数据挖掘技术在采摘机器人图像采集过程应用

针对采摘机器人果实识别速率较低导致采摘效率较低的问题,对数据挖掘技术在采摘机器人中图像采集过程的应用进行了分析。采摘机器人主要组成包括图像采集模块、运动控制模块、气压驱动模块、电源模块、微处理器模块和无线网传输模块。为了提升图像数据的处理速度,采用MR模型和决策树中的ID_(3)算法对图像数据进行处理,并构建决策树模型,对图像数据进行数据挖掘处理。为了验证该采摘机器人的性能,对其进行数据挖掘算法调试试验和采摘机器人性能试验,结果表明:该图像处理算法速度显著提升,采摘机器人性能稳定,采摘效果好。

Pathological and MR-DWI study of the acute hepatic injury model after stem cell transplantation

AIM： To investigate apparent diffusion coefficient （ADC） values as an indication of reconditioning of acute hepatic injury （AHI） after allogeneic mononuclear bone marrow cell （MBMC） transplantation. METHODS： Three groups were used in our study： a cell transplantation group （n = 21）, transplantation control group （n = 21） and normal control group （n = 10）. AHI model rabbits in the cell transplantation group were injected with 5 mL of MBMC suspension at multiple sites in the liver and the transplantation controls were injected with 5 mL D-Hanks solution. At the end of the 1st, 2nd and 4th wk, 7 rabbits were randomly selected from the cell transplantation group and trans- plantation control group for magnetic resonance diffu- sion-weighted imaging （MR-DWI） and measurement of the mean ADC values of injured livers. After MR-DWI examination, the rabbits were sacrificed and the livers subjected to pathological examination. Ten healthy rab- bits from the normal control group were used for MR- DWI examination and measurement of the mean ADC value of normal liver. RESULTS： At all time points, the liver pathological scores from the cell transplantation group were significantly lower than those in the transplantation control group （27.14± 1.46 vs 69.29± 6.16, 22.29 ± 2.29 vs 57.00± 1.53, 19.00 ± 2.31 vs 51.86 ± 6.04, P = 0.000）. The mean ADC values of the cell transplantation group were significantly higher than the transplantation con- trol group （（1.07± 0.07） ×10^-3 mm2/s vs （0.69 ± 0.05） ×10^-3 mm2/s, （1.41± 0.04） ×10^-3 mm2/s vs （0.84± 0.03） ×10^-3 mm2/s, （1.68 ± 0.04） ×10^-3 mm2/s vs （0.86± 0.04） ×10^-3 mm2/s, P = 0.000）. The pathological scores of the cell transplantation group and transplantation control group gradually decreased. However, their mean ADC values gradually increased to near that of the normal control. At the end of the 1st wk, the mean ADC values of the cell transplantation group and transplantation control group were significantly lower than those of the normal control group [（1.07 ± 0.07） ×10^-3 mm2/s vs （± 0.03） ×10^-3 mm2/s, （0.69± 0.05） ×10^-3 mm2/s vs （1.76 ± 0.03） ×10^-3 mm2/s, P = 0.000]. At any 2 time points, the pathological scores and the mean ADC values of the cell transplantation group were significantly different （P = 0.000）. At the end of the 1st wk, the pathological scores and the mean ADC values of the transplantation control group were significantly different from those at the end of the 2nd and 4th wk （P = 0.000）. However, there was no significant difference between the 2nd and 4th wk （P = 0.073 and 0.473, respectively）. The coefficient of correlation between the pathological score and the mean ADC value in the cell transplantation group was -0.883 （P = 0.000） and -0.762 （P = 0.000） in the transplantation control group. CONCLUSION： Tracking the longitudinally dynamic change in the mean ADC value of the AHI liver may reflect hepatic injury reconditioning after allogeneic MBMC transplantation.

A New Model Parameter Identification Technique for Magnetorheological Dampers

Magnetorheological （MR） Dampers offer rapid variation in damping properties, making them ideal in semi-active control of structures. They potentially offer highly reliable operation and can be viewed as fail safe, in that in the worst case, they become passive dampers. Perfect understanding of the response is necessary when implementing these in operation in conjunction with a control mechanism. There are many models used to predict the behavior of MR dampers. One of these is the Bouc-Wen model. It is extremely popular as it is numerically tractable, very versatile and can exhibit a wide range of hysteretic behavior. It is necessary to first identify the characteristic parameters of the model before response prediction is possible. However, characteristic parameters identification of the Bouc-Wen model needs an experimental base, which has its own limitations. The extraction of these characteristic parameters by trial and error and optimization techniques leaves significant difference between observed and simulated results. This paper deals with a new approach to extract characteristic parameters for the Bouc-Wen model.

MR image denoising method for brain surface 3D modeling

Three-dimensional(3D) modeling of medical images is a critical part of surgical simulation. In this paper, we focus on the magnetic resonance(MR) images denoising for brain modeling reconstruction, and exploit a practical solution. We attempt to remove the noise existing in the MR imaging signal and preserve the image characteristics. A wavelet-based adaptive curve shrinkage function is presented in spherical coordinates system. The comparative experiments show that the denoising method can preserve better image details and enhance the coefficients of contours. Using these denoised images, the brain 3D visualization is given through surface triangle mesh model, which demonstrates the effectiveness of the proposed method.