Influence Of Image Resolution Research Articles

A Remote Sensing Method for Crop Mapping Based on Multiscale Neighborhood Feature Extraction

Obtaining accurate and timely crop mapping is essential for refined agricultural refinement and food security. Due to the spectral similarity between different crops, the influence of image resolution, the boundary blur and spatial inconsistency that often occur in remotely sensed crop mapping, remotely sensed crop mapping still faces great challenges. In this article, we propose to extend a neighborhood window centered on the target pixel to enhance the receptive field of our model and extract the spatial and spectral features of different neighborhood sizes through a multiscale network. In addition, we also designed a coordinate convolutional module and a convolutional block attention module to further enhance the spatial information and spectral features in the neighborhoods. Our experimental results show that this method allowed us to obtain accuracy scores of 0.9481, 0.9115, 0.9307 and 0.8729 for OA, kappa coefficient, F1 score and IOU, respectively, which were better than those obtained using other methods (Resnet-18, MLP and RFC). The comparison of the experimental results obtained from different neighborhood window sizes shows that the spatial inconsistency and boundary blurring in crop mapping could be effectively reduced by extending the neighborhood windows. It was also shown in the ablation experiments that the coordinate convolutional and convolutional block attention modules played active roles in the network. Therefore, the method proposed in this article could provide reliable technical support for remotely sensed crop mapping.

Deep Learning Based Solar Flare Forecasting Model. II. Influence of Image Resolution

Due to the accumulation of solar observational data and the development of data-driven algorithms, deep learning methods are widely applied to build a solar flare forecasting model. Most of the works focus on how to design or select proper deep networks for the forecasting task. Nevertheless, the influence of image resolution on the learning based solar flare forecasting model has not been analyzed and discussed. In this Paper, we investigate the influence of the resolution of magnetograms on the accuracy of solar flare forecasting. We study the active regions by the Solar Dynamics Observatory/Helioseismic and Magnetic Imager (SDO/HMI) magnetograms from 2010 to 2019. Then, we downsample them to get a database containing active regions with several resolutions. Afterwards, three deep neural networks (i) AlexNet, (ii) ResNet-18, and (iii) SqueezeNet are implemented to evaluate the performance of solar flare forecasting compared to different resolutions of magnetogram. In experiments, we first did comparative experiments on our own simulated HMI database with different resolutions. Then we conducted experiments on two selected actual overlapping databases, Hinode–HMI and Michelson Doppler Imager–HMI, to reconfirm our conclusions. The experiment results show that all the selected deep learning networks are insensitive to the resolution to a certain extent. We visualized the regions of interest of the network from an interpretable perspective and found that the deep learning network pays more attention to the global features extracted from active regions that are not sensitive to local information in magnetograms.

Retrieval of rapeseed leaf area index using the PROSAIL model with canopy coverage derived from UAV images as a correction parameter

Leaf area index (LAI), which is an important structural parameter, plays a vital role in evaluating crop growth and yield. In this study, we used the canopy coverage (CC) derived from unmanned aerial vehicle (UAV) images as a correction parameter in the PROSAIL model coupled with a neural network (NN) to improve the accuracy of LAI inversion of rapeseed plots. CC had a significantly positive impact on the accuracy of LAI inversion especially in sparse canopy structure with the 22.24% decrease in the entire dataset and 35.76% decrease in the sparse canopy dataset. We then compared the inversion performances of an empirical statistical model (ESM) based on a vegetation index and the PROSAIL model incorporating CC correction for 2016 and 2018 datasets. The ESM performed better in modeling the 2016 dataset, but its accuracy was much lower for the 2018 dataset (2016: NRMSE = 0.131; 2018: NRSME = 0.348). Overall, the PROSAIL model was more robust over these two datasets (2016: NRMSE = 0.152; 2018: NRMSE = 0.168). In addition, the original-resolution images were resampled to six coarse resolutions to evaluate the influence of image resolution on the LAI inversion performance of the PROSAIL model. When pixel size increased to more than 10 cm, the inversion accuracy began to decrease dramatically. In conclusion, introducing a canopy coverage correction parameter in the PROSAIL model improved its performance in retrieving rapeseed LAI.

Ultrasound image resolution influences analysis of skeletal muscle composition.

Analysis of muscle composition using ultrasound requires standardization of several equipment settings (i.e. gain). However, the influence of image resolution, which is altered by imaging depth, on measures of muscle composition is unknown. We analysed rectus femoris muscle composition using ultrasound images captured from 32 males and females (aged 28±5years) at depths of 9.0, 7.3, 5.9 and 4.7cm. The transducer's orientation was fixed using a clamp during image acquisition to minimize movement. Across each image resolution, a region of interest encompassing the same anatomical area within the muscle was used for muscle composition analysis. Muscle composition was analysed using a combination of first-, second- and higher-order texture features. Muscle composition agreement across image resolutions was evaluated using a one-way ANOVA and intraclass correlation coefficients (ICC). Most muscle composition features displayed differences due to image resolution (p<.05). ICCs demonstrated poor-to-good agreement across different image resolutions. In general, higher resolution images (i.e. shallower imaging depth) demonstrated better agreement (ICC>0.90) compared to lower resolution images. Ultrasound image resolution influences muscle composition analysis. Image resolution should be fixed within and between individuals when evaluating muscle composition using ultrasound.

Effects of UAV Image Resolution, Camera Type, and Image Overlap on Accuracy of Biomass Predictions in a Tropical Woodland

Unmanned aerial systems (UASs) and photogrammetric structure from motion (SFM) algorithms can assist in biomass assessments in tropical countries and can be a useful tool in local greenhouse gas accounting. This study assessed the influence of image resolution, camera type and side overlap on prediction accuracy of biomass models constructed from ground-based data and UAS data in miombo woodlands in Malawi. We compared prediction accuracy of models reflecting two different image resolutions (10 and 15 cm ground sampling distance) and two camera types (NIR and RGB). The effect of two different side overlap levels (70 and 80%) was also assessed using data from the RGB camera. Multiple linear regression models that related the biomass on 37 field plots to several independent 3-dimensional variables derived from five UAS acquisitions were constructed. Prediction accuracy quantified by leave-one-out cross validation increased when using finer image resolution and RGB camera, while coarser resolution and NIR data decreased model prediction accuracy, although no significant differences were observed in absolute prediction error around the mean between models. The results showed that a reduction of side overlap from 80 to 70%, while keeping a fixed forward overlap of 90%, might be an option for reducing flight time and cost of acquisitions. Furthermore, the analysis of terrain slope effect in biomass predictions showed that error increases with steeper slopes, especially on slopes greater than 35%, but the effects were small in magnitude.

Influence of image resolution on the performance of remote breathing rate measurement using thermal imaging technique

The aim of this work was to evaluate the influence of a thermal image resolution on the performance of thermal imaging system for remote and unobtrusive breathing rate (BR) measurement. The thermal imager with original resolution of 640 × 480 was first used to establish the thermal imaging system. The thermal video database containing 60 videos was developed to validate the feasibility of this system for BR measurement. Subsequently, the nearest-neighbor interpolation was exerted to create the thermal videos with low resolutions. The Bland-Altman plots of BR showed that distances between the upper and lower 95% limits of agreements were 4.094 bpm, 10.595 bpm, 10.091 bpm, 15.168 bpm and 20.769 bpm for video resolutions of 640 × 480, 512 × 384, 320 × 240, 192 × 144 and 64 × 48, respectively. With respect to linear regression analysis, the determination coefficient (R2) of 0.898, 0.890, 0.900, 0.776 and 0.539 had been found for the resolutions of 640 × 480, 512 × 384, 320 × 240, 192 × 144 and 64 × 48, respectively. These two statistical approaches both demonstrated that the thermal resolution of 320 × 240 achieved the performance comparable to the high thermal resolution. Inspired by the above results, for the practical applications, we can choose the appropriate resolution by taking both the system performance and cost into account. In this case, the thermal video resolution of 320 × 240 could be selected for BR measurement in some application scenarios (i.e. lie detection) where the measurement distance is approximately 1.5 m. This research may greatly improve the design of the contactless measurement device for physiological signals.

Postural Photogrammetry as Promising Tool for Clinical Use: A Reliability Study

Many studies on postural photogrammetry had reported various intra-class correlation coefficients (ICC) across postural variable measurements, however no conclusive solution was given. This reliability and cross-sectional study was done in June 2016 at the Faculty of Medicine and Health Sciences, Universiti Malaysia Sabah. A total of 24 male adult subjects with mean age 28.5 years (± 4.8 years), body mass 24.97 kg (± 3.85 kg) and height 166.6 cm (± 6 cm) were evaluated for standing postural photogrammetry. Four sets of manually digitized posture image files (by 4 raters) were measured and statistically analyzed for inter-rater agreement as well as the influence of image resolution and camera height from the floor on various postural variable measurements. The ICC between 4 raters for all postural variables was excellent (the lowest ICC was 0.940 for Q Angle of the Right Knee measurements). Two-Way ANOVA showed that postural variable measurements were not affected by either image resolution or camera height from the floor. Scrupulous attempts done on standing postural photogrammetry amplified the potential for standing postural evaluation in clinical settings.

Evaluation of the U.S. Geological Survey Landsat Burned Area Essential Climate Variable across the Conterminous U.S. Using Commercial High-Resolution Imagery

The U.S. Geological Survey has produced the Landsat Burned Area Essential Climate Variable (BAECV) product for the conterminous United States (CONUS), which provides wall-to-wall annual maps of burned area at 30 m resolution (1984–2015). Validation is a critical component in the generation of such remotely sensed products. Previous efforts to validate the BAECV relied on a reference dataset derived from Landsat, which was effective in evaluating the product across its timespan but did not allow for consideration of inaccuracies imposed by the Landsat sensor itself. In this effort, the BAECV was validated using 286 high-resolution images, collected from GeoEye-1, QuickBird-2, Worldview-2 and RapidEye satellites. A disproportionate sampling strategy was utilized to ensure enough burned area pixels were collected. Errors of omission and commission for burned area averaged 22 ± 4% and 48 ± 3%, respectively, across CONUS. Errors were lowest across the western U.S. The elevated error of commission relative to omission was largely driven by patterns in the Great Plains which saw low errors of omission (13 ± 13%) but high errors of commission (70 ± 5%) and potentially a region-growing function included in the BAECV algorithm. While the BAECV reliably detected agricultural fires in the Great Plains, it frequently mapped tilled areas or areas with low vegetation as burned. Landscape metrics were calculated for individual fire events to assess the influence of image resolution (2 m, 30 m and 500 m) on mapping fire heterogeneity. As the spatial detail of imagery increased, fire events were mapped in a patchier manner with greater patch and edge densities, and shape complexity, which can influence estimates of total greenhouse gas emissions and rates of vegetation recovery. The increasing number of satellites collecting high-resolution imagery and rapid improvements in the frequency with which imagery is being collected means greater opportunities to utilize these sources of imagery for Landsat product validation.

Saliency detection on sampled images for tag ranking

Image saliency contributes to rank the unordered tags extracted from social media, but the existing saliency detection methods can hardly efficiently handle massive images in tag ranking. In this paper, we focus on improving the efficiency of saliency detection methods by applying them on the sampled images with suitable resolutions. We extensively investigate the influence of image resolution to saliency detection performance of the typical methods, and summarize a sampling strategy for different categories of salient object detection methods. Furthermore, we validate the effectiveness of the sampling strategy by applying the salient object detection methods on the sampled images with the selected resolutions in tag ranking. The experimental results show that sampling can significantly improve the efficiency of the existing salient object detection methods without obvious loss in effectiveness.

Electrical and fluid transport in consolidated sphere packs

We calculate geometrical and transport properties (electrical conductivity, permeability, specific surface area, and surface conductivity) of a family of model granular porous media from an image based representation of its microstructure. The models are based on the packing described by Finney and cover a wide range of porosities. Finite difference methods are applied to solve for electrical conductivity and hydraulic permeability. Two image processing methods are used to identify the pore-grain interface and to test correlations linking permeability to electrical conductivity. A three phase conductivity model is developed to compute surface conductivity associated with the grain-pore interface. Our results compare well against empirical models over the entire porosity range studied. We conclude by examining the influence of image resolution on our calculations.

Multi-temporal imaging using an unmanned aerial vehicle for monitoring a sunflower crop

The objective of this study is to determine the capability of an unmanned aerial vehicle system carrying a multispectral sensor to acquire multitemporal images during the growing season of a sunflower crop. Measurements were made at different times of the day and with different resolutions to estimate the normalised difference vegetation index (NDVI) and study its relationship with several indices related to crop status with the aim of generating useful information for application to precision agriculture techniques. NDVI was calculated from images acquired on four different dates during the cropping season. On two of these dates, two images were acquired to determine how the time of day when the images were taken influences NDVI value. To study the influence of image resolution on NDVI, the original images were resampled to 30 × 30 and 100 × 100 cm pixel sizes. The results showed that the linear regressions between NDVI and grain yield, aerial biomass and nitrogen content in the biomass were significant at the 99% confidence level, except during very early growth stages, whereas the time of day when the images were acquired, the classification process, and image resolution had no effect on the results. The methodology provides information that is related to crop yield from the very early stages of growth and its spatial variability within the crop field to be harvested, which can subsequently be used to prescribe the most appropriate management strategy on a site-specific basis.

Blood Oxygen Level–Dependent Magnetic Resonance Imaging of the Kidneys

The aim of this study was to quantify the influence of image resolution on the apparent transverse relaxivity (R2*) of the magnetic resonance (MR) signal in human renal tissue in vivo and in phantom measurements. This prospective study included 17 healthy volunteers (age, 32 ± 8 years, 6 women). Parametrical R2* maps were computed via monoexponential fitting of multiecho 2-dimensional fast-field echo data measured at 1.5 T (repetition time [TR], 150 milliseconds; flip angle [FA], 40°; minimum echo time [TE], 4.6 milliseconds; ΔTE, 5 milliseconds; 16 echoes) and at 3 T (TR, 140 milliseconds; FA, 70°; minimum TE, 2 milliseconds; ΔTE, 5 milliseconds; 16 echoes) with varying nominal volumes of the encoded voxels (from 5.76 to 36.0 mm). For each voxel size, mean R2* values were computed in regions of interest drawn in the left and right renal parenchyma. For data acquired using minimum voxel size, the mean R2* values were computed over the cortex and medulla separately. The squared 2-norm of the residuals was computed to evaluate the goodness of the pixel-wise exponential fits. Six multiecho MR images of a water phantom were acquired using a 2-dimensional fast-field echo sequence (FA, 50°; TR, 108 milliseconds; TE, 4 milliseconds; ΔTE, 20 milliseconds) at 3 T after shim adjustment and in the presence of a uniform background gradient of 40 μT/m. The nominal voxel size was varied in a range between 2 and 12.5 mm. Mean R2* values of 13.04 ± 0.71 s (right renal cortex) and 16.47 ± 1.92 s (right renal medulla) were computed at 1.5 T. At 3 T, the R2* of the right medulla was 28.27 ± 1.52 s and the cortical R2* was 19.22 ± 2.32 s. Comparable relaxivity values were found over the left kidney at both field strengths. Increasing R2* values were observed for increasing voxel volume in both the water phantom and renal tissue data. At a constant slice thickness of 4 mm, the decrease in the in-plane resolution from 1.2 × 1.2 mm to 3.0 × 3.0 mm led to a maximum increase of the renal R2* of 15% at 1.5 T and of 12% at 3 T. Increasing the slice thickness from 3 to 8 mm at a constant in-plane resolution of 1.5 × 1.5 mm resulted in a maximum increase of the renal R2* of 30% at 1.5 T and of 26% at 3 T. On the other hand, increasing the voxel size improved the goodness of the fit implied by the smaller residuals. The phantom experiments and in vivo acquisitions of healthy renal tissue documented a significant dependence of the apparent R2* relaxation rate on the spatial resolution of the MR imaging data. In clinical practice, the voxel volume for the quantification of renal R2* should be optimized in a compromise between minimizing the effects of macroscopic field inhomogeneity and maintaining a sufficiently high signal-to-noise ratio and goodness of fit. When comparing quantitative R2* among different publications, the influence of the spatial resolution should be taken into account.

Fractal analysis and its impact factors on pore structure of artificial cores based on the images obtained using magnetic resonance imaging

Pore structure is one of important factors affecting the properties of porous media , but it is difficult to describe the complexity of pore structure exactly. Fractal theory is an effective and available method for quantifying the complex and irregular pore structure. In this paper, the fractal dimension calculated by box-counting method based on fractal theory was applied to characterize the pore structure of artificial cores. The microstructure or pore distribution in the porous material was obtained using the nuclear magnetic resonance imaging (MRI). Three classical fractals and one sand packed bed model were selected as the experimental material to investigate the influence of box sizes, threshold value, and the image resolution when performing fractal analysis. To avoid the influence of box sizes, a sequence of divisors of the image was proposed and compared with other two algorithms (geometric sequence and arithmetic sequence) with its performance of partitioning the image completely and bringing the least fitted error. Threshold value selected manually and automatically showed that it plays an important role during the image binary processing and the minimum-error method can be used to obtain an appropriate or reasonable one. Images obtained under different pixel matrices in MRI were used to analyze the influence of image resolution. Higher image resolution can detect more quantity of pore structure and increase its irregularity. With benefits of those influence factors, fractal analysis on four kinds of artificial cores showed the fractal dimension can be used to distinguish the different kinds of artificial cores and the relationship between fractal dimension and porosity or permeability can be expressed by the model of D = a − b ln( x + c ). ► The proposed sequence of divisors in BC method can eliminate the border effect. ► Appropriate threshold value can be obtained using the minimum-error method. ► Higher image resolution brings more accurate estimation of fractal dimension. ► Fractal dimension can be used to distinguish the different kinds of artificial cores. ► Relationship between FD and porosity or permeability is expressed by D = a − b ln( x + c ).

APPLICATION OF IMAGE ANALYSIS TO ASSESSING CRITICAL PORE SIZE FOR PERMEABILITY PREDICTION OF CEMENT PASTE

This paper explores image analysis techniques that provide insight into the nature of pore structure as observed in backscattered electron images of polished sections. On the basis of mathematical morphology, the pore size distribution is characterised and the critical pore size is determined for cement paste at different hydration time. The influence of image resolution is investigated. The permeability of cement paste can be predicted according to General Effective Media (GEM) theory. Comparison between permeability estimation and experiment results reveals good agreement.

A new algorithm for automatic Rumex obtusifolius detection in digital images using colour and texture features and the influence of image resolution

In Gebhardt et al. (2006) an object-oriented image classification algorithm was introduced for detecting Rumex obtusifolius (RUMOB) and other weeds in mixed grassland swards, based on shape, colour and texture features. This paper describes a new algorithm that improves classification accuracy. The leaves of the typical grassland weeds (RUMOB, Taraxacum officinale, Plantago major) and other homogeneous regions were segmented automatically in digital colour images using local homogeneity and morphological operations. Additional texture and colour features were identified that contribute to the differentiation between grassland weeds using a stepwise discriminant analysis. Maximum-likelihood classification was performed on the variables retained after discriminant analysis. Classification accuracy was improved by up to 83% and Rumex detection rates of 93% were achieved. The effect of image resolution on classification results was investigated. The eight million pixel images were upscaled in six stages to create images with decreasing pixel resolution. Rumex detection rates of over 90% were obtained at almost all resolutions, and there was only moderate misclassification of other objects to RUMOB. Image processing time ranged from 45 s for the full resolution images to 2.5 s for the lowest resolution ones.

Method to classify xylem cell types using cross sectionsof 1-year-old branches in Mediterranean woody species

Measurements of anatomical parameters of wood are of great interest both for eco-physiological purposes and for technological applications. The aim of this paper is to describe a new method for classifying and measuring cell lumen of xylem, analysing cross sections under the light microscope. The proposed method is based on the application of digital image analysis on images of the cross sections of xylem in combination with graphical and statistical methods. The methodology was tested on 1-year-old branches of several woody species, both trees and shrubs, occurring in a Mediterranean natural ecosystem in southern Italy. The development of the procedure was based on statistical comparison between data collected according to four procedures: (a) manual identification and measurement of lumen diameter of conduits on longitudinal sections; (b) manual identification and measurement of lumen diameter of conduits on cross sections; (c) manual identification and measurement of lumen area of conduits on cross sections; and (d) automatic measurement of lumen area of conduits on cross sections. The influence of image resolution and that of the position of the selected area on cell classification and measurements were ascertained. The proposed method was proved to be specific to woody structures and allowed the construction of a model-graph that is species-specific. Interpretation of the model-graphs allows classification and hence measurement of identified cells.

Influence of image resolution and evaluation algorithm on estimates of the lacunarity of porous media

In recent years, experience has demonstrated that the classical fractal dimensions are not sufficient to describe uniquely the interstitial geometry of porous media. At least one additional index or dimension is necessary. Lacunarity, a measure of the degree to which a data set is translationally invariant, is a possible candidate. Unfortunately, several approaches exist to evaluate it on the basis of binary images of the object under study, and it is unclear to what extent the lacunarity estimates that these methods produce are dependent on the resolution of the images used. In the present work, the gliding-box algorithm of Allain and Cloitre [Phys. Rev. A 44, 3552 (1991)] and two variants of the sandbox algorithm of Chappard et al. [J. Pathol. 195, 515 (2001)], along with three additional algorithms, are used to evaluate the lacunarity of images of a textbook fractal, the Sierpinski carpet, of scanning electron micrographs of a thin section of a European soil, and of light transmission photographs of a Togolese soil. The results suggest that lacunarity estimates, as well as the ranking of the three tested systems according to their lacunarity, are affected strongly by the algorithm used, by the resolution of the images to which these algorithms are applied, and, at least for three of the algorithms (producing scale-dependent lacunarity estimates), by the scale at which the images are observed. Depending on the conditions under which the estimation of the lacunarity is carried out, lacunarity values range from 1.02 to 2.14 for the three systems tested, and all three of the systems used can be viewed alternatively as the most or the least "lacunar." Some of this indeterminacy and dependence on image resolution is alleviated in the averaged lacunarity estimates yielded by Chappard et al.'s algorithm. Further research will be needed to determine if these lacunarity estimates allow an improved, unique characterization of porous media.

A basic study of a face-to-face image presentation system

Recently, broadband networks have made great progress and various interactive visual communication systems including video conferencing, video telephones and video networks for meetings are becoming increasingly popular. However, unnatural eye contact in the displays used by these communication systems has presented a barrier to natural conversation between communicators. We confirmed the possibility of realizing the feeling of eye contact that is part of natural face-to-face conversation. This paper also describes experimental results on the permissible eye-line angle that gives a sense of eye contact and the influence of image resolution.

Analysis of the influence of image resolution on the discriminating power of trabecular bone architectural parameters

In the study the influence of image resolution on the discriminating power of different parameters quantifying the architecture of trabecular bone is investigated. High-resolution images of 200-μm-thick sections of young and old trabecular bone are collected. Different architectural parameters are evaluated and shown to be statistically different in both groups of images. Then the resolution of the images is artificially degraded to the level comparable with resolution achievable in vivo. It is shown that although the errors of evaluation are quite large for low-resolution images, the statistical difference present in the original data is still observed for parameters that depend only on the global characteristics of trabecular structure or marrow space, or depend only linearly on the number and area of disconnected parts of marrow space or trabecular structure. The parameters that fulfill such conditions are Euler number, mean area of marrow cavities, star volume and trabecular spacing. The statistical difference disappears for parameters that depend on the number and area of disconnected parts of marrow space or trabecular structure in a nonlinear manner, that is, marrow and trabecular disconnection probabilities, two-point distance and second moment of marrow cavities area distribution. It is shown that the mechanism that leads to the losing of discriminating power does not depend crucially on the estimation error or noise level. This mechanism, which involves an interplay between changes of mean values and standard deviations, is the manifestation of partial volume effect amplified with image segmentation.

Description and classification of soil structure using distance transform data

SummaryClassification of soil based on structure is useful for conveying information about physical properties and soil processes. The distance transform is an image analysis technique suitable for quantifying soil structure. An analysis of distance transform data, in the form of cumulative area distribution curves for previously published images of soil structures of various types, is presented. The images were used to derive a quantitative classification of structure using maximum distance of solid from a macropore (Dmax, measured), the distance from macropore space containing 50% of the solid area (k, derived by fitting a sigmoidal function to the cumulative area distribution curve), the total interface length between pore and solid per area of sample (IA, measured), the porosity or the proportion of pores per area of sample (PA, measured) and the pore distribution characteristic (n, derived by fitting a sigmoidal function to the cumulative area distribution curve) which is related to the number, continuity and distribution of pores. The influence of image resolution was investigated, and within limits found to be fairly small. The final classification of soil structure was based on the hypothesized relations between the descriptors and structure‐forming processes.