498 lines
17 KiB
Python
498 lines
17 KiB
Python
import os
|
||
from pathlib import Path
|
||
import numpy as np
|
||
import supervision as sv
|
||
import cv2
|
||
from concurrent.futures import ThreadPoolExecutor, as_completed
|
||
|
||
import os
|
||
import cv2
|
||
from pathlib import Path
|
||
|
||
def generate_video_to_objects(
|
||
obj_dict: dict[dict],
|
||
input_video_path: str,
|
||
output_dir: str,
|
||
) -> None:
|
||
"""
|
||
根据 obj_dict 中的物体信息，从原视频中截取附近帧，生成新视频
|
||
|
||
:param obj_dict: 包含物体信息的字典，每个元素结构为:
|
||
"0": {
|
||
"class_id": 4,
|
||
"start_frame": 551,
|
||
"end_frame": 597
|
||
},
|
||
:param input_video_path: 输入视频文件路径
|
||
:param output_dir: 输出视频目录
|
||
"""
|
||
# 最低秒数
|
||
min_seconds = 2
|
||
# 前后额外帧数
|
||
extra_frames = 5
|
||
|
||
print(f"开始抽取物体视频: {input_video_path}")
|
||
print(f"输出目录: {output_dir}")
|
||
# 确保输出目录存在
|
||
Path(output_dir).mkdir(parents=True, exist_ok=True)
|
||
|
||
# 1. 打开原始视频
|
||
backends = [
|
||
(cv2.CAP_FFMPEG, 'FFmpeg'),
|
||
(cv2.CAP_DSHOW, 'DirectShow'),
|
||
(cv2.CAP_ANY, 'Default')
|
||
]
|
||
|
||
cap = None
|
||
for backend, backend_name in backends:
|
||
try:
|
||
cap = cv2.VideoCapture(input_video_path, backend)
|
||
if cap.isOpened():
|
||
if backend == cv2.CAP_FFMPEG:
|
||
try:
|
||
cap.set(cv2.CAP_PROP_HW_ACCELERATION, cv2.VIDEO_ACCELERATION_ANY)
|
||
hw_accel = cap.get(cv2.CAP_PROP_HW_ACCELERATION)
|
||
print(f"FFmpeg硬件加速: {'已启用' if hw_accel > 0 else '未启用'}")
|
||
except Exception as e:
|
||
print(f"设置硬件加速失败: {e}")
|
||
print(f"使用后端: {backend_name}")
|
||
break
|
||
except Exception as e:
|
||
print(f"尝试{backend_name}后端失败: {e}")
|
||
continue
|
||
|
||
if not cap or not cap.isOpened():
|
||
raise Exception(f"无法打开视频文件: {input_video_path}")
|
||
|
||
# 2. 获取视频参数
|
||
fps = cap.get(cv2.CAP_PROP_FPS)
|
||
width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
|
||
height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
|
||
total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
|
||
min_frames = int(min_seconds * fps)
|
||
|
||
fourcc = cv2.VideoWriter_fourcc(*'avc1')
|
||
|
||
# 3. 预处理所有物体的帧范围 + 初始化写入器
|
||
total_objects = len(obj_dict)
|
||
print(f"总共需要处理 {total_objects} 个物体")
|
||
|
||
writers = {} # 存储所有视频写入器
|
||
obj_ranges = {} # 存储每个物体的起止帧
|
||
|
||
for track_id, track_data in obj_dict.items():
|
||
start_idx = max(0, track_data["start_frame"] - extra_frames)
|
||
end_idx = min(total_frames - 1, track_data["end_frame"] + extra_frames)
|
||
|
||
# 保证最小长度
|
||
if end_idx - start_idx + 1 < min_frames:
|
||
end_idx = start_idx + min_frames - 1
|
||
|
||
output_path = os.path.join(output_dir, f"obj_{int(track_id):03d}.mp4")
|
||
out = cv2.VideoWriter(output_path, fourcc, fps, (width, height))
|
||
|
||
if not out.isOpened():
|
||
raise Exception(f"无法初始化视频写入器: {output_path}")
|
||
|
||
writers[track_id] = out
|
||
obj_ranges[track_id] = (start_idx, end_idx)
|
||
print(f"物体 {track_id}: 帧 {start_idx} ~ {end_idx} -> {output_path}")
|
||
|
||
# 4. 单轮遍历视频，一次性写入所有需要的帧（最高效）
|
||
frame_idx = 0
|
||
print_interval = 50
|
||
|
||
while cap.isOpened():
|
||
# 读取一帧
|
||
ret, frame = cap.read()
|
||
if not ret:
|
||
break
|
||
|
||
# 遍历所有物体，判断当前帧是否需要写入
|
||
for track_id, (start, end) in obj_ranges.items():
|
||
if start <= frame_idx <= end:
|
||
writers[track_id].write(frame)
|
||
|
||
# 进度打印
|
||
if frame_idx % print_interval == 0:
|
||
progress = (frame_idx / total_frames) * 100
|
||
print(f"处理帧: {frame_idx}/{total_frames} ({progress:.1f}%)")
|
||
|
||
frame_idx += 1
|
||
|
||
# 5. 释放所有写入器
|
||
for track_id, out in writers.items():
|
||
out.release()
|
||
start, end = obj_ranges[track_id]
|
||
print(f"物体 {track_id} 完成 | 总帧数: {end - start + 1}")
|
||
|
||
# 6. 释放资源
|
||
cap.release()
|
||
cv2.destroyAllWindows()
|
||
print(f"\n✅ 所有物体视频生成完成！目录: {output_dir}")
|
||
|
||
def process_track_id(
|
||
track_id: int,
|
||
frame_list: list[tuple[int, list[int]]],
|
||
input_video_path: str,
|
||
output_video_root: str,
|
||
frame_width: int,
|
||
frame_height: int,
|
||
target_fps: int,
|
||
frame_interval: int
|
||
) -> str:
|
||
"""
|
||
处理单个track_id，生成对应的视频
|
||
"""
|
||
# 计算需要生成的总帧数（确保覆盖所有物体帧且不少于两秒）
|
||
min_frames_for_2s = 25 * 2 # 2秒 @ 25fps
|
||
object_based_frames = len(frame_list) * frame_interval
|
||
max_output_frame = max(object_based_frames, min_frames_for_2s)
|
||
|
||
# 创建输出视频路径
|
||
output_path = os.path.join(output_video_root, f"{track_id}.mp4")
|
||
|
||
# 尝试使用GPU硬件编码
|
||
try:
|
||
# 对于不同平台的GPU编码，使用不同的fourcc
|
||
# Windows平台使用h264_nvenc或h264_amf
|
||
# 如果GPU编码不可用，会回退到CPU编码
|
||
fourcc = cv2.VideoWriter_fourcc(*'h264') # type: ignore
|
||
out = cv2.VideoWriter(output_path, fourcc, target_fps, (frame_width, frame_height))
|
||
# 检查是否成功打开
|
||
if not out.isOpened():
|
||
# 尝试其他编码方式
|
||
fourcc = cv2.VideoWriter_fourcc(*'mp4v') # type: ignore
|
||
out = cv2.VideoWriter(output_path, fourcc, target_fps, (frame_width, frame_height))
|
||
except Exception:
|
||
# 异常时使用CPU编码
|
||
fourcc = cv2.VideoWriter_fourcc(*'mp4v') # type: ignore
|
||
out = cv2.VideoWriter(output_path, fourcc, target_fps, (frame_width, frame_height))
|
||
|
||
if not out.isOpened():
|
||
print(f"无法创建视频文件: {output_path}")
|
||
return f"失败: {output_path}"
|
||
|
||
# 打开原视频（每个线程独立打开，避免线程安全问题）
|
||
cap = cv2.VideoCapture(input_video_path)
|
||
if not cap.isOpened():
|
||
out.release()
|
||
return f"失败: 无法打开视频 {input_video_path}"
|
||
|
||
# 生成视频帧
|
||
current_output_frame = 0
|
||
obj_frame_idx = 0
|
||
|
||
while current_output_frame < max_output_frame:
|
||
# 检查当前输出帧是否是5的倍数
|
||
if current_output_frame % frame_interval == 0 and obj_frame_idx < len(frame_list):
|
||
# 这是需要放置物体帧的位置
|
||
original_frame_id, xyxy = frame_list[obj_frame_idx]
|
||
|
||
# 设置原视频读取位置
|
||
cap.set(cv2.CAP_PROP_POS_FRAMES, original_frame_id)
|
||
ret, frame = cap.read()
|
||
|
||
if not ret:
|
||
# 读取失败，使用黑色帧
|
||
output_frame = np.zeros((frame_height, frame_width, 3), dtype=np.uint8)
|
||
else:
|
||
# 有数据，截取对应区域
|
||
x1, y1, x2, y2 = map(int, xyxy)
|
||
# 确保坐标在有效范围内
|
||
x1 = max(0, min(x1, frame_width))
|
||
y1 = max(0, min(y1, frame_height))
|
||
x2 = max(0, min(x2, frame_width))
|
||
y2 = max(0, min(y2, frame_height))
|
||
|
||
# 创建黑色背景
|
||
output_frame = np.zeros((frame_height, frame_width, 3), dtype=np.uint8)
|
||
# 将截取的区域放到输出帧中（保持原位置）
|
||
if x2 > x1 and y2 > y1:
|
||
cropped = frame[y1:y2, x1:x2]
|
||
output_frame[y1:y2, x1:x2] = cropped
|
||
|
||
# 移到下一个物体帧
|
||
obj_frame_idx += 1
|
||
else:
|
||
# 剩余帧留黑
|
||
output_frame = np.zeros((frame_height, frame_width, 3), dtype=np.uint8)
|
||
|
||
# 写入帧
|
||
out.write(output_frame)
|
||
current_output_frame += 1
|
||
|
||
# 释放资源
|
||
out.release()
|
||
cap.release()
|
||
print(f"已生成视频: {output_path}, 共 {current_output_frame} 帧")
|
||
return f"成功: {output_path}"
|
||
|
||
|
||
def frame_all_to_obj_vid(
|
||
json_data: dict,
|
||
input_video_path: str,
|
||
output_video_root: str,
|
||
) -> None:
|
||
"""
|
||
根据标注数据从原视频中截取物体，生成ai读取专用视频
|
||
|
||
参数:
|
||
json_data: 标注数据
|
||
input_video_path: 原视频路径
|
||
output_video_root: 输出视频根目录
|
||
"""
|
||
# 确保输出目录存在
|
||
os.makedirs(output_video_root, exist_ok=True)
|
||
|
||
# 1. 从 json_data 中提取数据，按 track_id 组织
|
||
track_dict: dict[int, list[tuple[int, list[int]]]] = {}
|
||
# 遍历每一帧
|
||
for frame_id_str, detections in json_data.items():
|
||
frame_id = int(frame_id_str)
|
||
for det in detections:
|
||
track_id = det.get("track_id", -1)
|
||
xyxy = det.get("xyxy", [0, 0, 0, 0])
|
||
if track_id not in track_dict:
|
||
track_dict[track_id] = []
|
||
track_dict[track_id].append((frame_id, xyxy))
|
||
|
||
# 2. 获取视频信息（只需要获取一次）
|
||
temp_cap = cv2.VideoCapture(input_video_path)
|
||
if not temp_cap.isOpened():
|
||
raise ValueError(f"无法打开视频: {input_video_path}")
|
||
|
||
frame_width = int(temp_cap.get(cv2.CAP_PROP_FRAME_WIDTH))
|
||
frame_height = int(temp_cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
|
||
temp_cap.release()
|
||
|
||
# 目标fps为25，每5帧取一帧（0, 5, 10...）
|
||
target_fps = 25
|
||
frame_interval = 5 # 每隔5帧取一帧
|
||
|
||
# 3. 使用多线程并行处理多个track_id
|
||
# 根据CPU核心数设置线程池大小
|
||
max_workers = min(os.cpu_count() or 4, len(track_dict))
|
||
print(f"使用 {max_workers} 个线程并行处理")
|
||
|
||
with ThreadPoolExecutor(max_workers=max_workers) as executor:
|
||
# 提交所有任务
|
||
future_to_track = {
|
||
executor.submit(
|
||
process_track_id,
|
||
track_id,
|
||
frame_list,
|
||
input_video_path,
|
||
output_video_root,
|
||
frame_width,
|
||
frame_height,
|
||
target_fps,
|
||
frame_interval
|
||
):
|
||
track_id for track_id, frame_list in track_dict.items()
|
||
}
|
||
|
||
# 等待所有任务完成
|
||
for future in as_completed(future_to_track):
|
||
track_id = future_to_track[future]
|
||
try:
|
||
result = future.result()
|
||
print(f"Track ID {track_id}: {result}")
|
||
except Exception as e:
|
||
print(f"Track ID {track_id} 处理失败: {str(e)}")
|
||
|
||
# def frame_all_to_obj_vid(
|
||
# json_data: dict,
|
||
# input_video_path: str,
|
||
# output_video_root: str,
|
||
# ) -> None:
|
||
# """
|
||
# 根据标注数据从原视频中截取物体，生成ai读取专用视频
|
||
|
||
# 参数:
|
||
# json_data: 标注数据
|
||
# input_video_path: 原视频路径
|
||
# output_video_root: 输出视频根目录
|
||
# """
|
||
# # 确保输出目录存在
|
||
# os.makedirs(output_video_root, exist_ok=True)
|
||
|
||
# # 1. 从 json_data 中提取数据，按 track_id 组织
|
||
# track_dict: dict[int, list[tuple[int, list[int]]]] = {}
|
||
# # 遍历每一帧
|
||
# for frame_id_str, detections in json_data.items():
|
||
# frame_id = int(frame_id_str)
|
||
# for det in detections:
|
||
# track_id = det.get("track_id", -1)
|
||
# xyxy = det.get("xyxy", [0, 0, 0, 0])
|
||
# if track_id not in track_dict:
|
||
# track_dict[track_id] = []
|
||
# track_dict[track_id].append((frame_id, xyxy))
|
||
|
||
# # 2. 打开原视频
|
||
# cap = cv2.VideoCapture(input_video_path)
|
||
# if not cap.isOpened():
|
||
# raise ValueError(f"无法打开视频: {input_video_path}")
|
||
|
||
# # 获取原视频信息
|
||
# original_fps = cap.get(cv2.CAP_PROP_FPS)
|
||
# total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
|
||
# frame_width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
|
||
# frame_height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
|
||
|
||
# # 目标fps为25，每5帧取一帧（0, 5, 10...）
|
||
# target_fps = 25
|
||
# frame_interval = 5 # 每隔5帧取一帧
|
||
|
||
# # 为每个 track_id 生成视频
|
||
# for track_id, frame_list in track_dict.items():
|
||
# # 计算需要生成的总帧数（确保覆盖所有物体帧且不少于两秒）
|
||
# min_frames_for_2s = 25 * 2 # 2秒 @ 25fps
|
||
# object_based_frames = len(frame_list) * frame_interval
|
||
# max_output_frame = max(object_based_frames, min_frames_for_2s)
|
||
|
||
# # 创建输出视频路径
|
||
# output_path = os.path.join(output_video_root, f"{track_id}.mp4")
|
||
|
||
# # 创建视频写入器
|
||
# fourcc = cv2.VideoWriter_fourcc(*'mp4v') # type: ignore
|
||
# out = cv2.VideoWriter(output_path, fourcc, target_fps, (frame_width, frame_height))
|
||
|
||
# if not out.isOpened():
|
||
# print(f"无法创建视频文件: {output_path}")
|
||
# continue
|
||
|
||
# # 生成视频帧
|
||
# current_output_frame = 0
|
||
# obj_frame_idx = 0
|
||
|
||
# while current_output_frame < max_output_frame:
|
||
# # 检查当前输出帧是否是5的倍数
|
||
# if current_output_frame % frame_interval == 0 and obj_frame_idx < len(frame_list):
|
||
# # 这是需要放置物体帧的位置
|
||
# original_frame_id, xyxy = frame_list[obj_frame_idx]
|
||
|
||
# # 设置原视频读取位置
|
||
# cap.set(cv2.CAP_PROP_POS_FRAMES, original_frame_id)
|
||
# ret, frame = cap.read()
|
||
|
||
# if not ret:
|
||
# # 读取失败，使用黑色帧
|
||
# output_frame = np.zeros((frame_height, frame_width, 3), dtype=np.uint8)
|
||
# else:
|
||
# # 有数据，截取对应区域
|
||
# x1, y1, x2, y2 = map(int, xyxy)
|
||
# # 确保坐标在有效范围内
|
||
# x1 = max(0, min(x1, frame_width))
|
||
# y1 = max(0, min(y1, frame_height))
|
||
# x2 = max(0, min(x2, frame_width))
|
||
# y2 = max(0, min(y2, frame_height))
|
||
|
||
# # 创建黑色背景
|
||
# output_frame = np.zeros((frame_height, frame_width, 3), dtype=np.uint8)
|
||
# # 将截取的区域放到输出帧中（保持原位置）
|
||
# if x2 > x1 and y2 > y1:
|
||
# cropped = frame[y1:y2, x1:x2]
|
||
# output_frame[y1:y2, x1:x2] = cropped
|
||
|
||
# # 移到下一个物体帧
|
||
# obj_frame_idx += 1
|
||
# else:
|
||
# # 剩余帧留黑
|
||
# output_frame = np.zeros((frame_height, frame_width, 3), dtype=np.uint8)
|
||
|
||
# # 写入帧
|
||
# out.write(output_frame)
|
||
# current_output_frame += 1
|
||
|
||
# # 释放视频写入器
|
||
# out.release()
|
||
# print(f"已生成视频: {output_path}, 共 {current_output_frame} 帧")
|
||
|
||
# # 释放原视频
|
||
# cap.release()
|
||
|
||
|
||
def create_mian_vid_for_ai(
|
||
input_video_path: str,
|
||
output_folder: str
|
||
) -> str:
|
||
"""
|
||
将原始视频的第0,1,2...帧映射到新视频的0,5,10...帧，其他帧留黑
|
||
|
||
参数:
|
||
input_video_path: 原始视频路径
|
||
output_folder: 输出文件夹路径
|
||
返回:
|
||
str: 输出视频路径
|
||
"""
|
||
# 确保输出目录存在
|
||
os.makedirs(output_folder, exist_ok=True)
|
||
|
||
# 构建输出视频路径
|
||
output_video_path = os.path.join(output_folder, "mian_vid_ai.mp4")
|
||
|
||
# 打开原视频
|
||
cap = cv2.VideoCapture(input_video_path)
|
||
if not cap.isOpened():
|
||
raise ValueError(f"无法打开视频: {input_video_path}")
|
||
|
||
# 获取原视频信息
|
||
original_fps = cap.get(cv2.CAP_PROP_FPS)
|
||
total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
|
||
frame_width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
|
||
frame_height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
|
||
|
||
# 目标fps为25，每5帧取一帧（0, 5, 10...）
|
||
target_fps = 25
|
||
frame_interval = 5
|
||
|
||
# 计算输出视频的总帧数
|
||
# 确保覆盖所有原始帧且不少于两秒
|
||
min_frames_for_2s = 25 * 2 # 2秒 @ 25fps
|
||
max_output_frame = max(total_frames * frame_interval, min_frames_for_2s)
|
||
|
||
# 创建视频写入器
|
||
fourcc = cv2.VideoWriter_fourcc(*'mp4v') # type: ignore
|
||
out = cv2.VideoWriter(output_video_path, fourcc, target_fps, (frame_width, frame_height))
|
||
|
||
if not out.isOpened():
|
||
raise ValueError(f"无法创建视频文件: {output_video_path}")
|
||
|
||
# 生成视频帧
|
||
current_output_frame = 0
|
||
original_frame_idx = 0
|
||
|
||
while current_output_frame < max_output_frame:
|
||
# 检查当前输出帧是否是5的倍数
|
||
if current_output_frame % frame_interval == 0 and original_frame_idx < total_frames:
|
||
# 这是需要放置原始帧的位置
|
||
# 设置原视频读取位置
|
||
cap.set(cv2.CAP_PROP_POS_FRAMES, original_frame_idx)
|
||
ret, frame = cap.read()
|
||
|
||
if not ret:
|
||
# 读取失败，使用黑色帧
|
||
output_frame = np.zeros((frame_height, frame_width, 3), dtype=np.uint8)
|
||
else:
|
||
# 有数据，使用原始帧
|
||
output_frame = frame
|
||
|
||
# 移到下一个原始帧
|
||
original_frame_idx += 1
|
||
else:
|
||
# 剩余帧留黑
|
||
output_frame = np.zeros((frame_height, frame_width, 3), dtype=np.uint8)
|
||
|
||
# 写入帧
|
||
out.write(output_frame)
|
||
current_output_frame += 1
|
||
|
||
# 释放资源
|
||
cap.release()
|
||
out.release()
|
||
print(f"已生成视频: {output_video_path}, 共 {current_output_frame} 帧")
|
||
|
||
return output_video_path
|
||
|