Ceres Solver 中文文档
  • 😜Ceres Solver 中文文档
  • Why Ceres ?
  • Installation
  • Tutorial
    • Non-linear Least Squares
      • Introduction
      • Hello World
      • Derivatives
        • Numeric Derivatives
        • Analytic Derivatives
        • More About Derivatives
      • Powell’s Function
      • Curve Fitting
      • Robust Curve Fitting
      • Bundle Adjustment
      • Other Examples
    • General Unconstrained Minimization
      • General Unconstrained Minimization
  • On Derivatives
    • Spivak Notation
    • Analytic Derivatives
    • Numeric Derivatives
    • Automatic Derivatives
    • Interfacing with Automatic Differentiation
    • Using Inverse & Implicit Function Theorems
  • Modeling Non-linear Least Squares
    • Introduction
    • Main Class Interface
      • CostFunction
      • SizeCostFunction
      • AutoDiffCostFunction
      • DynamicAutoDiffCostFunction
      • NumericDiffCostFunction
      • DynamicNumericDifferCostFunction
      • CostFunctionToFunctor
      • DynamicCostFunctionToFunctor
      • ConditionedCostFunction
      • GradientChecker
      • NormalPrior
      • LossFunction
      • Manifold
      • AutoDIffManifold
      • Problem
      • EvaluatationCallback
      • Rotation
      • Cubic Interpolation
        • CubicInterpolator
        • BiCubicInterpolator
  • Solveing Non-linear Least Squares
    • Introduction
    • Trust Region Methodd
    • Line Search Methods
    • Linear Solvers
    • Mixed Precision Solves
    • Preconditioners
    • Ordering
    • Main Class Interfaces
      • Solver::Options
      • ParameterBlockOrdering
      • IterationSummary
      • IterationCallback
      • CRSMatrix
      • Solver::Summary
  • Covariance Estimation
    • Introduction
    • Gauge Invariance
    • Covariance
    • Rank of the Jacobian
      • Options
      • Covariance
      • GetCovarianceBlock
      • GetCovarianceBlockInTangentSpace
    • Example Usage
Powered by GitBook
On this page
  1. Tutorial
  2. Non-linear Least Squares

Curve Fitting

到目前为止,我们看到的例子都是没有数据的简单优化问题。最小二乘法和非线性最小二乘法的最初目的是对数据进行曲线拟合。我们使用曲线方程 y=e0.3x+0.1y=e^{0.3x+0.1}y=e0.3x+0.1 来生成样本数据,并且对每个样本都添加一个标准差为 σ=0.2\sigma=0.2σ=0.2 的高斯噪声,我们来拟合如下曲线方程:

y=emx+cy=e^{mx+c}y=emx+c

其中 m,cm,cm,c 使我们要通过样本进行估计的参数。和前面一样,第一步仍然是定义一个模板函数去评估每一个参数块的残差,这里有多少个样本就有多少个残差项,但是每个残差项的 Cost 计算方式和 ParameterBlock 的维度都是一样的。

struct ExponentialResidual {
  ExponentialResidual(double x, double y)
      : x_(x), y_(y) {}

  template <typename T>
  bool operator()(const T* const m, const T* const c, T* residual) const {
    residual[0] = y_ - exp(m[0] * x_ + c[0]);
    return true;
  }

 private:
  // Observations for a sample.
  const double x_;
  const double y_;
};

假定观测数据存放在一个 2n 大小的数组中,我们称为 data。然后进行优化问题的构造,就是为每个观测创建一个 CostFunction。

double m = 0.0;
double c = 0.0;

Problem problem;
for (int i = 0; i < kNumObservations; ++i) {
  CostFunction* cost_function =
       new AutoDiffCostFunction<ExponentialResidual, 1, 1, 1>
           (data[2 * i], data[2 * i + 1]);
  problem.AddResidualBlock(cost_function, nullptr, &m, &c);
}
iter      cost      cost_change  |gradient|   |step|    tr_ratio  tr_radius  ls_iter  iter_time  total_time
   0  1.211734e+02    0.00e+00    3.61e+02   0.00e+00   0.00e+00  1.00e+04       0    5.34e-04    2.56e-03
   1  1.211734e+02   -2.21e+03    0.00e+00   7.52e-01  -1.87e+01  5.00e+03       1    4.29e-05    3.25e-03
   2  1.211734e+02   -2.21e+03    0.00e+00   7.51e-01  -1.86e+01  1.25e+03       1    1.10e-05    3.28e-03
   3  1.211734e+02   -2.19e+03    0.00e+00   7.48e-01  -1.85e+01  1.56e+02       1    1.41e-05    3.31e-03
   4  1.211734e+02   -2.02e+03    0.00e+00   7.22e-01  -1.70e+01  9.77e+00       1    1.00e-05    3.34e-03
   5  1.211734e+02   -7.34e+02    0.00e+00   5.78e-01  -6.32e+00  3.05e-01       1    1.00e-05    3.36e-03
   6  3.306595e+01    8.81e+01    4.10e+02   3.18e-01   1.37e+00  9.16e-01       1    2.79e-05    3.41e-03
   7  6.426770e+00    2.66e+01    1.81e+02   1.29e-01   1.10e+00  2.75e+00       1    2.10e-05    3.45e-03
   8  3.344546e+00    3.08e+00    5.51e+01   3.05e-02   1.03e+00  8.24e+00       1    2.10e-05    3.48e-03
   9  1.987485e+00    1.36e+00    2.33e+01   8.87e-02   9.94e-01  2.47e+01       1    2.10e-05    3.52e-03
  10  1.211585e+00    7.76e-01    8.22e+00   1.05e-01   9.89e-01  7.42e+01       1    2.10e-05    3.56e-03
  11  1.063265e+00    1.48e-01    1.44e+00   6.06e-02   9.97e-01  2.22e+02       1    2.60e-05    3.61e-03
  12  1.056795e+00    6.47e-03    1.18e-01   1.47e-02   1.00e+00  6.67e+02       1    2.10e-05    3.64e-03
  13  1.056751e+00    4.39e-05    3.79e-03   1.28e-03   1.00e+00  2.00e+03       1    2.10e-05    3.68e-03
Ceres Solver Report: Iterations: 13, Initial cost: 1.211734e+02, Final cost: 1.056751e+00, Termination: CONVERGENCE
Initial m: 0 c: 0
Final   m: 0.291861 c: 0.131439

从 m=0,c=0m=0,c=0m=0,c=0 为初值进行优化,目标函数值为 1.211734e+02。最终 Ceres 找到的解为 m=0.291861,c=0.131439m=0.291861,c=0.131439m=0.291861,c=0.131439,目标函数值为 1.056。和真值 m=0.3,c=0.1m=0.3,c=0.1m=0.3,c=0.1 比起来有略微的差别,但是由于噪声的存在,结果其实在可接受的范围内。实际上,如果用m=0.3,c=0.1m=0.3,c=0.1m=0.3,c=0.1 来评估目标函数会发现 Cost 比 1.056 还要大。下图展示了拟合结果。

Footnotes

PreviousPowell’s FunctionNextRobust Curve Fitting

Last updated 1 year ago

编译并运行 可以得到如下结果:

examples/curve_fitting.cc
examples/curve_fitting.cc
Least Square Curve Fitting