We implemented a change detection method based on multitemporal object using chi-square and iterative trimming to find the changed in past. Trapping sample data does not obey Gaussian distribution, result is ideal. To fix this problem, we proposed law of cosines with box-whisker plot. First, incremental segmentation segment image selecting optimum combination spatial, spectral, contextual features by Optimum Index Factor method, constructed feature space different time images. Then, was used...

This paper presents a novel method for hyperspectral image classification based on the minimum noise fraction (MNF) and an approach combining support vector machine (SVM) linear discriminant analysis (LDA). A new SVM/LDA algorithm is used classification. First, we use MNF to reduce dimension extract features of image, then transform extracted features. Next, train result transformation, optimize parameters through cross-validation grid search method, get optimal classifier. Finally, this...

The digital elevation model (DEM) generates a simulation of ground terrain in certain range with the usage 3D point cloud data. It is an important source spatial modeling information. Due to various reasons, however, generated DEM has data holes. Based on algorithm deep learning, this paper aims train generation (DGM) complete void filling task. A amount and randomly mask are taken as network inputs, along which reconstruction loss generative adversarial (GAN) used assist training, so...

A hyperspectral image classification method based on probabilistic weighted fusion of multiple spectral-spatial features is proposed in this letter. First, dimensionality reduction and feature extraction images are conducted by minimum noise fraction. Then, two composed through a combination texture multiscale morphological with characteristic obtained Next, support vector machine classifier employed to classify each feature. Finally, model established applied for the output every single The...

The robust and rapid matching of oblique UAV images urban area remains a challenge until today. method proposed in this paper, Nicer Affine Invariant Feature (NAIF), calculates the image view an by making full use rough Exterior Orientation (EO) elements image, then recovers to rectified doing inverse affine transform, left over SIFT method. significance test left-right validation have applied process reduce rate mismatching. Experiments conducted on demonstrate that NAIF takes about same...

This paper proposes a method of semiautomatic right-angle building extraction from very high resolution (VHR) aerial images, based on graph cuts with star shape constraint and regularization. In the proposed method, first, image block containing target is obtained by seed line humans. Next, preprocessed bilateral filtering then simple linear iterative clustering oversegmentation was used to segment into small regions. Then regions that contain pixels belonging are merged as foreground area,...

For the calculation of bulk volume, conventional method is to use RTK or total station measure a certain density three-dimensional coordinate points on surface heap, and then grid DTM calculate volume. However, in practice, many objects be measured are large inconvenient at close range, so methods difficult adapt this time. In addition, obtained by low, which affects accuracy result, cost-effectiveness ratio low. Fortunately, field computer vision, Structure from Motion Patch-based...

In order to overcome errors caused by season, shooting angle, and other factors in multitemporal remote sensing image change detection, a method based on old-temporal vector data new-temporal is proposed under the premise that changes of images are normally less than unchanged ones. Getting object through incremental segmentation constraints previous data, we extract its textural spectral features get dataset transform principal component analysis. After this, isolation forest used calculate...

This paper proposes a method of hyperspectral image dimensionality reduction based on automatic subspace partition, k-means clustering mutual information and adaptive band selection. first division is used to determine the initial subspace, in various through between variance K - means center from two adjacent selection their difference absolute minimum as boundary delimit molecular space, then each obtained by applying index, get biggest index big small order according at last three bands....

The change detection method based on multi-temporal object was implemented by chi-square test and Gaussian distribution iteration to find the changed in past. However, trapped sample data does not obey distribution, effect is ideal. In order fix this problem, a law of cosines with box-whisker plot proposed. First, feature space different time images constructed. Then, used calculate index every object. objects are identified through analyzing at last. High-resolution remote sensing GF-1 as...

Abstract The quantitative study of terrestrial water storage change (TWSC) and its spatial temporal dynamics has been an important topic in resources research management. In this study, we invert the characteristics TWSC China based on GRACE time-varying gravity field model explore attribution analysis, using Theil- Sen Median trend, Mann-Kendall test geographic detector, as well geographically temporally weighted regression (GTWR), combined with meteorological drought indicators other data...

In order to retrieve the positioning image efficiently and quickly from a large number of different images realize three-dimensional spatial positioning, in this article, based on photogrammetry computer vision theory, new method big data under bag words model guidance is proposed. The consists two parts: retrieving positioning. First, complete retrieval by feature extraction, K-means clustering, building other processes, thus improve efficiency matching. Second, achieve interior exterior...

In this paper, based on the theory of down-scaling, we propose methods for linear mixed model disaggregate pixels in coarse resolution images. Exploiting information about within pixel each component fractional cover derived from high spatial classification map, Sub-pixel reflectance different land-cover classes are calculated by solving a system equations image and producing subpixel level NDVI time series curve component. Results showed that application algorithm provided good estimates...

Camera calibration is essential to obtaining three-dimensional information from two-dimensional image, this paper combines the method of photogrammetry and computer vision, put forward a kind camera self-calibration based on hierarchical reconstruction bundle adjustment. The projective obtained by SVD measurement matrix, Kruppa equation are deduced for calculating parameters, then upgrade Euclidean reconstruction. Executing overall optimization solve inner orientation elements lens...

For features extraction of image measurement and reconstruction from industrial parts images, this paper puts forward a mathematical model guided by CAD designed data which can detect lines with high accuracy. Firstly, based on getting line initial values projecting object through the conversion 3D coordinate image, then extracting sequence edge points least squares template matching (LSTM) acquiring equations principle fitting, finally precisely obtaining contours solving intersections...