Factors including welding sequence being input of the network , and welding distortions being output can make the resulting network which is used for the model of welding distortion and its fast forecasting 以焊接顺序等因素作为网络输入，以钢结构板材凹凸变形为输出，训练生成网络焊接变形模型，可用于预测船舶钢结构焊接变形量的快速预测。
To deal with the self - regulation problem coming from the accidental fault of the measurement sensor , the hot - deformation model of air preheater built by elman neural network is used to replace the measurement function of gap sensor 为了解决测量传感器意外损坏后的自动调节问题，我们使用elman动态神经网络建立了空气预热器的热变形模型，用来在故障情况下代替间隙传感器的测量功能。
Though the result edge drawn by the deformable model will be influenced by a large amount of noises , the model has merits that traditional methods do n ' t have . the principal merit is that the object dealt with by the model is a curve in succession 虽然大量的噪声会影响gvfsnake模型的边缘提取结果，但变形模型有传统模型无法相比的优点，即处理的对象本身就是一个连续的曲线，从而避免了检测出边缘后进行连接造成的错误。
Based on the above purpose , in this paper , aiming at the questions in the past landslide prediction , such as sole research technique , backward method and theory achievement with weak practical , application and so on , time prediction and evaluation of landslides is emphatically researched from quantitative , qualitative and numerical modeling three aspects on the basis of widely consulting the domestic and foreign documentation and material related to landslides forecast . according to landslide monitor material , landslide quantitative prediction is realized by using landslide prediction and evaluation models ; on the basis of experts " experience knowledge , landslide qualitative prediction is realized by using the method and way of expert system ; landslide numerical prediction is realized by using simulation method to establish geology - mechanics - distortion model ( gmd ) 基于上述目的，本论文针对过去滑坡预报的研究方法单一、手段落后，理论成果的实际应用性不强等问题，在广泛查阅国内外有关滑坡预测预报文献资料的基础上，着重从定量、定性和数值模拟三个方面对滑坡的时间预测预报开展研究：依据滑坡的监测资料，运用滑坡预测预报模型实现了滑坡的定量预报；以专家的经验知识为依据，运用专家系统的方法和手段实现了滑坡的定性预报；利用数值模拟手段，通过建立滑坡的地质?力学?变形模型( gmd模型)实现了滑坡的数值预报。
In the application of mini - micro - robot visual perception , there is a need for fish - eye lenses for capturing wide field of view for navigation . though fish - eye lenses provide a wide field of view ( 180 ) , they introduce significant distortion in images and the acquired images are quite warped , which makes conventional camera calibration algorithms no longer work well . this paper presents an accurate calibration framework for fish - eye lens ( a high distortion lens ) camera stereo vision system . the accurate calibration model is formulated with radial distortion , decentering distortion and thin prism distortion based on the fisheye deformation model . using fish - eye and non - linear camera model , the author employs levenberg - marquardt method to realize precise non - linear calibration for wide - view - scene dense depth image recovery 鱼眼镜头成像立体视觉系统在微小型机器人视觉导航和近距离大视场物体识别与定位中有着广泛的应用.尽管鱼眼镜头摄像机具有很大的视场角(接近180 ) ，但同时也引入严重的图像变形，常规的摄像机标定方法无法使用.该文提出一种标定鱼眼镜头摄像机立体视觉系统的方法.在鱼眼镜头变形模型的基础上，通过考虑鱼眼镜头成像的径向变形、偏心变形和薄棱镜变形，建立了鱼眼镜头成像的精确成像模型,然后，利用非线性迭代算法，精确求解摄像机外部参数、内部参数.实验表明，使用该方法得到的立体视觉系统参数满足精确恢复大场景稠密深度图的要求
Three factors , that beam reinforcement relative diameter , concrete intensity , axial compression ratio are considered in the model of slip of longitudinal reinforcement passing through joint , and the correlation between shear deformation and cycle number are simulated four factors that concrete intensity , axial compression ratio , shear compression ratio , hoop reinforcement characteristic number are considered in the model of joint shear deformation , and the correlation between joint shear deformation and cycle number are also simulated 贯穿节点梁筋的滑移模型考虑了贯穿梁筋的相对直径、混凝土强度及轴压比三个因素，并根据试验结果模拟出滑移量随循环次数的变化规律。节点剪切变形模型考虑了混凝土强度、轴压比、剪压比、节点配箍特征值四个因素，并根据试验结果模拟出节点剪切变形量随循环次数的变化规律。
The main contents are as follows : 1 . the experimental data of autogenous volume deformation of concrete with mgo admixture ( mgo concrete for short ) was comprehensively analyzed , then the exponent model and hyperbola model for autogenous volume deformation of mgo concrete were discussed . at the same time , the autogenous volume deformation model of concrete that has 4 . 5 % mgo content was developed 主要研究内容如下： 1 、对外掺mgo混凝土的自生体积变形试验资料进行综合分析研究，在此基础上，探讨了外掺mgo混凝土自生体积变形的指数模型和双曲线模型，并建立了外掺mgo掺量为4 . 5混凝土自生体积变形模型。
In these two experiments , it includes test scheme . outer precision estimation , inner precision estimation . the data collecting is in accordance with ii order leveling measuring and ii order distance network . these two experiments test the reasonable of datum consistency and the method of mean gap building deformation model 该实验验证了对于基准问题的分析和平均间隙法对位移量的探测情况；在水平位移变形监测网的实验中，数据采集按二等测边网测量精度进行，证实了用平均间隙法判断稳定点，进而确定变形模型的合理性。
Considering the joint core deformation effect , the existing cantilever program was assembled with the slip deformation model of longitudinal reinforcement passing through joint and the shear deformation model of joint to obtain the program for simulation of hysteretic characteristics of beam - column subassemblage 本文考虑节点区变形的影响，为原有程序组装了节点区贯穿节点梁筋的滑移变形模型和节点剪切变形模型，得到梁柱组合体在低周交变加载下的梁端荷载-位移滞回规律模拟程序。
Through approximating the deformation space by using curve and surface offset techniques , we obtain a unified space deformation model . thus the sampling problem is solve naturally at the same time , we provide a unified kernel for the implementation of different space deformation models 通过对曲线、曲面控制变形中的变形空间进行等距逼近的方法，我们实现了一个一维、二维、三维控制手段的统一空间变形模型，这种统一不但自然地解决了曲线、曲面控制变形中的采样问题，同时为实现各种空间变形提供了一个统一的数学模型。
