Mercurial > hgrepos > Python2 > PyMuPDF
view mupdf-source/thirdparty/zxing-cpp/core/src/GridSampler.cpp @ 2:b50eed0cc0ef upstream
ADD: MuPDF v1.26.7: the MuPDF source as downloaded by a default build of PyMuPDF 1.26.4.
The directory name has changed: no version number in the expanded directory now.
| author | Franz Glasner <fzglas.hg@dom66.de> |
|---|---|
| date | Mon, 15 Sep 2025 11:43:07 +0200 |
| parents | |
| children |
line wrap: on
line source
/* * Copyright 2016 Nu-book Inc. * Copyright 2016 ZXing authors * Copyright 2020 Axel Waggershauser */ // SPDX-License-Identifier: Apache-2.0 #include "GridSampler.h" #ifdef PRINT_DEBUG #include "LogMatrix.h" #include "BitMatrixIO.h" #endif namespace ZXing { #ifdef PRINT_DEBUG LogMatrix log; #endif DetectorResult SampleGrid(const BitMatrix& image, int width, int height, const PerspectiveTransform& mod2Pix) { return SampleGrid(image, width, height, {ROI{0, width, 0, height, mod2Pix}}); } DetectorResult SampleGrid(const BitMatrix& image, int width, int height, const ROIs& rois) { #ifdef PRINT_DEBUG LogMatrix log; static int i = 0; LogMatrixWriter lmw(log, image, 5, "grid" + std::to_string(i++) + ".pnm"); #endif if (width <= 0 || height <= 0) return {}; for (auto&& [x0, x1, y0, y1, mod2Pix] : rois) { // Precheck the corners of every roi to bail out early if the grid is "obviously" not completely inside the image auto isInside = [&mod2Pix = mod2Pix, &image](int x, int y) { return image.isIn(mod2Pix(centered(PointI(x, y)))); }; if (!mod2Pix.isValid() || !isInside(x0, y0) || !isInside(x1 - 1, y0) || !isInside(x1 - 1, y1 - 1) || !isInside(x0, y1 - 1)) return {}; } BitMatrix res(width, height); for (auto&& [x0, x1, y0, y1, mod2Pix] : rois) { for (int y = y0; y < y1; ++y) for (int x = x0; x < x1; ++x) { auto p = mod2Pix(centered(PointI{x, y})); // Due to a "numerical instability" in the PerspectiveTransform generation/application it has been observed // that even though all boundary grid points get projected inside the image, it can still happen that an // inner grid points is not. See #563. A true perspective transformation cannot have this property. // The following check takes 100% care of the issue and turned out to be less of a performance impact than feared. // TODO: Check some mathematical/numercial property of mod2Pix to determine if it is a perspective transforation. if (!image.isIn(p)) return {}; #ifdef PRINT_DEBUG log(p, 3); #endif #if 0 int sum = 0; for (int dy = -1; dy <= 1; ++dy) for (int dx = -1; dx <= 1; ++dx) sum += image.get(p + PointF(dx, dy)); if (sum >= 5) #else if (image.get(p)) #endif res.set(x, y); } } #ifdef PRINT_DEBUG printf("width: %d, height: %d\n", width, height); // printf("%s", ToString(res).c_str()); #endif auto projectCorner = [&](PointI p) { for (auto&& [x0, x1, y0, y1, mod2Pix] : rois) if (x0 <= p.x && p.x <= x1 && y0 <= p.y && p.y <= y1) return PointI(mod2Pix(PointF(p)) + PointF(0.5, 0.5)); return PointI(); }; return {std::move(res), {projectCorner({0, 0}), projectCorner({width, 0}), projectCorner({width, height}), projectCorner({0, height})}}; } } // ZXing
