The ring fit reported the SCATTER of its measurements (rms, and the beam-centre standard error that follows from it) but nothing about how well the fit pinned each parameter. Those two part company exactly where a calibration is worth doubting: as the rings run out, the tilt and the beam centre stop being separable - both displace a ring's radius as cos(phi) and only the way that amplitude scales with radius tells them apart - so the fit can sit tightly on the few points it has while being free to spend tens of pixels of beam centre on a tilt the data do not support. Take the covariance of the converged problem from Ceres and report it. Measured on a LaB6 distance series, the fitted tilt is 50 sigma at 110 mm and 0.1 sigma at 500 mm, where only two rings reach the detector; at 500 mm the fit quotes its own beam centre to +-180 px and its tilt to +-2.9 deg on a 0.35 deg value, and the correlation between them is 1.000. Nothing acts on this yet - it is printed so the next change can gate on it. Ceres returns the bare (J'J)^-1 of an unweighted problem, so it is scaled by chi2 per degree of freedom; that leaves the sigmas in pixels, mm and radians whatever unit the residual is stated in. Cost is one 5x5 SVD per run, below the noise of the surrounding I/O. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01NfuDvf5ipV3Hi8TiCUKD27
303 lines
11 KiB
C++
303 lines
11 KiB
C++
// SPDX-FileCopyrightText: 2025 Filip Leonarski, Paul Scherrer Institute <filip.leonarski@psi.ch>
|
|
// SPDX-License-Identifier: GPL-3.0-only
|
|
|
|
#include "../../common/JFJochMath.h"
|
|
#include "AssignSpotsToRings.h"
|
|
|
|
#include <vector>
|
|
#include <cmath>
|
|
#include <tuple>
|
|
#include <algorithm>
|
|
#include <iostream>
|
|
|
|
#include "../../common/CrystalLattice.h"
|
|
|
|
FindCircleCenterResult FindCircleCenter(const std::vector<SpotToSave> &v, int64_t width, int64_t height, int64_t max_spots) {
|
|
if ((width <= 0) || (height <= 0) || (max_spots <= 0))
|
|
throw JFJochException(JFJochExceptionCategory::InputParameterInvalid, "Invalid image size");
|
|
|
|
std::vector<int64_t> vote(width * height, 0);
|
|
|
|
// Limit to only first 250 spots, given the algorithm is N^3
|
|
auto task_size = std::min<size_t>(v.size(), max_spots);
|
|
|
|
for (int i = 0; i < task_size; i++) {
|
|
for (int j = i+1; j < task_size; j++) {
|
|
for (int k = j+1; k < task_size; k++) {
|
|
// Calculation via determinants
|
|
float a = v[i].x * (v[j].y - v[k].y) - v[i].y * (v[j].x - v[k].x) + v[j].x * v[k].y - v[k].x * v[j].y;
|
|
if (std::abs(a) < 1e-10)
|
|
continue; // Points are collinear
|
|
|
|
float x1_sq = v[i].x*v[i].x + v[i].y*v[i].y;
|
|
float x2_sq = v[j].x*v[j].x + v[j].y*v[j].y;
|
|
float x3_sq = v[k].x*v[k].x + v[k].y*v[k].y;
|
|
|
|
float bx = x1_sq * (v[j].y - v[k].y) + x2_sq * (v[k].y - v[i].y) + x3_sq * (v[i].y - v[j].y);
|
|
float by = x1_sq * (v[k].x - v[j].x) + x2_sq * (v[i].x - v[k].x) + x3_sq * (v[j].x - v[i].x);
|
|
|
|
int64_t cx = std::lround(bx / (2.0f * a));
|
|
int64_t cy = std::lround(by / (2.0f * a));
|
|
|
|
if ((cx >= 0) && (cx < width) && (cy >= 0) && (cy < height))
|
|
vote[cx + cy * width]++;
|
|
}
|
|
}
|
|
}
|
|
|
|
int64_t total_votes = 0;
|
|
int64_t max_votes = 0;
|
|
int64_t cx = 0;
|
|
int64_t cy = 0;
|
|
for (int64_t y = 0; y < height; y++) {
|
|
for (int64_t x = 0; x < width; x++) {
|
|
total_votes += vote[x + y * width];
|
|
if (vote[x + y * width] > max_votes) {
|
|
max_votes = vote[x + y * width];
|
|
cx = x;
|
|
cy = y;
|
|
}
|
|
}
|
|
}
|
|
|
|
return {
|
|
.total_votes = total_votes,
|
|
.votes_for_beam_center = max_votes,
|
|
.x = static_cast<float>(cx),
|
|
.y = static_cast<float>(cy)
|
|
};
|
|
}
|
|
|
|
// Very simple 1D DBSCAN on radii (works for rings)
|
|
std::vector<std::vector<int>> ClusterSpotsIntoRings(const std::vector<float>& r, float eps, int minPts) {
|
|
size_t n = r.size();
|
|
|
|
std::vector<int> labels(n, -1); // -1 = unvisited, -2 = noise
|
|
int cluster_id = 0;
|
|
|
|
for (int i=0; i<n; i++) {
|
|
if (labels[i] != -1) continue; // already visited
|
|
|
|
// find neighbors within eps in radius
|
|
std::vector<int> neighbors;
|
|
for (int j=0; j<n; j++) {
|
|
if (std::fabs(r[i] - r[j]) <= eps) neighbors.push_back(j);
|
|
}
|
|
|
|
if ((int)neighbors.size() < minPts) {
|
|
labels[i] = -2; // noise
|
|
continue;
|
|
}
|
|
|
|
// start new cluster
|
|
labels[i] = cluster_id;
|
|
std::vector<int> seeds = neighbors;
|
|
for (size_t k=0; k<seeds.size(); k++) {
|
|
int j = seeds[k];
|
|
if (labels[j] == -2) labels[j] = cluster_id;
|
|
if (labels[j] != -1) continue;
|
|
labels[j] = cluster_id;
|
|
|
|
// expand cluster
|
|
std::vector<int> nbrs2;
|
|
for (int m=0; m<n; m++) {
|
|
if (std::fabs(r[j] - r[m]) <= eps) nbrs2.push_back(m);
|
|
}
|
|
if ((int)nbrs2.size() >= minPts) {
|
|
seeds.insert(seeds.end(), nbrs2.begin(), nbrs2.end());
|
|
}
|
|
}
|
|
cluster_id++;
|
|
}
|
|
|
|
// Collect results
|
|
std::vector<std::vector<int>> clusters(cluster_id);
|
|
for (int i=0; i<n; i++) {
|
|
if (labels[i] >= 0)
|
|
clusters[labels[i]].push_back(i);
|
|
}
|
|
return clusters;
|
|
}
|
|
|
|
float median(std::vector<float> v) {
|
|
if (v.empty()) return std::numeric_limits<float>::quiet_NaN();
|
|
size_t n = v.size();
|
|
std::nth_element(v.begin(), v.begin()+n/2, v.end());
|
|
float m = v[n/2];
|
|
if (n % 2 == 0) {
|
|
auto it = std::max_element(v.begin(), v.begin()+n/2);
|
|
m = 0.5f*(m + *it);
|
|
}
|
|
return m;
|
|
}
|
|
|
|
std::vector<RingClusters> AnalyzeClusters(const std::vector<float>& r, const std::vector<std::vector<int>> &clusters) {
|
|
std::vector<RingClusters> ret;
|
|
|
|
for (const auto & cluster : clusters) {
|
|
std::vector<float> cluster_r;
|
|
for (const auto &idx : cluster)
|
|
cluster_r.push_back(r[idx]);
|
|
|
|
if (cluster_r.size() < 2) continue;
|
|
float m = median(cluster_r);
|
|
ret.push_back({cluster, m, -1});
|
|
}
|
|
|
|
// sort by observed radius
|
|
if (!ret.empty())
|
|
std::sort(ret.begin(), ret.end(), [](const RingClusters& a, const RingClusters& b)
|
|
{ return a.R_obs < b.R_obs; });
|
|
return ret;
|
|
}
|
|
|
|
namespace {
|
|
bool reflection_present(ReflectionCondition condition, int h, int k, int l) {
|
|
const bool all_odd = (h % 2 != 0) && (k % 2 != 0) && (l % 2 != 0);
|
|
const bool all_even = (h % 2 == 0) && (k % 2 == 0) && (l % 2 == 0);
|
|
switch (condition) {
|
|
case ReflectionCondition::FaceCentred:
|
|
return all_odd || all_even;
|
|
case ReflectionCondition::Diamond:
|
|
return all_odd || (all_even && ((h + k + l) % 4 == 0));
|
|
case ReflectionCondition::All:
|
|
default:
|
|
return true;
|
|
}
|
|
}
|
|
}
|
|
|
|
std::vector<float> CalculateXtalRings(const UnitCell &cell, ReflectionCondition condition, int hkl_max) {
|
|
CrystalLattice latt(cell);
|
|
|
|
Coord Astar = latt.Astar();
|
|
Coord Bstar = latt.Bstar();
|
|
Coord Cstar = latt.Cstar();
|
|
|
|
std::vector<float> u;
|
|
// Both signs of h and k: only for a diagonal metric does |h a* + k b*| equal |h a* - k b*|, so on a
|
|
// triclinic cell (silver behenate) the positive octant alone misses more rings than it finds. l stays
|
|
// non-negative because hkl and -h-k-l are the same ring.
|
|
for (int h = -hkl_max; h <= hkl_max; h++) {
|
|
for (int k = -hkl_max; k <= hkl_max; k++) {
|
|
for (int l = 0; l <= hkl_max; l++) {
|
|
if (h == 0 && k == 0 && l == 0) continue;
|
|
if (!reflection_present(condition, h, k, l)) continue;
|
|
auto p = Astar * h + Bstar * k + Cstar * l;
|
|
float Q = 2.0f * PI * p.Length();
|
|
u.push_back(Q);
|
|
}
|
|
}
|
|
}
|
|
std::sort(u.begin(), u.end());
|
|
|
|
// Deduplicate (since e.g. (100), (010), (001) all give sqrt(1))
|
|
u.erase(std::unique(u.begin(), u.end(),
|
|
[](float a, float b){ return std::fabs(a-b) < 1e-6; }),
|
|
u.end());
|
|
|
|
return u;
|
|
}
|
|
|
|
|
|
std::vector<float> CalculateCubicXtalRings(float a, int hkl_max) {
|
|
return CalculateXtalRings(UnitCell(a,a,a,90,90,90), ReflectionCondition::All, hkl_max);
|
|
}
|
|
|
|
float GuessDetectorDistance(const DiffractionGeometry& geom, float ring_radius_pxl, float d_A) {
|
|
float sin_theta = geom.GetWavelength_A() / (2 * d_A);
|
|
if (sin_theta < 0 || sin_theta > 1)
|
|
throw JFJochException(JFJochExceptionCategory::InputParameterInvalid, "Geometry makes no sense");
|
|
float theta = asinf(sin_theta);
|
|
float radius_mm = ring_radius_pxl * geom.GetPixelSize_mm();
|
|
float det_dist_mm = radius_mm / tanf(2.0f * theta);
|
|
return det_dist_mm;
|
|
}
|
|
|
|
std::vector<RingClusters> GuessInitialGeometry(DiffractionGeometry &geom, const std::vector<SpotToSave> &v, float largest_ring_d_A) {
|
|
// Reset rotations. The model assumes these are very small in any case!
|
|
geom.PoniRot1_rad(0.0).PoniRot2_rad(0.0).PoniRot3_rad(0.0);
|
|
|
|
auto center = FindCircleCenter(v);
|
|
if (center.votes_for_beam_center < 20)
|
|
throw JFJochException(JFJochExceptionCategory::CalibrationError, "Beam center not found");
|
|
|
|
geom.BeamX_pxl(center.x).BeamY_pxl(center.y);
|
|
|
|
std::vector<float> radii(v.size());
|
|
for (int i = 0; i < v.size(); i++)
|
|
radii[i] = std::hypot(v[i].x - center.x, v[i].y - center.y);
|
|
|
|
auto clusters = ClusterSpotsIntoRings(radii);
|
|
if (clusters.empty())
|
|
throw JFJochException(JFJochExceptionCategory::CalibrationError, "Couldn't find spot clusters");
|
|
|
|
auto cluster_annot = AnalyzeClusters(radii, clusters);
|
|
|
|
float det_distance = GuessDetectorDistance(geom, cluster_annot[0].R_obs, largest_ring_d_A);
|
|
geom.DetectorDistance_mm(det_distance);
|
|
return cluster_annot;
|
|
}
|
|
|
|
void GuessGeometry(DiffractionGeometry &geom, const std::vector<SpotToSave> &v, const std::vector<float> &ring_q,
|
|
bool refine_tilt) {
|
|
if (ring_q.empty())
|
|
throw JFJochException(JFJochExceptionCategory::InputParameterInvalid, "No calibrant rings given");
|
|
auto cluster_annot = GuessInitialGeometry(geom, v, 2 * PI / ring_q[0]);
|
|
|
|
std::vector<RingOptimizerInput> optimizer_input;
|
|
|
|
int ring_idx = 0;
|
|
int cluster_idx = 0;
|
|
|
|
while (cluster_idx < cluster_annot.size() && ring_idx < ring_q.size()) {
|
|
|
|
float obs_q = 2 * PI / geom.PxlToRes(cluster_annot[cluster_idx].R_obs);
|
|
if (std::fabs(ring_q[ring_idx] - obs_q) < 0.1) {
|
|
std::cout << "Found ring " << ring_idx
|
|
<< " with q_ring = " << ring_q[ring_idx]
|
|
<< " and q_obs = " << obs_q
|
|
<< " diff = " << std::fabs(ring_q[ring_idx] - obs_q) << std::endl;
|
|
for (const auto &spot: cluster_annot[cluster_idx].spots)
|
|
optimizer_input.push_back({v[spot].x, v[spot].y, ring_q[ring_idx]});
|
|
ring_idx++;
|
|
cluster_idx++;
|
|
} else {
|
|
std::cout << "Cannot match " << ring_idx
|
|
<< " with q_ring = " << ring_q[ring_idx]
|
|
<< " and q_obs = " << obs_q
|
|
<< " diff = " << std::fabs(ring_q[ring_idx] - obs_q) << std::endl;
|
|
if (ring_q[ring_idx] < obs_q)
|
|
ring_idx++;
|
|
else
|
|
cluster_idx++;
|
|
}
|
|
}
|
|
|
|
RingOptimizer optimizer(geom, refine_tilt);
|
|
geom = optimizer.Run(optimizer_input);
|
|
}
|
|
|
|
std::vector<RingOptimizerInput> AssignSpotsToRings(const DiffractionGeometry &geom,
|
|
const std::vector<SpotToSave> &v,
|
|
const std::vector<float> &ring_q) {
|
|
std::vector<RingOptimizerInput> optimizer_input;
|
|
|
|
for (const auto& s: v) {
|
|
float q_obs = 2 * PI / geom.PxlToRes(s.x, s.y);
|
|
for (const auto &q : ring_q) {
|
|
if (std::fabs(q - q_obs) < 0.1) {
|
|
optimizer_input.push_back({s.x, s.y, q});
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
return optimizer_input;
|
|
}
|
|
|
|
void OptimizeGeometry(DiffractionGeometry &geom, const std::vector<SpotToSave> &v, const std::vector<float> &ring_q,
|
|
bool refine_tilt, RingFitUncertainty *unc) {
|
|
RingOptimizer optimizer(geom, refine_tilt);
|
|
geom = optimizer.Run(AssignSpotsToRings(geom, v, ring_q), unc);
|
|
}
|