float simple_liner_regression(np::ndarray a, np::ndarray b, np::ndarray c) {
int nd1 = a.get_nd();
int nd2 = b.get_nd();
if (nd1 != 1 || nd2 != 1)
throw std::runtime_error("a and b must be 1-dimensional");
if ( (a.get_dtype() != np::dtype::get_builtin<double>()) ||
(b.get_dtype() != np::dtype::get_builtin<double>()) )
throw std::runtime_error("a and b must be float64 array");
size_t N = a.shape(0);
if ( N != b.shape(0) )
throw std::runtime_error(" a and b must be same size");
double *p = reinterpret_cast<double *>(a.get_data());
std::vector<float> x;
for(int i=0;i<N;i++) x.push_back(*p++);
double *q = reinterpret_cast<double *>(b.get_data());
std::vector<float> y;
for(int i=0;i<N;i++) y.push_back(*q++);
// 回帰系数の計算
float a1 = calc_covariance(x,y) / calc_variance(x);
float a0 = calc_mean(y) - a1 * calc_mean(x);
double *r = reinterpret_cast<double *>(c.get_data());
*r = a0; r++;
*r = a1;
}
float covariance(np::ndarray a, np::ndarray b) {
int nd1 = a.get_nd();
int nd2 = b.get_nd();
if (nd1 != 1 || nd2 != 1)
throw std::runtime_error("a and b must be 1-dimensional");
if ( (a.get_dtype() != np::dtype::get_builtin<double>()) ||
(b.get_dtype() != np::dtype::get_builtin<double>()) )
throw std::runtime_error("a and b must be float64 array");
size_t N = a.shape(0);
if ( N != b.shape(0) )
throw std::runtime_error(" a and b must be same size");
double *p = reinterpret_cast<double *>(a.get_data());
std::vector<float> x;
for(int i=0;i<N;i++) x.push_back(*p++);
double *q = reinterpret_cast<double *>(b.get_data());
std::vector<float> y;
for(int i=0;i<N;i++) y.push_back(*q++);
return calc_covariance(x,y);
}
int main(int argc, char *argv[])
{
int i, j; /* Loop control variables */
int bands; /* Number of image bands */
double *mu; /* Mean vector for image bands */
double **covar; /* Covariance Matrix */
double *eigval;
double **eigmat;
int *inp_fd;
int scale, scale_max, scale_min;
struct GModule *module;
struct Option *opt_in, *opt_out, *opt_scale;
/* initialize GIS engine */
G_gisinit(argv[0]);
module = G_define_module();
G_add_keyword(_("imagery"));
G_add_keyword(_("image transformation"));
G_add_keyword(_("PCA"));
module->description = _("Principal components analysis (PCA) "
"for image processing.");
/* Define options */
opt_in = G_define_standard_option(G_OPT_R_INPUTS);
opt_in->description = _("Name of two or more input raster maps");
opt_out = G_define_option();
opt_out->label = _("Base name for output raster maps");
opt_out->description =
_("A numerical suffix will be added for each component map");
opt_out->key = "output_prefix";
opt_out->type = TYPE_STRING;
opt_out->key_desc = "string";
opt_out->required = YES;
opt_scale = G_define_option();
opt_scale->key = "rescale";
opt_scale->type = TYPE_INTEGER;
opt_scale->key_desc = "min,max";
opt_scale->required = NO;
opt_scale->answer = "0,255";
opt_scale->label =
_("Rescaling range for output maps");
opt_scale->description =
_("For no rescaling use 0,0");
opt_scale->guisection = _("Rescale");
if (G_parser(argc, argv))
exit(EXIT_FAILURE);
/* determine number of bands passed in */
for (bands = 0; opt_in->answers[bands] != NULL; bands++) ;
if (bands < 2)
G_fatal_error(_("Sorry, at least 2 input bands must be provided"));
/* default values */
scale = 1;
scale_min = 0;
scale_max = 255;
/* get scale parameters */
set_output_scale(opt_scale, &scale, &scale_min, &scale_max);
/* allocate memory */
covar = G_alloc_matrix(bands, bands);
mu = G_alloc_vector(bands);
inp_fd = G_alloc_ivector(bands);
eigmat = G_alloc_matrix(bands, bands);
eigval = G_alloc_vector(bands);
/* open and check input/output files */
for (i = 0; i < bands; i++) {
char tmpbuf[128];
sprintf(tmpbuf, "%s.%d", opt_out->answer, i + 1);
G_check_input_output_name(opt_in->answers[i], tmpbuf, GR_FATAL_EXIT);
inp_fd[i] = Rast_open_old(opt_in->answers[i], "");
}
G_verbose_message(_("Calculating covariance matrix..."));
calc_mu(inp_fd, mu, bands);
calc_covariance(inp_fd, covar, mu, bands);
for (i = 0; i < bands; i++) {
for (j = 0; j < bands; j++) {
covar[i][j] =
covar[i][j] /
((double)((Rast_window_rows() * Rast_window_cols()) - 1));
G_debug(3, "covar[%d][%d] = %f", i, j, covar[i][j]);
}
}
G_math_d_copy(covar[0], eigmat[0], bands*bands);
G_debug(1, "Calculating eigenvalues and eigenvectors...");
G_math_eigen(eigmat, eigval, bands);
#ifdef PCA_DEBUG
/* dump eigen matrix and eigen values */
dump_eigen(bands, eigmat, eigval);
#endif
G_debug(1, "Ordering eigenvalues in descending order...");
G_math_egvorder(eigval, eigmat, bands);
G_debug(1, "Transposing eigen matrix...");
G_math_d_A_T(eigmat, bands);
/* write output images */
write_pca(eigmat, inp_fd, opt_out->answer, bands, scale, scale_min,
scale_max);
/* write colors and history to output */
for (i = 0; i < bands; i++) {
char outname[80];
sprintf(outname, "%s.%d", opt_out->answer, i + 1);
/* write colors and history to file */
write_support(bands, outname, eigmat, eigval);
/* close output file */
Rast_unopen(inp_fd[i]);
}
/* free memory */
G_free_matrix(covar);
G_free_vector(mu);
G_free_ivector(inp_fd);
G_free_matrix(eigmat);
G_free_vector(eigval);
exit(EXIT_SUCCESS);
}
