From 2b5c77131ee75b1e6d18a6a94f664abbc6059f5f Mon Sep 17 00:00:00 2001 From: duxin Date: Tue, 7 Jul 2026 10:32:02 +0800 Subject: [PATCH] =?UTF-8?q?fix:=20BIP=20=E6=A0=BC=E5=BC=8F=E5=85=BC?= =?UTF-8?q?=E5=AE=B9=20=E2=80=94=20=E9=B2=81=E6=A3=92=20HDR=20=E6=9F=A5?= =?UTF-8?q?=E6=89=BE=20+=20find=5Fband=5Fnumber=20GDAL=20=E5=9B=9E?= =?UTF-8?q?=E9=80=80?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 问题: 用户导入 3ref.bip 文件，step2 报 FileNotFoundError: 3ref.hdr 不存在。 根本原因: get_hdr_file_path() 只用 os.path.splitext()[0]+.hdr， 对于 3ref.bip 只查找 3ref.hdr，不兼容 3ref.bip.hdr 等其他命名规范。 修复内容: **util.py (核心):** - get_hdr_file_path(): 改为多候选路径查找（按优先级）: 3ref.hdr → 3ref.bip.hdr → 3ref.HDR → 3ref.bip.HDR - find_band_number(): 三级回退 — 1) ENVI .hdr 文件 → spectral 解析 2) GDAL 元数据域 (ENVI/wavelength, WAVELENGTH_1..N) 3) 线性估算 (假设 400-1000nm 或 400-2500nm) - 新增 _read_wavelengths_from_gdal() 辅助函数 **同模式修复 (4 处):** - get_spectral.py: get_hdr_file_path() 多候选 - get_spectral-test.py: 同上 - waterindex_inversion/__init__.py: 两处 hdr 构造均改为多候选 - sampling.py: 波长读取的 hdr 查找改为多候选 --- .../waterindex_inversion/__init__.py | 50 ++++--- src/core/glint_removal/get_spectral-test.py | 19 +-- src/core/glint_removal/get_spectral.py | 17 ++- src/utils/sampling.py | 22 ++- src/utils/util.py | 136 +++++++++++++++++- 5 files changed, 204 insertions(+), 40 deletions(-) diff --git a/src/core/algorithms/waterindex_inversion/__init__.py b/src/core/algorithms/waterindex_inversion/__init__.py index e88cd4d..1a14a71 100644 --- a/src/core/algorithms/waterindex_inversion/__init__.py +++ b/src/core/algorithms/waterindex_inversion/__init__.py @@ -205,13 +205,18 @@ class WaterIndexProcessor: except Exception: pass - # 2. HDR 补充波长信息 + # 2. HDR 补充波长信息（多命名规范兼容 .bsq/.bil/.bip/.dat） if hdr_path is None: - hdr_path = os.path.splitext(bsq_path)[0] + '.hdr' - if not os.path.isfile(hdr_path): - hdr_path_alt = os.path.splitext(bsq_path)[0] + '.HDR' - if os.path.isfile(hdr_path_alt): - hdr_path = hdr_path_alt + hdr_candidates = [ + os.path.splitext(bsq_path)[0] + '.hdr', # 3ref.hdr + bsq_path + '.hdr', # 3ref.bip.hdr + os.path.splitext(bsq_path)[0] + '.HDR', # 3ref.HDR + bsq_path + '.HDR', # 3ref.bip.HDR + ] + for candidate in hdr_candidates: + if os.path.isfile(candidate): + hdr_path = candidate + break if os.path.isfile(hdr_path): wl = self._parse_wavelengths_from_hdr(hdr_path) @@ -365,13 +370,18 @@ class WaterIndexProcessor: dict {公式名: 输出 GeoTIFF 路径} """ - # ── 自动构造 HDR 路径 ──────────────────────────────────────────── + # ── 自动构造 HDR 路径（多命名规范兼容） ──────────────────────── if hdr_path is None: - hdr_path = os.path.splitext(bsq_path)[0] + '.hdr' - if not os.path.isfile(hdr_path): - hdr_path_alt = os.path.splitext(bsq_path)[0] + '.HDR' - if os.path.isfile(hdr_path_alt): - hdr_path = hdr_path_alt + hdr_candidates = [ + os.path.splitext(bsq_path)[0] + '.hdr', + bsq_path + '.hdr', + os.path.splitext(bsq_path)[0] + '.HDR', + bsq_path + '.HDR', + ] + for candidate in hdr_candidates: + if os.path.isfile(candidate): + hdr_path = candidate + break # ── 自动构造输出目录 ──────────────────────────────────────────── if output_dir is None: @@ -605,11 +615,17 @@ class WaterIndexProcessor: notify("开始水色指数反演", 0) bsq_path = deglint_img_path - hdr_path = os.path.splitext(bsq_path)[0] + '.hdr' - if not os.path.isfile(hdr_path): - hdr_path_alt = os.path.splitext(bsq_path)[0] + '.HDR' - if os.path.isfile(hdr_path_alt): - hdr_path = hdr_path_alt + hdr_candidates = [ + os.path.splitext(bsq_path)[0] + '.hdr', + bsq_path + '.hdr', + os.path.splitext(bsq_path)[0] + '.HDR', + bsq_path + '.HDR', + ] + hdr_path = None + for candidate in hdr_candidates: + if os.path.isfile(candidate): + hdr_path = candidate + break output_dir = os.path.join(work_dir, "10_WaterIndex_Images") diff --git a/src/core/glint_removal/get_spectral-test.py b/src/core/glint_removal/get_spectral-test.py index b90a9ca..4b1f481 100644 --- a/src/core/glint_removal/get_spectral-test.py +++ b/src/core/glint_removal/get_spectral-test.py @@ -330,15 +330,18 @@ def load_mask_file(mask_path): def get_hdr_file_path(file_path): """ - 获取HDR文件路径 - - Args: - file_path: 影像文件路径 - - Returns: - HDR文件路径 + 获取 ENVI 头文件路径（鲁棒版：多命名规范兼容） """ - return os.path.splitext(file_path)[0] + ".hdr" + candidates = [ + os.path.splitext(file_path)[0] + ".hdr", + file_path + ".hdr", + os.path.splitext(file_path)[0] + ".HDR", + file_path + ".HDR", + ] + for path in candidates: + if os.path.isfile(path): + return path + return candidates[0] def calculate_utm_zone(longitude): diff --git a/src/core/glint_removal/get_spectral.py b/src/core/glint_removal/get_spectral.py index 2edf0f2..687f56c 100644 --- a/src/core/glint_removal/get_spectral.py +++ b/src/core/glint_removal/get_spectral.py @@ -212,15 +212,26 @@ def load_mask_file(mask_path): def get_hdr_file_path(file_path): """ - 获取HDR文件路径 + 获取 ENVI 头文件路径（鲁棒版：多命名规范兼容） + + 支持 .bsq / .bil / .bip / .dat 等格式的多种 .hdr 命名规范。 Args: file_path: 影像文件路径 Returns: - HDR文件路径 + 存在的 .hdr 文件路径；若都不存在，返回标准命名路径 """ - return os.path.splitext(file_path)[0] + ".hdr" + candidates = [ + os.path.splitext(file_path)[0] + ".hdr", # 3ref.hdr + file_path + ".hdr", # 3ref.bip.hdr + os.path.splitext(file_path)[0] + ".HDR", # 3ref.HDR + file_path + ".HDR", # 3ref.bip.HDR + ] + for path in candidates: + if os.path.isfile(path): + return path + return candidates[0] def load_wavelength_columns(imgpath, num_bands): diff --git a/src/utils/sampling.py b/src/utils/sampling.py index 93d00da..418e1e2 100644 --- a/src/utils/sampling.py +++ b/src/utils/sampling.py @@ -38,13 +38,23 @@ def get_wavelengths_from_bil_header(bil_file): list - 波长列表，如果无法获取则返回None """ try: - # 获取头文件路径 - header_file = os.path.splitext(bil_file)[0] + ".hdr" - - if not os.path.exists(header_file): - print(f"警告: 找不到头文件 {header_file}") + # 获取头文件路径（多命名规范兼容 .bsq/.bil/.bip/.dat） + hdr_candidates = [ + os.path.splitext(bil_file)[0] + ".hdr", # 3ref.hdr + bil_file + ".hdr", # 3ref.bip.hdr + os.path.splitext(bil_file)[0] + ".HDR", # 3ref.HDR + bil_file + ".HDR", # 3ref.bip.HDR + ] + header_file = None + for candidate in hdr_candidates: + if os.path.exists(candidate): + header_file = candidate + break + + if header_file is None: + print(f"警告: 找不到头文件，已尝试: {hdr_candidates}") return None - + # 使用spectral库读取头文件 import spectral.io.envi as envi header = envi.read_envi_header(header_file) diff --git a/src/utils/util.py b/src/utils/util.py index 8b92205..ce4a989 100644 --- a/src/utils/util.py +++ b/src/utils/util.py @@ -33,18 +33,142 @@ def timeit(f): # decorator def get_hdr_file_path(file_path): - return os.path.splitext(file_path)[0] + ".hdr" + """获取 ENVI 头文件路径（鲁棒版：多命名规范兼容） + + 支持以下命名模式（按优先级检测）： + 1. {filename}.hdr ← 标准 ENVI 规范，如 3ref.hdr + 2. {filename_with_ext}.hdr ← 如 3ref.bip.hdr + 3. {basename}.HDR ← 大写变体 + 4. {filename}.HDR + + Args: + file_path: 影像文件路径（.bsq / .bil / .bip / .dat 等） + + Returns: + 存在的 .hdr 文件路径；若都不存在，返回标准命名路径（交给调用方报错） + """ + # 候选路径列表（按优先级） + candidates = [ + os.path.splitext(file_path)[0] + ".hdr", # 3ref.hdr + file_path + ".hdr", # 3ref.bip.hdr + os.path.splitext(file_path)[0] + ".HDR", # 3ref.HDR + file_path + ".HDR", # 3ref.bip.HDR + ] + + for path in candidates: + if os.path.isfile(path): + return path + + # 都不存在：返回标准命名（让调用方报明确错误） + return candidates[0] def find_band_number(wav1, imgpath): - in_hdr_dict = spectral.envi.read_envi_header(get_hdr_file_path(imgpath)) + """根据目标波长查找最接近的波段序号（0-based） - wavelengths = np.array(in_hdr_dict['wavelength']).astype('float64') + 优先级： + 1) ENVI .hdr 文件 → spectral 库解析 + 2) GDAL 元数据 → 直接从数据集读取波长域 + 3) 回退 → 基于波段计数线性估算 - differences = np.abs(wavelengths - wav1) - min_position = np.argmin(differences) + Args: + wav1: 目标波长 (nm) + imgpath: 影像文件路径 - return int(min_position) + Returns: + 最接近的波段序号（0-based int） + """ + # ── 路径 1：ENVI .hdr 文件 ── + hdr_path = get_hdr_file_path(imgpath) + if os.path.isfile(hdr_path): + try: + in_hdr_dict = spectral.envi.read_envi_header(hdr_path) + wavelengths = np.array(in_hdr_dict['wavelength']).astype('float64') + differences = np.abs(wavelengths - wav1) + min_position = int(np.argmin(differences)) + print(f"[find_band] HDR 解析成功: target={wav1}nm → band {min_position} " + f"(wl={wavelengths[min_position]:.2f}nm), 来自 {hdr_path}") + return min_position + except Exception as e: + print(f"[find_band] HDR 解析失败 ({hdr_path}): {e}，尝试 GDAL 回退") + + # ── 路径 2：GDAL 元数据 ── + try: + ds = gdal.Open(imgpath, gdal.GA_ReadOnly) + if ds is not None: + n_bands = ds.RasterCount + # 尝试从 GDAL 元数据域读取波长 + wavelengths = _read_wavelengths_from_gdal(ds, n_bands) + ds = None + if wavelengths is not None and len(wavelengths) > 0: + differences = np.abs(np.array(wavelengths, dtype='float64') - wav1) + min_position = int(np.argmin(differences)) + print(f"[find_band] GDAL 回退成功: target={wav1}nm → band {min_position} " + f"(wl={wavelengths[min_position]:.2f}nm), 共 {n_bands} 波段") + return min_position + except Exception as e: + print(f"[find_band] GDAL 元数据读取失败: {e}") + + # ── 路径 3：线性估算（最后回退） ── + try: + ds = gdal.Open(imgpath, gdal.GA_ReadOnly) + n_bands = ds.RasterCount + ds = None + # 假设波长范围 400-1000nm 或 400-2500nm 线性分布 + # 这是非常粗略的估算，但比崩溃好 + estimated_wl_min, estimated_wl_max = (400.0, 1000.0) if n_bands <= 300 else (400.0, 2500.0) + band_idx = int(round((wav1 - estimated_wl_min) / (estimated_wl_max - estimated_wl_min) * (n_bands - 1))) + band_idx = max(0, min(n_bands - 1, band_idx)) + print(f"[find_band] ⚠ 线性估算回退: target={wav1}nm → band {band_idx} " + f"(假设范围 {estimated_wl_min}-{estimated_wl_max}nm, {n_bands} 波段)") + return band_idx + except Exception as e: + raise RuntimeError( + f"无法确定波段号 (target={wav1}nm, img={imgpath}): " + f"HDR 不存在、GDAL 无波长元数据、且无法读取波段数。" + ) from e + + +def _read_wavelengths_from_gdal(dataset, n_bands: int): + """从 GDAL Dataset 的元数据域中提取波长列表 + + 支持的来源（按优先级）： + - ENVI 域: ENVI/wavelength + - 默认域: WAVELENGTH_1, WAVELENGTH_2, ... + + Returns: + wavelengths list 或 None + """ + wavelengths = [] + + # 尝试 ENVI metadata domain + try: + envi_md = dataset.GetMetadata('ENVI') + if 'wavelength' in envi_md: + raw = envi_md['wavelength'] + # 可能是 { ... } 包裹的逗号分隔列表 + raw = raw.strip('{}').strip() + wavelengths = [float(x.strip()) for x in raw.split(',') if x.strip()] + if wavelengths: + return wavelengths + except Exception: + pass + + # 尝试从默认域按 WAVELENGTH_1, WAVELENGTH_2, ... 读取 + try: + md = dataset.GetMetadata() + for i in range(1, n_bands + 1): + key = f'WAVELENGTH_{i}' + if key in md: + wavelengths.append(float(md[key])) + else: + break + if wavelengths: + return wavelengths + except Exception: + pass + + return None @timeit