the difference image, and reliable change information is obtained by integrating NDVI and shadow information of the building. Three sets of WorldView and QuickBird satellite images are used to compare the proposed method with three state-of-the-art methods. The experimental results show that the average accuracy of the proposed method can reach 92.80%, which is 12.66% higher than that of compared methods.

Object-Based Method for Urban Extraction through Using Quick Bird Satellite Imagery, LiDAR Data and Digital Urban Geomatics Techniques

Urban extraction mapping has become increasingly important in recent years and particularity extraction urban features based on remotely sensed data such as high-resolution imagery and LiDAR data. Though the researchers used the high spatial resolution image to extract urban area but he methods are still complexand still there are challenges associated with combining data that were acquired over differing time periods using inconsistent standards. So, this study will focus on the extraction of urban area based on an object-based classification method with integration of Quickbird satellite image and digital surface elevation (DSM) extracted from LiDAR data for the Rusafa city of Baghdad, Iraq. All the processes were done in eCognition and ArcGIS software for feature extraction and mapping, respectively. The overall methodological steps proposed in this research for the extraction of urban area using object-based method. In addition of that both the image data and LiDAR-derived DSM were integrated based on the eCognition software for extraction urban map of Rusafa city, Baghdad. Finally, the results indicated that the Artificial Neural Networks (ANN) model achieved the highest training and testing accuracies and performed the best compared to RF and Support Vector Machines (SVM) methods. And also, the results showed that the Artificial Neural Networks (ANN) had capability to extract the boundaries of the buildings and other urban features more accurately than the other two methods. This could be interpreted as the Artificial Neural Networks (ANN) model can learn complex features by the optimization process of the model and its multi-level feature extraction property.

PSGAN: A Generative Adversarial Network for Remote Sensing Image Pan-Sharpening

This article addresses the problem of remote sensing image pan-sharpening from the perspective of generative adversarial learning. We propose a novel deep neural network-based method named pansharpening GAN (PSGAN). To the best of our knowledge, this is one of the first attempts at producing high-quality pan-sharpened images with generative adversarial networks (GANs). The PSGAN consists of two components: a generative network (i.e., generator) and a discriminative network (i.e., discriminator). The generator is designed to accept panchromatic (PAN) and multispectral (MS) images as inputs and maps them to the desired high-resolution (HR) MS images, and the discriminator implements the adversarial training strategy for generating higher fidelity pan-sharpened images. In this article, we evaluate several architectures and designs, namely, two-stream input, stacking input, batch normalization layer, and attention mechanism to find the optimal solution for pan-sharpening. Extensive experiments on QuickBird, GaoFen-2, and WorldView-2 satellite images demonstrate that the proposed PSGANs not only are effective in generating high-quality HR MS images and superior to state-of-the-art methods but also generalize well to full-scale images.

Unsupervised Change Detection Using Spectrum-Trend and Shape Similarity Measure

The emergence of very high resolution (VHR) images contributes to big challenges in change detection. It is hard for traditional pixel-level approaches to achieve satisfying performance due to radiometric difference. This work proposes a novel feature descriptor that is based on spectrum-trend and shape context for VHR remote sensing images. The proposed method is mainly composed of two aspects. The spectrum-trend graph is generated first, and then the shape context is applied in order to describe the shape of spectrum-trend. By constructing spectrum-trend graph, spatial and spectral information is integrated effectively. The approach is performed and assessed by QuickBird and SPOT-5 satellite images. The quantitative analysis of comparative experiments proves the effectiveness of the proposed technique in dealing with the radiometric difference and improving the accuracy of change detection. The results indicate that the overall accuracy and robustness are both boosted. Moreover, this work provides a novel viewpoint for discriminating changed and unchanged pixels by comparing the shape similarity of local spectrum-trend.

Evaluasi Pemanfaatan Citra Tegak Satelit Resolusi Tinggi untuk Percepatan Pembuatan Peta Blok Pajak Bumi dan Bangunan (PBB-P2) secara Partisipatif di Desa Triharjo, Kabupaten Bantul, DIY

The recent advances of remote sensing technology development can support government needs including cadastral related purposes. One of the purposes is to supply geospatial data especially land parcel maps for land and tax value inventory. This research tries to evaluate the utilization of Quickbird Satellite image for updating land and building (PBB-P2) block maps in Bantul Regency, Yogyakarta Province of Indonesia. The image was recorded in 2014 with resolution of 0.6 meter and has been ortho-rectified. Evaluation is conducted by analyzing the utilization of the image as the base map for deriving land parcel maps according to the Decree of Tax Director General number: KEP-533/PJ/2000. The process was firstly done by digitizing land parcel boundaries on Quickbird image to produce block maps. The process was done by the community through participatory mapping. The resulting block maps then were compared with land registration map from National Land Agency. The evaluation was focused on the accuracy of land parcel position as well as the resulting land areas. For this purpose the land registration map is regarded as the reference data. By focusing on the 51 sample data in Triharjo Village, statistical test was conducted to evaluate the accuracy of the resulting areas of the digitized land parcel map. The results show that 47 samples comply with a tolerance requirement of 10% according to the Decree of Tax Directorate number KEP-533/PJ/2000’. In addition, the t-test analysis with a significant level of 5% indicates that there is no significant difference of area between the digitized land parcel data with land registration map. Based on the results, it is concluded that ortho-rectified QuickBird image is sufficient for producing of land and building tax (PBB-P2) block maps.

Water Supply System Assessment in Urban Areas Using Geospatial Techniques: A Case Study of Baghdad City

This study shows, how to linked GPS, RS, LiDAR and GIS mapping facilities of urban distribute network, positioning of the water projects and the storage water reservoirs for the water supply systems, with consideration of the earth’s surface properties, land-use change and ownership. Moreover, the use of the Global Positioning System (GPS) with Remote Sensing imagery in surveying, modernizing and developing Baghdad city complex water supply network is discussed. The data of land-use and current water supply system was generated using the Quick Bird Satellite Image (0.6m resolution data, 2015) with a map of the city of Baghdad at scale 1:25000 and field update survey. The necessary water supply system maps and data were collected from the Mayoralty of Baghdad (MoB). The Digital Elevation Model (DEM) and the contour map were generated from the LiDAR point cloud. All these techniques and geospatial data were mapped and analyzed in ArcGIS 10.2.2 software. The outputs of the study were the water distribution network with the position of projects, reservoirs, pipelines, etc. The interpretation of the geospatial database with analytical tools allowed the water distribution network to be planned more economically and effectively.

The Contribution of Multispectral Satellite Image to Shallow Water Bathymetry Mapping on the Coast of Misano Adriatico, Italy

The results of absolute satellite-derived bathymetry (SDB) are presented in the current study. A comparative analysis was conducted on empirical methods in order to explore the potential of SDB in shallow water on the coast of Misano, Italy. Operations were carried out by relying on limited in situ water depth data to extract and calibrate bathymetry from a QuickBird satellite image acquired on a highly dynamic coastal environment. The image was processed using the log-band ratio and optimal band ratio analysis (OBRA) methods. Preprocessing steps included the conversion of the raw satellite image into top of atmosphere reflectance, spatial filtering, land and water classification, the determination of the optimal OBRA spectral band pairs, and the estimation of relative SDB. Furthermore, calibration and vertical referencing were performed via in situ bathymetry acquired in November 2007. The relative bathymetry obtained from different band ratios were vertically referenced to the local datum using in situ water depth in order to obtain absolute SDB. The coefficient of determination (R2) and vertical root mean square error (RMSE) were computed for each method. A strong correlation with in situ field bathymetry was observed for both methods, with R2 = 0.8682 and RMSE = 0.518 m for the log-band ratio method and R2 = 0.8927–0.9108 and RMSE = 0.35 m for the OBRA method. This indicated a high degree of confidence of the SDB results obtained for the study area, with a high performance of the OBRA method for SDB mapping in turbid water.

DV3+HED+: a DCNN-based framework to monitor temporary works and ESAs in railway construction project using VHR satellite images

Current very high-resolution (VHR) satellite images enable the detailed monitoring of the Earth and can capture the ongoing works of railway construction. We presented an integrated framework applied to monitoring railway construction in China using QuickBird, Gaofen-2, and Google Earth VHR satellite images. We also constructed a deep convolutional neural network-based semantic segmentation network to label the temporary works, such as borrow and spoil areas, camp, beam yard, and environmentally sensitive areas such as resident houses throughout the whole railway construction project using VHR satellite images. In addition, we employed the holistically nested edge detection subnetwork to refine the boundary details and a cross entropy loss function to fit the sample class disequilibrium problem. Our semantic segmentation network was trained on 471, validated on 58, and tested on the 58 VHR true color images. The experiment results showed that compared with the existing state-of-the-art DeepLabV3 plus approach, our approach has improvements with an overall accuracy of more than 80%.

Automated Extraction of Slum Built-up Areas from Multispectral Imageries

Slum areas are dense urban areas in which the building size is quite small, and the buildings are interconnected with each other. Also, there is a lot of variation in the texture of slum area buildings, which makes the extraction of individual buildings in slum areas a tough task. In this paper, a methodology has been proposed which aims to extract slum built-up areas using multispectral satellite images using MATLAB software. In the proposed methodology, two building masks have been prepared from the input image by using threshold value and Laws’ texture energy measure. After that, another building mask has been prepared by using these two masks and vegetation, non-building areas and shadow areas have been removed from it, which finally results in the detection of slum built-up areas. The proposed methodology has been applied on three subsets of QuickBird satellite image containing slum built-up areas. For accuracy assessment, the total slum area extracted from the proposed methodology has been compared with the total area obtained by manually digitized buildings. The overall accuracy of slum built-up extraction with respect to area has been found to be more than 83%. Due to resemblance of building and road texture, some over-extraction of road as slum built-up areas has also been observed in only one subset image. No over-detection has been found in other two subset images.

Spatial Analysis of Forest–Tundra Ecotones Reveals the Influence of Topography and Vegetation on Alpine Treeline Patterns in the Subarctic

The response of vegetation to recent climate change is a central theme in contemporary biogeography. A tenet of this research is that forests will advance upslope and northward as climate warms, replacing tundra communities. There has been considerable variation, however, in the pace and extent of recent change, particularly in alpine regions. The objectives of this study were to determine how and why tree spatial patterns vary across different topographic features, and to use spatial patterns to infer the mechanisms governing treeline dynamics in an alpine region of subarctic Canada. We mapped trees across different elevations, slope aspects, and slope angles in two QuickBird satellite images, and in field plots established within the bounds of each image. We then quantified the degree of clustering among trees using Ripley’s K(t) statistic. We also classified each image into vegetation classes and used class-level landscape metrics to quantify the degree of treeline abruptness in different topographic settings. We found that clustering of stems was more common on south than north aspects, likely due to the high occurrence of overwinter damage on the former. Treelines were also more abrupt on south aspects, likely because high tall shrub abundance on these slopes inhibits tree seedling establishment. We conclude that the spatial patterns of subarctic alpine treelines are strongly influenced by both physical and biological factors that vary strongly with slope aspect. The response of treelines to future climate change will likely be highly variable at the landscape scale, despite experiencing similar climatic conditions. Key Words: forest–tundra ecotone, FragStats, landscape metrics, QuickBird, spatial pattern analysis, Yukon.

Detection of ships in inland river using high-resolution optical satellite imagery based on mixture of deformable part models

Ship detection using optical satellite imagery is of great significance in many applications such as traffic surveillance, pollution monitoring, etc. So far, a lot of ship detection methods have been developed for images covering open sea, offshore area and harbors. Compared to the ship detection in sea and offshore area, it is more difficult to detect ships in inland river due to several challenges. First of all, many ships in inland river are clustered together and hard to be separated from each other. Secondly, ships lying alongside the pier are very likely to be recognized as part of the pier. Thirdly, ships in inland river is usually smaller than those in the sea. A hierarchical method is proposed to detect the ships in inland river in this paper. The Regions of Interest (ROIs) are firstly extracted based on water–land segmentation using multi-spectral information. Then two kinds of ship candidates are extracted based on the panchromatic band. The isolated ships are detected by analyzing the shape of connected components and the clustered ships are detected by using mixtures multi-scale Deformable Part Models (DPM) and Histogram of Oriented Gradient (HOG). At last, a Back Propagation Neural Network (BPNN) is trained to classify the ship candidates using the multi-spectral bands. The experiments using Quickbird satellite images show that our approach is effective in ship detection and performs particularly well in separating the ships clustered together and staying alongside the pier.