Search results

So two complementary edge cues [10], that is, the gradient and the template matching with a template derived from the input data are integrated as weights in (1).
Experimental data and setting The experiments are operated on two pair of images.
Three image partitions and their local details of the first group of data, got from PAN, MS and a union set of results of PAN and MS Figure 5.
Table 1 presents the spatial metrics for the second group of data, Table 2 and Table 3 presented the spectral metrics for the first group of data.
Zhang, "Multi-source remote sensing data fusion: status and trends," International Journal of Image and Data Fusion, vol. 1, pp. 5-24, 2010

The multi-layer perceptron (MLP) of the neural network structure was trained to reveal the underlying pattern within the data sample.
The sample data was obtained from a sensitivity analysis drive from FEA.
A sensitivity analysis has been conducted to generate the sample data.
Sample data available was split in half for training and validation.
(Left) Meta-model ANN of the torque ripple output response ( black dots are the sampling data calculated during sensitivity analysis).

b) Retransmission, caused by data collision, loses and so on, is also a source of energy waste.
c) Nodes monitor and receive data which is sent to other nodes, result to unnecessary energy waste.
In this article, we optimize the S-MAC protocol in the side of channel competition and data transmission.
After the node, who own the channel, sending the data, recalculating and setting the priority again.
New frame will replace the original data whose position is near to the head of queue.

The Y-doped films exhibit an elevated electron Fermi level,which may enhance band bending to lower the density of empty trap states.Because of this Y-doping, the Dsscs can alleviate the decay of light to electric energy conversion efficiency due to light intensity reduction.
X-ray photoelectron spectroscopy data were obtained with an ESCALab220i-XL electron spectrometer from VG Scientific using 300 W Al Kα radiation.

Verification of the proposed approach is based on statistical distribution of the measured surface roughness data.
There are few data about how any active surface and its morphology do affect on process of the material shaping into definite product.
Advantage of such a progress is that it enables to measure data resulting from milling operation.
Distribution of all the measured Rz data are compared in Fig. 6, where two limiting cases are shown.
Though Rz data in the 3D contour milling produce similar distribution of the machined surface as that of Morphed Milling, a disadvantage of the 3D contour milling is found out in the very high mean data of about Rz=8.8 μm.

It is noted that the FEA analysis was benchmarked to the hoop strain data at the expense of the radial data.
The SHD stress results were calculated using the strain-data averaging method as described in reference [4].
Again, the largest differences between the DHD data and the reference solution occur at the peak stress locations.
Investigations into potential improvements to the FEA reference solutions, particularly in the elastic-plastic transition region, and the use of the eigenstrain method, reference [5], to improve the DHD data reduction are planned as part of ongoing DHD development and validation studies.
Hill, "Improved Data Reduction for the Deep-Hole Method of Residual Stress Measurement," Journal of Strain Analysis, vol. 38, no. 1, 2003, pp 65-78

Such reduction of the shearing stress is considered to be.
The actual and FEM data of the micro-stage center displacement were compared and the results are shown in Fig. 9.
They also show the deviation of the bite tip of micro-stage displacement between the actual data and FEM analysis data.
The results in Fig. 9 and Table 5 show that the FEM displacement data are similar to the actual displacement data obtained by operating the micro-stage; the error ratio was 3.53%.
The deviation between FEM data and actual data obtained by driving the micro stage was 3.53%, which was less than the allowable engineering deviation with use of FEM; this low deviation value verifies the validity of the FEM used in our study. 5.

Optimization and reduction of zeolite A synthesis costs are the focus of several studies.
XRD data and refinement show that the obtained material presents 99.84% crystallinity, average crystallite size of 54.92 nm, and a semi-quantitative percentage of 79% zeolite A.

This article takes “1+4” the aluminum belt hot tandem rolling line as the object of study, based on the scene deviation data, has researched the deviation regular of tandem rolling strap, has established the BP neural network forecast model of the rolling process.
From the actualdeviation data, the change of rolling parameters in F4 rolling system, often has a decisive impact on the final deviation of the strip.
Then we analyse to the actual rolling production deviation data in one aluminum company.
This aluminum had been rolled at 12:35:34 on June 25th, the average wedge of the product is 3.12mm,average wedge is 5μm, belt speed is 5.16m / s, rolling length is 764m.We can obtain the characteristic value of the steady-state deviation of these data
On basis of distribution rule of actual rolling data, tail deviation was divided into some kinds of patterns, obtained characteristic parameters and mapping matrix by BP neural network to predict tail deviation pattern.

Then, we completed the 3-Dimensional localization after data fusion by weighting the 2-Dimensional localization result.
Table 1 Comparison of actual data with measurement data and measurement error in simulation experiment number test point coordinates error actual measurement 1 67,24,85 67.05,24.00,84.97 0.50 2 327,101,34 327.21,101.07,33.87 0.26 3 136,250,7 135.86,249.78,7.00 0.26 4 312,169,54 312.23,169.07,54.01 0.24 5 88,139,55 87.93,139.00,55.03 0.0 6 370,54,68 370.41,53.79,68.12 0.24 7 409,233,58 409.02,232.96,58.12 0.13 8 141,233,10 141.03,233.02,8.96 1.04 9 432,57,88 432.02,57.11,88.00 0.11 10 127,442,56 127.42,442.25,56.13 0.51 From the Table 1, we can get not only the actual data and measurement data of the 10 test points mentioned above, but also the actual distance and measured distance of them to each base station and the actual angle and measured angle of the receives signals.
In addition, we can calculate the average localization error of this algorithm which is 0.337. 3-Dimensional simulation has been made with the data above using Matlab.
Now, with the same conditions and data as feasibility experiments, localize the test point using the two algorithms, and compare the coordinate errors of these three algorithms, as shown in Figure 6.
In addition, with the same data acquisition environment and data sources, the 3-Dimensional localization algorithm based on weighted fusion with projection coordination has the smallest error.