Merge branch 'master' of https://gitea.swiftsnake.cn/yipai-tech/SupervisorAI
This commit is contained in:
@@ -4,6 +4,9 @@
|
||||
|
||||
import cv2
|
||||
import base64
|
||||
import json
|
||||
import os
|
||||
import subprocess
|
||||
import time
|
||||
import threading
|
||||
import queue
|
||||
@@ -12,7 +15,9 @@ from typing import Dict, Any, Callable
|
||||
from concurrent.futures import ThreadPoolExecutor
|
||||
|
||||
from common import constants
|
||||
from common.type_mapping import get_alert_label
|
||||
from utils.logger import get_logger
|
||||
from utils.hls_utils import get_segments_before_current, parse_segment_info
|
||||
|
||||
logger = get_logger(__name__)
|
||||
|
||||
@@ -67,6 +72,12 @@ class BaseFrameProcessorWorker(threading.Thread):
|
||||
max_workers=post_workers,
|
||||
thread_name_prefix="alert_post"
|
||||
)
|
||||
|
||||
# MP4缓存 {segment_path: mp4_path}
|
||||
self._mp4_cache: Dict[str, str] = {}
|
||||
|
||||
# 启动视频文件清理线程
|
||||
self._start_cleanup_thread()
|
||||
|
||||
def _encode_image_to_base64(self, img) -> str:
|
||||
"""图像编码为 Base64"""
|
||||
@@ -80,10 +91,10 @@ class BaseFrameProcessorWorker(threading.Thread):
|
||||
将 msg 中的 result_type 从数组展开为多个独立的 msg
|
||||
|
||||
Args:
|
||||
msg: 原始消息，result_type 为数组
|
||||
msg: 原始消息，result_type 为 action code 字符串数组
|
||||
|
||||
Returns:
|
||||
msg 列表，每个 msg 的 result_type 为数组中的单个元素
|
||||
msg 列表，每个 msg 的 result_type 为包含 action_code 和 action_name 的对象
|
||||
"""
|
||||
result_types = msg.get("result_type", [])
|
||||
if not isinstance(result_types, list):
|
||||
@@ -94,15 +105,18 @@ class BaseFrameProcessorWorker(threading.Thread):
|
||||
return [msg]
|
||||
|
||||
result = []
|
||||
for r_type in result_types:
|
||||
for action_code in result_types:
|
||||
new_msg = msg.copy()
|
||||
new_msg["result_type"] = r_type
|
||||
new_msg["result_type"] = {
|
||||
"action_code": action_code,
|
||||
"action_name": get_alert_label(action_code)
|
||||
}
|
||||
result.append(new_msg)
|
||||
|
||||
return result
|
||||
|
||||
def _post_alert(self, msg: dict):
|
||||
"""异步发送告警 POST 请求（在线程池中执行）"""
|
||||
"""异步发送告警 POST 请求（在线程池中执行）- 旧接口，保留备用"""
|
||||
try:
|
||||
response = requests.post(constants.ALERT_PUSH_URL, json=msg, timeout=5.0)
|
||||
if response.status_code == 200:
|
||||
@@ -112,6 +126,218 @@ class BaseFrameProcessorWorker(threading.Thread):
|
||||
except Exception as e:
|
||||
print(f"[ERROR] POST alert request failed: {e}")
|
||||
|
||||
def _post_alert_with_video(self, msg: dict, video_path: str = None):
|
||||
"""
|
||||
异步发送告警 POST 请求（带视频，multipart/form-data）
|
||||
|
||||
Args:
|
||||
msg: 消息内容
|
||||
video_path: 视频文件路径（可选）
|
||||
"""
|
||||
try:
|
||||
if video_path and os.path.exists(video_path):
|
||||
# 有视频，使用 multipart/form-data 上传
|
||||
with open(video_path, 'rb') as f:
|
||||
files = {
|
||||
'video': f,
|
||||
'metadata': (None, json.dumps(msg))
|
||||
}
|
||||
response = requests.post(constants.ALERT_PUSH_URL, files=files, timeout=10.0)
|
||||
else:
|
||||
# 无视频，也使用 multipart/form-data
|
||||
files = {
|
||||
'metadata': (None, json.dumps(msg))
|
||||
}
|
||||
response = requests.post(constants.ALERT_PUSH_URL, files=files, timeout=5.0)
|
||||
|
||||
if response.status_code == 200:
|
||||
logger.info(f"[INFO] POST alert sent successfully for actions: {msg.get('result_type')}")
|
||||
else:
|
||||
logger.warning(f"[WARN] POST alert failed with status: {response.status_code}")
|
||||
except Exception as e:
|
||||
logger.error(f"[ERROR] POST alert request failed: {e}")
|
||||
|
||||
def _start_cleanup_thread(self):
|
||||
"""启动视频文件清理线程"""
|
||||
def cleanup_loop():
|
||||
while not self.stop_event.is_set():
|
||||
try:
|
||||
self._cleanup_expired_files()
|
||||
except Exception as e:
|
||||
logger.error(f"[ERROR] Cleanup thread error: {e}")
|
||||
# 每10分钟检查一次
|
||||
self.stop_event.wait(600)
|
||||
|
||||
thread = threading.Thread(target=cleanup_loop, daemon=True, name="video_cleanup")
|
||||
thread.start()
|
||||
logger.info("[INFO] Video cleanup thread started")
|
||||
|
||||
def _cleanup_expired_files(self):
|
||||
"""清理过期的视频文件"""
|
||||
output_dir = constants.VIDEO_CLIP_OUTPUT_DIR
|
||||
if not output_dir or not os.path.exists(output_dir):
|
||||
return
|
||||
|
||||
retention_seconds = constants.VIDEO_CLIP_RETENTION_SECONDS
|
||||
current_time = time.time()
|
||||
|
||||
try:
|
||||
for filename in os.listdir(output_dir):
|
||||
if not filename.endswith('.mp4'):
|
||||
continue
|
||||
filepath = os.path.join(output_dir, filename)
|
||||
if os.path.isfile(filepath):
|
||||
file_mtime = os.path.getmtime(filepath)
|
||||
if current_time - file_mtime > retention_seconds:
|
||||
try:
|
||||
os.remove(filepath)
|
||||
logger.info(f"[INFO] Cleaned up expired video: {filename}")
|
||||
except Exception as e:
|
||||
logger.error(f"[ERROR] Failed to delete {filename}: {e}")
|
||||
except Exception as e:
|
||||
logger.error(f"[ERROR] Cleanup expired files error: {e}")
|
||||
|
||||
def _create_or_get_video_clip(self, segment_path: str, segment_duration: float = None) -> str | None:
|
||||
"""
|
||||
创建或获取视频剪辑
|
||||
|
||||
Args:
|
||||
segment_path: 当前TS分片路径
|
||||
segment_duration: 当前分片时长（秒）
|
||||
|
||||
Returns:
|
||||
MP4文件路径，失败返回 None
|
||||
"""
|
||||
if not segment_path:
|
||||
return None
|
||||
|
||||
# 检查缓存
|
||||
if segment_path in self._mp4_cache:
|
||||
cached_path = self._mp4_cache[segment_path]
|
||||
if os.path.exists(cached_path):
|
||||
return cached_path
|
||||
else:
|
||||
# 缓存失效，移除
|
||||
del self._mp4_cache[segment_path]
|
||||
|
||||
# 解析分片信息，构建MP4路径
|
||||
camera_id, timestamp, seq = parse_segment_info(segment_path)
|
||||
if not camera_id:
|
||||
logger.warning(f"[WARN] Failed to parse segment info: {segment_path}")
|
||||
return None
|
||||
|
||||
output_dir = constants.VIDEO_CLIP_OUTPUT_DIR
|
||||
if not output_dir:
|
||||
logger.warning("[WARN] VIDEO_CLIP_OUTPUT_DIR not configured")
|
||||
return None
|
||||
|
||||
# 确保输出目录存在
|
||||
os.makedirs(output_dir, exist_ok=True)
|
||||
|
||||
# MP4文件名
|
||||
mp4_filename = f"{camera_id}_{timestamp}_{seq}.mp4"
|
||||
mp4_path = os.path.join(output_dir, mp4_filename)
|
||||
|
||||
# 检查是否已存在
|
||||
if os.path.exists(mp4_path):
|
||||
self._mp4_cache[segment_path] = mp4_path
|
||||
return mp4_path
|
||||
|
||||
# 计算需要回溯的分片数量
|
||||
clip_duration = constants.VIDEO_CLIP_DURATION_SECONDS
|
||||
default_segment_duration = constants.VIDEO_CLIP_DEFAULT_SEGMENT_DURATION
|
||||
|
||||
effective_duration = segment_duration if segment_duration else default_segment_duration
|
||||
if effective_duration <= 0:
|
||||
effective_duration = default_segment_duration
|
||||
|
||||
n_segments = int(clip_duration / effective_duration) + 1
|
||||
|
||||
# 获取需要合并的分片
|
||||
ts_files = get_segments_before_current(segment_path, n_segments)
|
||||
if not ts_files:
|
||||
logger.warning(f"[WARN] No segments found for clip: {segment_path}")
|
||||
return None
|
||||
|
||||
# 合并TS为MP4
|
||||
if self._merge_ts_to_mp4(ts_files, mp4_path):
|
||||
self._mp4_cache[segment_path] = mp4_path
|
||||
return mp4_path
|
||||
|
||||
return None
|
||||
|
||||
def _merge_ts_to_mp4(self, ts_files: list, output_path: str) -> bool:
|
||||
"""
|
||||
使用 ffmpeg 合并 TS 分片为 MP4
|
||||
|
||||
Args:
|
||||
ts_files: TS文件路径列表（按时间顺序）
|
||||
output_path: 输出MP4路径
|
||||
|
||||
Returns:
|
||||
是否成功
|
||||
"""
|
||||
if not ts_files:
|
||||
return False
|
||||
|
||||
try:
|
||||
# 构建 concat 字符串
|
||||
concat_str = "|".join(ts_files)
|
||||
|
||||
# ffmpeg 命令
|
||||
cmd = [
|
||||
'ffmpeg',
|
||||
'-i', f'concat:{concat_str}',
|
||||
'-c', 'copy',
|
||||
'-y', # 覆盖输出文件
|
||||
output_path
|
||||
]
|
||||
|
||||
# 执行命令
|
||||
result = subprocess.run(
|
||||
cmd,
|
||||
capture_output=True,
|
||||
timeout=60 # 60秒超时
|
||||
)
|
||||
|
||||
if result.returncode == 0:
|
||||
logger.info(f"[INFO] Created video clip: {output_path}")
|
||||
return True
|
||||
else:
|
||||
logger.error(f"[ERROR] ffmpeg failed: {result.stderr.decode()}")
|
||||
return False
|
||||
|
||||
except subprocess.TimeoutExpired:
|
||||
logger.error(f"[ERROR] ffmpeg timeout for {output_path}")
|
||||
return False
|
||||
except FileNotFoundError:
|
||||
logger.error("[ERROR] ffmpeg not found, please install ffmpeg")
|
||||
return False
|
||||
except Exception as e:
|
||||
logger.error(f"[ERROR] Failed to merge TS files: {e}")
|
||||
return False
|
||||
|
||||
def _process_alert_with_video(self, msg: dict, segment_path: str, segment_duration: float):
|
||||
"""
|
||||
处理告警（含视频剪辑）- 在线程池中执行
|
||||
|
||||
Args:
|
||||
msg: 基础消息
|
||||
segment_path: 当前TS分片路径
|
||||
segment_duration: 当前分片时长
|
||||
"""
|
||||
# 尝试创建/获取视频剪辑
|
||||
mp4_path = None
|
||||
if segment_path:
|
||||
mp4_path = self._create_or_get_video_clip(segment_path, segment_duration)
|
||||
|
||||
# 展开 result_type
|
||||
expanded_msgs = self._expand_msg_by_result_type(msg)
|
||||
|
||||
# 发送每个展开后的消息
|
||||
for expanded_msg in expanded_msgs:
|
||||
self._post_alert_with_video(expanded_msg, mp4_path)
|
||||
|
||||
def _create_detector(self, params):
|
||||
"""创建检测器实例"""
|
||||
# 使用 type(self) 访问类属性，避免 lambda 被绑定 self 参数
|
||||
@@ -203,13 +429,24 @@ class BaseFrameProcessorWorker(threading.Thread):
|
||||
try:
|
||||
self.ws_queue.put(msg, timeout=1.0)
|
||||
if push_actions and len(push_actions) > 0:
|
||||
# 异步发送 POST 请求（提交到线程池）
|
||||
# 构建消息
|
||||
post_msg = msg.copy()
|
||||
post_msg['type'] = self.POST_TYPE
|
||||
# 展开 result_type 为多个独立的 msg
|
||||
expanded_msgs = self._expand_msg_by_result_type(post_msg)
|
||||
for expanded_msg in expanded_msgs:
|
||||
self.post_executor.submit(self._post_alert, expanded_msg)
|
||||
|
||||
#备用backup
|
||||
#self.post_executor.submit(self._post_alert, post_msg)
|
||||
|
||||
# 获取视频相关信息（仅HLS模式有）
|
||||
segment_path = item.get("segment_path")
|
||||
segment_duration = item.get("segment_duration")
|
||||
|
||||
# 提交到线程池执行（包含视频剪辑和POST）
|
||||
self.post_executor.submit(
|
||||
self._process_alert_with_video,
|
||||
post_msg,
|
||||
segment_path,
|
||||
segment_duration
|
||||
)
|
||||
except queue.Full:
|
||||
logger.warning("[WARN] ws_send_queue full, drop frame message")
|
||||
|
||||
|
||||
@@ -372,12 +372,12 @@ class KadianDetector:
|
||||
|
||||
# 情况1：通过时间太短 -> Ignore (Too Fast)
|
||||
if duration_frames < self.frame_thresh_car_min_duration:
|
||||
print(f"ALARM: Car {car_id} passed too fast -> Regarded as Ignore Checked!")
|
||||
logger.info(f"ALARM: Car {car_id} passed too fast -> Regarded as Ignore Checked!")
|
||||
self.fast_pass_alerts[car_id] = self.current_frame_idx + int(self.ignore_show_seconds * self.fps)
|
||||
|
||||
# 情况2：时间够长，但没检查后备箱 -> Unchecked Trunk
|
||||
elif not car_info['is_checked']:
|
||||
print(f"ALARM: Car {car_id} left without checking trunk!")
|
||||
logger.info(f"ALARM: Car {car_id} left without checking trunk!")
|
||||
self.unchecked_trunk_alerts[car_id] = self.current_frame_idx + int(
|
||||
self.openTrunk_show_seconds * self.fps)
|
||||
|
||||
|
||||
@@ -1,515 +0,0 @@
|
||||
|
||||
import cv2
|
||||
import numpy as np
|
||||
from typing import Dict, Any
|
||||
import threading
|
||||
import queue
|
||||
|
||||
from biz.base_frame_processor import BaseFrameProcessorWorker
|
||||
|
||||
# -------------------------- Kadian 检测相关导入 --------------------------
|
||||
from algorithm.common.npu_yolo_onnx_person_car_phone import YOLOv8_ONNX # 主检测模型（人/车/后备箱/手机）
|
||||
from algorithm.common.npu_yolo_pose_onnx import YOLOv8_Pose_ONNX # Pose 专用模型
|
||||
from yolox.tracker.byte_tracker import BYTETracker
|
||||
|
||||
from utils.logger import get_logger
|
||||
logger = get_logger(__name__)
|
||||
|
||||
# ========================= 配置区 =========================
|
||||
# Kadian 模型路径与ROI（可根据实际情况修改）
|
||||
DETECT_MODEL_PATH = 'YOLO_Weight/Kadian.onnx'
|
||||
#POSE_MODEL_PATH = 'YOLO_Weight/yolov8l-pose.onnx'
|
||||
|
||||
# 默认相对ROI（与原文件一致）
|
||||
#ROI_RELATIVE = np.array([
|
||||
# [0.10989583333333333, 0.006481481481481481],
|
||||
# [0.421875, 0.005555555555555556],
|
||||
# [0.9921875, 0.9888888888888889],
|
||||
# [0.3411458333333333, 0.9861111111111112]
|
||||
#])
|
||||
|
||||
# ROI_RELATIVE=np.array([
|
||||
# [0.15,0.001],
|
||||
# [0.5,0.001],
|
||||
# [1.0,0.8],
|
||||
# [0.35,1.0]
|
||||
# ])
|
||||
|
||||
ROI_RELATIVE=np.array([
|
||||
[0.12,0.0],
|
||||
[0.3,0.0],
|
||||
[0.5,0.2],
|
||||
[1.0, 0.95],
|
||||
[1.0,1.0],
|
||||
[0.42,1.0]
|
||||
])
|
||||
|
||||
|
||||
ALERT_PUSH_INTERVAL = 5.0
|
||||
|
||||
# 输入尺寸
|
||||
PERSON_CAR_INPUT_SIZE = 640
|
||||
|
||||
|
||||
|
||||
RTSP_TARGET_FPS = 10.0
|
||||
|
||||
|
||||
|
||||
class KadianDetector:
|
||||
def __init__(self, roi_points=ROI_RELATIVE):
|
||||
# 模型加载 - 仅保留主检测器，删除pose_detector
|
||||
self.detector = YOLOv8_ONNX(
|
||||
DETECT_MODEL_PATH,
|
||||
conf_threshold=0.25,
|
||||
iou_threshold=0.45,
|
||||
input_size=PERSON_CAR_INPUT_SIZE
|
||||
)
|
||||
|
||||
# 跟踪器配置
|
||||
class TrackerArgs:
|
||||
track_thresh = 0.3 # 必须大于等于yolo的conf_threshold
|
||||
track_buffer = 40
|
||||
match_thresh = 0.85
|
||||
mot20 = True
|
||||
|
||||
self.fps = RTSP_TARGET_FPS
|
||||
self.tracker = BYTETracker(TrackerArgs(), frame_rate=self.fps)
|
||||
self.track_role = {} # 跟踪ID到类别的映射
|
||||
|
||||
|
||||
# ROI 处理（支持相对/绝对）
|
||||
self.roi_points = np.array(roi_points, dtype=np.float64) if roi_points is not None else None
|
||||
|
||||
# ===================== 超参数设置 (仅保留车/后备箱相关) =====================
|
||||
# 后备箱检查判定阈值
|
||||
self.TIME_THRESHOLD_TRUNK_OPEN = 0.1
|
||||
# 车辆最小停留时间阈值 (小于此时间视为无人检查/直接通过)
|
||||
self.TIME_THRESHOLD_CAR_MIN_DURATION = 3.0
|
||||
# Car 丢帧/ID维持缓冲
|
||||
self.TIME_TOLERANCE_CAR = 2.0
|
||||
|
||||
# police丢失阈值
|
||||
self.TIME_TOLERANCE_POLICE = 3.0
|
||||
# police状态判定阈值 (累计秒数)
|
||||
self.TIME_THRESHOLD_NOBODY = 5.0
|
||||
self.TIME_THRESHOLD_ONLY_ONE = 5.0
|
||||
|
||||
# --- 计算对应的帧数阈值 ---
|
||||
self.frame_thresh_trunk_valid = int(self.TIME_THRESHOLD_TRUNK_OPEN * self.fps)
|
||||
self.frame_thresh_car_min_duration = int(self.TIME_THRESHOLD_CAR_MIN_DURATION * self.fps)
|
||||
self.frame_buffer_limit_car = int(self.TIME_TOLERANCE_CAR * self.fps)
|
||||
self.frame_buffer_limit_police = int(self.TIME_TOLERANCE_POLICE * self.fps)
|
||||
self.frame_thresh_nobody = int(self.TIME_THRESHOLD_NOBODY * self.fps)
|
||||
self.frame_thresh_only_one = int(self.TIME_THRESHOLD_ONLY_ONE * self.fps)
|
||||
|
||||
# 显示相关阈值
|
||||
self.ignore_show_seconds = 0.2 # 未检测的警告显示时长
|
||||
self.openTrunk_show_seconds = 0.2 # 打开后备箱的警告显示时长
|
||||
self.police_show_seconds = 0.2 # 警察在场警告显示时长
|
||||
|
||||
# 状态变量初始化
|
||||
self.current_frame_idx = 0
|
||||
self.width = 0
|
||||
self.height = 0
|
||||
|
||||
|
||||
# 车辆注册表 (字典)
|
||||
self.roi_car_registry = {}
|
||||
# 违规车辆记录
|
||||
self.unchecked_trunk_alerts = {} # 后备箱未检
|
||||
self.fast_pass_alerts = {} # 通过过快
|
||||
|
||||
# 警察注册表 (字典)
|
||||
self.roi_police_registry = {}
|
||||
# 警察在场告警记录
|
||||
self.nobody_alerts = {} # 无人在场
|
||||
self.only_one_alerts = {} # 单人在场
|
||||
# 累计帧数计数器
|
||||
self.nobody_frames = 0 # 累计无人在场帧数
|
||||
self.only_one_frames = 0 # 累计单人在场帧数
|
||||
|
||||
# 打印超参数
|
||||
print(f"\n超参数设置:")
|
||||
print(f" FPS: {self.fps:.2f}")
|
||||
print(f" 判定 'Trunk Checked' 需累计检测: {self.frame_thresh_trunk_valid} 帧")
|
||||
print(f" 判定 'Too Fast' 最小停留: {self.frame_thresh_car_min_duration} 帧")
|
||||
|
||||
def _get_roi_points(self, frame_width: int, frame_height: int):
|
||||
"""
|
||||
每帧动态计算正确的 ROI 绝对坐标，并确保类型为 np.int32
|
||||
用于 pointPolygonTest 和 polylines
|
||||
"""
|
||||
if self.roi_points is None:
|
||||
raise ValueError("ROI points must be provided; cannot be None.")
|
||||
|
||||
if self.roi_points.max() <= 1.0:
|
||||
# 相对坐标 → 转换为绝对
|
||||
roi_abs = self.roi_points * np.array([frame_width, frame_height])
|
||||
else:
|
||||
# 绝对坐标，直接使用
|
||||
roi_abs = self.roi_points.copy()
|
||||
|
||||
# 强制转为 int32（关键！解决 OpenCV 断言错误）
|
||||
return roi_abs.astype(np.int32)
|
||||
|
||||
def check_point_in_roi(self, roi_points, point):
|
||||
"""判断点是否在ROI内"""
|
||||
return cv2.pointPolygonTest(roi_points, point, False) >= 0
|
||||
|
||||
def compute_iou(self, boxA, boxB):
|
||||
"""计算两个框的IOU"""
|
||||
# box = [x1, y1, x2, y2]
|
||||
xA = max(boxA[0], boxB[0])
|
||||
yA = max(boxA[1], boxB[1])
|
||||
xB = min(boxA[2], boxB[2])
|
||||
yB = min(boxA[3], boxB[3])
|
||||
|
||||
interW = max(0, xB - xA)
|
||||
interH = max(0, yB - yA)
|
||||
interArea = interW * interH
|
||||
|
||||
boxAArea = (boxA[2] - boxA[0]) * (boxA[3] - boxA[1])
|
||||
boxBArea = (boxB[2] - boxB[0]) * (boxB[3] - boxB[1])
|
||||
|
||||
unionArea = boxAArea + boxBArea - interArea
|
||||
if unionArea == 0:
|
||||
return 0.0
|
||||
|
||||
return interArea / unionArea
|
||||
|
||||
def draw_alert(self, frame, text, color=(0, 0, 255), sub_text=None, offset_y=0):
|
||||
"""在右上角绘制警告文字 (支持垂直偏移，防止文字重叠)"""
|
||||
font_scale = 1.5
|
||||
thickness = 3
|
||||
font = cv2.FONT_HERSHEY_SIMPLEX
|
||||
|
||||
(text_w, text_h), _ = cv2.getTextSize(text, font, font_scale, thickness)
|
||||
x = self.width - text_w - 20
|
||||
y = 50 + text_h + offset_y # 增加 Y 轴偏移
|
||||
|
||||
cv2.rectangle(frame, (x - 10, y - text_h - 10), (x + text_w + 10, y + 10), (0, 0, 0), -1)
|
||||
cv2.putText(frame, text, (x, y), font, font_scale, color, thickness)
|
||||
|
||||
if sub_text:
|
||||
cv2.putText(frame, sub_text, (x, y + 40), font, 0.7, (200, 200, 200), 2)
|
||||
|
||||
def is_point_in_box(self, point, box):
|
||||
"""判断点是否在框内"""
|
||||
px, py = point
|
||||
x1, y1, x2, y2 = box
|
||||
return x1 < px < x2 and y1 < py < y2
|
||||
|
||||
def process_frame(self, frame, camera_id: int, timestamp: float) -> Dict[str, Any]:
|
||||
h, w = frame.shape[:2]
|
||||
self.width, self.height = w, h
|
||||
|
||||
self.current_frame_idx += 1
|
||||
# ========= 每帧动态获取正确的 ROI（int32）=========
|
||||
roi_points_int32 = self._get_roi_points(w, h) # shape: (4, 2), dtype: int32
|
||||
roi_points_draw = roi_points_int32.reshape((-1, 1, 2)) # shape: (4, 1, 2) 用于绘制
|
||||
|
||||
current_time_sec = timestamp
|
||||
|
||||
# ========= 主检测（删除pose检测）=========
|
||||
detections = self.detector(frame)
|
||||
|
||||
dets_xyxy = []
|
||||
dets_roles = []
|
||||
dets_for_tracker = []
|
||||
|
||||
# ========= 当前帧所有警告列表 ==========
|
||||
current_frame_alerts = [] # 每帧清空，重新收集
|
||||
|
||||
if detections:
|
||||
for det in detections:
|
||||
x1, y1, x2, y2, conf, cls_id = det # x1,y1:左上角，x2,y2:右下角
|
||||
dets_xyxy.append([x1, y1, x2, y2])
|
||||
dets_for_tracker.append([x1, y1, x2, y2, conf])
|
||||
|
||||
# 更新类别映射：0=Car,1=OpenTrunk,2=Passerby,3=Police
|
||||
if cls_id == 0:
|
||||
dets_roles.append("car")
|
||||
elif cls_id == 1:
|
||||
dets_roles.append("opentrunk")
|
||||
elif cls_id == 2:
|
||||
dets_roles.append("passerby") # 路人
|
||||
elif cls_id == 3:
|
||||
dets_roles.append("police") # 警察
|
||||
|
||||
dets = np.array(dets_for_tracker, dtype=np.float32) if len(dets_for_tracker) else np.empty((0, 5))
|
||||
|
||||
# 跟踪器更新
|
||||
tracks = self.tracker.update(
|
||||
dets,
|
||||
[self.height, self.width],
|
||||
[self.height, self.width]
|
||||
)
|
||||
|
||||
# ========= 绘制 ROI =========
|
||||
cv2.polylines(frame, [roi_points_draw], isClosed=True, color=(255, 0, 0), thickness=3)
|
||||
|
||||
# ========= 单帧统计变量 =========
|
||||
current_roi_trunk_count = 0 # 仅保留后备箱统计
|
||||
current_roi_police_count = 0 # ROI内警察数量
|
||||
|
||||
# 临时存储本帧的目标，用于后续关联分析
|
||||
current_cars = [] # {'id':, 'box':}
|
||||
current_trunks = [] # (cx, cy)
|
||||
|
||||
# ========= 处理跟踪结果 =========
|
||||
for t in tracks:
|
||||
tid = t.track_id
|
||||
REVALIDATE_FRAME_INTERVAL = 10
|
||||
|
||||
# 定期重新匹配跟踪ID的类别
|
||||
if (self.current_frame_idx % REVALIDATE_FRAME_INTERVAL == 0) or (tid not in self.track_role):
|
||||
best_iou = 0
|
||||
best_role = "unknown"
|
||||
t_box = list(map(float, t.tlbr)) # [x1,y1,x2,y2]
|
||||
|
||||
for i, box in enumerate(dets_xyxy):
|
||||
iou_val = self.compute_iou(t_box, box)
|
||||
if iou_val > best_iou:
|
||||
best_iou = iou_val
|
||||
best_role = dets_roles[i]
|
||||
if best_iou > 0.1:
|
||||
self.track_role[tid] = best_role
|
||||
else:
|
||||
self.track_role[tid] = "unknown"
|
||||
|
||||
role = self.track_role.get(tid, "unknown")
|
||||
x1, y1, x2, y2 = map(int, t.tlbr)
|
||||
cx, cy = (x1 + x2) // 2, (y1 + y2) // 2
|
||||
|
||||
# 定义不同类别的颜色（仅标框，不告警）
|
||||
if role == "car":
|
||||
color = (0, 255, 0) # 绿色
|
||||
label = f"Car:{tid}"
|
||||
# 仅处理ROI内的车辆
|
||||
if self.check_point_in_roi(roi_points_int32, (cx, cy)):
|
||||
current_cars.append({'id': tid, 'box': [x1, y1, x2, y2]})
|
||||
# 车辆注册表初始化
|
||||
if tid not in self.roi_car_registry:
|
||||
self.roi_car_registry[tid] = {
|
||||
'first_seen': self.current_frame_idx,
|
||||
'last_seen': self.current_frame_idx,
|
||||
'trunk_frames': 0,
|
||||
'is_checked': False,
|
||||
}
|
||||
else:
|
||||
self.roi_car_registry[tid]['last_seen'] = self.current_frame_idx
|
||||
label += " IN"
|
||||
elif role == "opentrunk":
|
||||
color = (255, 165, 0) # 橙色
|
||||
label = "OpenTrunk"
|
||||
if self.check_point_in_roi(roi_points_int32, (cx, cy)):
|
||||
current_roi_trunk_count += 1
|
||||
current_trunks.append((cx, cy))
|
||||
label += " IN"
|
||||
elif role == "passerby":
|
||||
color = (255, 255, 0) # 黄色（仅标框，不告警）
|
||||
label = "Passerby"
|
||||
elif role == "police":
|
||||
color = (0, 255, 255) # 青色
|
||||
label = "Police"
|
||||
if self.check_point_in_roi(roi_points_int32, (cx, cy)):
|
||||
current_roi_police_count += 1
|
||||
# 警察注册表初始化
|
||||
if tid not in self.roi_police_registry:
|
||||
self.roi_police_registry[tid] = {
|
||||
'first_seen': self.current_frame_idx,
|
||||
'last_seen': self.current_frame_idx,
|
||||
}
|
||||
else:
|
||||
self.roi_police_registry[tid]['last_seen'] = self.current_frame_idx
|
||||
label += " IN"
|
||||
else:
|
||||
color = (255, 255, 255) # 白色
|
||||
label = "Unknown"
|
||||
|
||||
# 绘制检测框和标签（所有类别都标框，仅车/后备箱有逻辑）
|
||||
cv2.rectangle(frame, (x1, y1), (x2, y2), color, 2)
|
||||
cv2.putText(frame, label, (x1, y1 - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.6, color, 2)
|
||||
|
||||
# ==========================================
|
||||
# 关联分析: 哪个后备箱属于哪辆车?
|
||||
# ==========================================
|
||||
for car_info in current_cars:
|
||||
c_id = car_info['id']
|
||||
c_box = car_info['box']
|
||||
trunk_found_for_this_car = False
|
||||
|
||||
for t_pt in current_trunks:
|
||||
if self.is_point_in_box(t_pt, c_box):
|
||||
trunk_found_for_this_car = True
|
||||
break
|
||||
|
||||
if trunk_found_for_this_car:
|
||||
self.roi_car_registry[c_id]['trunk_frames'] += 1
|
||||
if self.roi_car_registry[c_id]['trunk_frames'] >= self.frame_thresh_trunk_valid:
|
||||
self.roi_car_registry[c_id]['is_checked'] = True
|
||||
|
||||
# ==========================================
|
||||
# 维护车辆注册表 & 生成离场报警
|
||||
# ==========================================
|
||||
active_car_ids = []
|
||||
cars_to_remove = []
|
||||
|
||||
for car_id, info in self.roi_car_registry.items():
|
||||
last_seen = info['last_seen']
|
||||
if (self.current_frame_idx - last_seen) <= self.frame_buffer_limit_car:
|
||||
active_car_ids.append(car_id)
|
||||
else:
|
||||
cars_to_remove.append(car_id)
|
||||
|
||||
# 处理离场车辆，生成违规告警
|
||||
for car_id in cars_to_remove:
|
||||
car_info = self.roi_car_registry[car_id]
|
||||
duration_frames = car_info['last_seen'] - car_info['first_seen']
|
||||
|
||||
# 情况1：通过时间太短 -> Ignore (Too Fast)
|
||||
if duration_frames < self.frame_thresh_car_min_duration:
|
||||
print(f"ALARM: Car {car_id} passed too fast -> Regarded as Ignore Checked!")
|
||||
self.fast_pass_alerts[car_id] = self.current_frame_idx + int(self.ignore_show_seconds * self.fps)
|
||||
|
||||
# 情况2：时间够长，但没检查后备箱 -> Unchecked Trunk
|
||||
elif not car_info['is_checked']:
|
||||
print(f"ALARM: Car {car_id} left without checking trunk!")
|
||||
self.unchecked_trunk_alerts[car_id] = self.current_frame_idx + int(
|
||||
self.openTrunk_show_seconds * self.fps)
|
||||
|
||||
del self.roi_car_registry[car_id]
|
||||
|
||||
effective_car_count = len(active_car_ids)
|
||||
|
||||
# ==========================================
|
||||
# 维护警察注册表
|
||||
# ==========================================
|
||||
active_police_ids = []
|
||||
polices_to_remove = []
|
||||
|
||||
for police_id, info in self.roi_police_registry.items():
|
||||
last_seen = info['last_seen']
|
||||
if (self.current_frame_idx - last_seen) <= self.frame_buffer_limit_police:
|
||||
active_police_ids.append(police_id)
|
||||
else:
|
||||
polices_to_remove.append(police_id)
|
||||
|
||||
for police_id in polices_to_remove:
|
||||
del self.roi_police_registry[police_id]
|
||||
|
||||
effective_police_count = len(active_police_ids)
|
||||
|
||||
# ==========================================
|
||||
# 显示调试信息和报警 (仅保留车/后备箱相关)
|
||||
# ==========================================
|
||||
# 调试信息
|
||||
debug_info = f"Cars: {len(active_car_ids)} | Trunk: {current_roi_trunk_count} | Police: {effective_police_count} | Nobody:{self.nobody_frames}/{self.frame_thresh_nobody} | OnlyOne:{self.only_one_frames}/{self.frame_thresh_only_one}"
|
||||
cv2.putText(frame, debug_info, (20, 40), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (0, 255, 0), 2)
|
||||
|
||||
# 报警偏移量（防止重叠）
|
||||
alert_offset = 0
|
||||
|
||||
# A. 显示 Trunk Checked (车辆已检查后备箱)
|
||||
# for car_id in active_car_ids:
|
||||
# if car_id in self.roi_car_registry and self.roi_car_registry[car_id]['is_checked']:
|
||||
# current_frame_alerts.append({
|
||||
# 'time': current_time_sec,
|
||||
# 'action': "Trunk Checked",
|
||||
# })
|
||||
# self.draw_alert(frame, "Trunk Checked!!", (0, 255, 0), offset_y=alert_offset)
|
||||
# alert_offset += 100
|
||||
# break # 只显示一次
|
||||
|
||||
# B. 显示 Unchecked Trunk (离场未检查后备箱)
|
||||
expired_alerts = [cid for cid, end_frame in self.unchecked_trunk_alerts.items() if
|
||||
self.current_frame_idx > end_frame]
|
||||
for cid in expired_alerts:
|
||||
del self.unchecked_trunk_alerts[cid]
|
||||
|
||||
if len(self.unchecked_trunk_alerts) > 0:
|
||||
alert_text = f"Unchecked Trunk! (ID:{list(self.unchecked_trunk_alerts.keys())})"
|
||||
current_frame_alerts.append({
|
||||
'time': current_time_sec,
|
||||
'action': "Unchecked Trunk",
|
||||
})
|
||||
self.draw_alert(frame, alert_text, (0, 0, 255), offset_y=alert_offset)
|
||||
alert_offset += 100
|
||||
|
||||
# C. 显示 Ignore (通过过快)
|
||||
expired_fast_alerts = [cid for cid, end_frame in self.fast_pass_alerts.items() if
|
||||
self.current_frame_idx > end_frame]
|
||||
for cid in expired_fast_alerts:
|
||||
del self.fast_pass_alerts[cid]
|
||||
|
||||
if len(self.fast_pass_alerts) > 0:
|
||||
alert_text = f"Ignore: (ID:{list(self.fast_pass_alerts.keys())})"
|
||||
current_frame_alerts.append({
|
||||
'time': current_time_sec,
|
||||
'action': "Ignore",
|
||||
})
|
||||
self.draw_alert(frame, alert_text, (0, 0, 255), offset_y=alert_offset)
|
||||
alert_offset += 100
|
||||
|
||||
# D. 显示警察在场状态 (Nobody/Only One)
|
||||
# 清理过期的 Nobody 告警
|
||||
expired_nobody = [k for k, v in self.nobody_alerts.items() if self.current_frame_idx > v]
|
||||
for k in expired_nobody:
|
||||
del self.nobody_alerts[k]
|
||||
|
||||
# 清理过期的 Only One 告警
|
||||
expired_only_one = [k for k, v in self.only_one_alerts.items() if self.current_frame_idx > v]
|
||||
for k in expired_only_one:
|
||||
del self.only_one_alerts[k]
|
||||
|
||||
if effective_car_count > 0:
|
||||
# 更新累计帧数
|
||||
if effective_police_count == 0:
|
||||
self.nobody_frames += 1
|
||||
self.only_one_frames = 0
|
||||
elif effective_police_count == 1:
|
||||
self.only_one_frames += 1
|
||||
self.nobody_frames = 0
|
||||
else:
|
||||
self.nobody_frames = 0
|
||||
self.only_one_frames = 0
|
||||
else:
|
||||
self.nobody_frames = 0
|
||||
self.only_one_frames = 0
|
||||
|
||||
if effective_police_count == 0 and self.nobody_frames >= self.frame_thresh_nobody:
|
||||
alert_text = "Nobody"
|
||||
if "Nobody" not in self.nobody_alerts:
|
||||
self.nobody_alerts["Nobody"] = self.current_frame_idx + int(self.police_show_seconds * self.fps)
|
||||
current_frame_alerts.append({
|
||||
'time': current_time_sec,
|
||||
'action': "Nobody",
|
||||
})
|
||||
#self.draw_alert(frame, alert_text, (0, 0, 255), offset_y=alert_offset)
|
||||
alert_offset += 100
|
||||
elif effective_police_count == 1 and self.only_one_frames >= self.frame_thresh_only_one:
|
||||
alert_text = "Only One"
|
||||
if "Only One" not in self.only_one_alerts:
|
||||
self.only_one_alerts["Only One"] = self.current_frame_idx + int(self.police_show_seconds * self.fps)
|
||||
current_frame_alerts.append({
|
||||
'time': current_time_sec,
|
||||
'action': "Only One",
|
||||
})
|
||||
#self.draw_alert(frame, alert_text, (255, 165, 0), offset_y=alert_offset)
|
||||
alert_offset += 100
|
||||
|
||||
return {
|
||||
"image": frame,
|
||||
"alerts": current_frame_alerts,
|
||||
}
|
||||
|
||||
|
||||
# ========================= 帧处理线程 =========================
|
||||
class FrameProcessorWorker(BaseFrameProcessorWorker):
|
||||
"""卡点检测帧处理线程"""
|
||||
|
||||
# 子类配置
|
||||
DETECTOR_FACTORY = lambda params: KadianDetector(params)
|
||||
POST_TYPE = 1
|
||||
TARGET_FPS = RTSP_TARGET_FPS
|
||||
259
biz/prison/ab_biz.py
Normal file
259
biz/prison/ab_biz.py
Normal file
@@ -0,0 +1,259 @@
|
||||
import cv2
|
||||
import numpy as np
|
||||
import base64
|
||||
from typing import Dict, Any
|
||||
import threading
|
||||
import time
|
||||
import queue
|
||||
import requests
|
||||
from biz.base_frame_processor import BaseFrameProcessorWorker
|
||||
|
||||
# -------------------------- Kadian 检测相关导入 --------------------------
|
||||
from algorithm.common.npu_yolo_onnx_person_car_phone import YOLOv8_ONNX # 主检测模型（人/车/后备箱/手机）
|
||||
from common.constants import ALERT_PUSH_URL
|
||||
|
||||
from yolox.tracker.byte_tracker import BYTETracker
|
||||
|
||||
|
||||
# ========================= 配置区 =========================
|
||||
# Kadian 模型路径与ROI（可根据实际情况修改）
|
||||
detector_model_path = 'YOLO_Weight/bag_model.onnx'
|
||||
|
||||
# 输入尺寸
|
||||
input_size = 640
|
||||
|
||||
RTSP_TARGET_FPS = 10.0
|
||||
|
||||
# 新增：告警推送频率限制（秒）
|
||||
ALERT_PUSH_INTERVAL = 5.0 # 相同action 5秒内仅推送一次
|
||||
|
||||
|
||||
class AbDetector:
|
||||
def __init__(self, params=None):
|
||||
# 摄像头额外参数
|
||||
self.params = params if params is not None else {}
|
||||
|
||||
# 模型加载
|
||||
|
||||
self.detector = YOLOv8_ONNX(detector_model_path, conf_threshold=0.5, iou_threshold=0.45,
|
||||
input_size=input_size)
|
||||
|
||||
# ByteTracker
|
||||
class TrackerArgs:
|
||||
track_thresh = 0.25
|
||||
track_buffer = 30
|
||||
match_thresh = 0.8
|
||||
mot20 = False
|
||||
|
||||
self.track_role = {}
|
||||
|
||||
self.fps = RTSP_TARGET_FPS
|
||||
|
||||
self.tracker = BYTETracker(TrackerArgs(), frame_rate=self.fps)
|
||||
|
||||
# ==========================================
|
||||
# 超参数设置 (Hyperparameters)
|
||||
# ==========================================
|
||||
self.TIME_THRESHOLD_BLACKBAG = 1.0 # 黑包判定时长（秒）
|
||||
self.TIME_TOLERANCE_BLACKBAG = 0.5 # 黑包丢失缓冲时间
|
||||
|
||||
# 转换为帧数阈值
|
||||
self.frame_thresh_blackbag = int(self.TIME_THRESHOLD_BLACKBAG * self.fps)
|
||||
self.frame_buffer_blackbag = int(self.TIME_TOLERANCE_BLACKBAG * self.fps)
|
||||
|
||||
print(f"\n超参数设置:")
|
||||
print(f" FPS: {self.fps:.2f}")
|
||||
print(f" 判定 'BlackBag Detected' 需累计检测: {self.frame_thresh_blackbag} 帧")
|
||||
print(f" 黑包丢失缓冲帧数: {self.frame_buffer_blackbag} 帧")
|
||||
|
||||
# ==========================================
|
||||
# 状态变量初始化
|
||||
# ==========================================
|
||||
self.current_frame_idx = 0
|
||||
|
||||
# 黑包检测状态
|
||||
self.blackbag_detection_frames = 0
|
||||
self.blackbag_missing_frames = 0
|
||||
self.blackbag_alert_active = False
|
||||
|
||||
# 人员统计变量
|
||||
self.current_person_count = 0
|
||||
|
||||
|
||||
def compute_iou(self,boxA, boxB):
|
||||
# box = [x1, y1, x2, y2]
|
||||
xA = max(boxA[0], boxB[0])
|
||||
yA = max(boxA[1], boxB[1])
|
||||
xB = min(boxA[2], boxB[2])
|
||||
yB = min(boxA[3], boxB[3])
|
||||
|
||||
interW = max(0, xB - xA)
|
||||
interH = max(0, yB - yA)
|
||||
interArea = interW * interH
|
||||
|
||||
boxAArea = (boxA[2] - boxA[0]) * (boxA[3] - boxA[1])
|
||||
boxBArea = (boxB[2] - boxB[0]) * (boxB[3] - boxB[1])
|
||||
|
||||
unionArea = boxAArea + boxBArea - interArea
|
||||
if unionArea == 0:
|
||||
return 0.0
|
||||
|
||||
return interArea / unionArea
|
||||
|
||||
def draw_alert(self, frame, text, color=(0, 0, 255), sub_text=None, offset_y=0):
|
||||
"""在右上角绘制警告文字 (支持垂直偏移，防止文字重叠)"""
|
||||
font_scale = 1.5
|
||||
thickness = 3
|
||||
font = cv2.FONT_HERSHEY_SIMPLEX
|
||||
|
||||
(text_w, text_h), _ = cv2.getTextSize(text, font, font_scale, thickness)
|
||||
x = self.width - text_w - 20
|
||||
y = 50 + text_h + offset_y # 增加 Y 轴偏移
|
||||
|
||||
cv2.rectangle(frame, (x - 10, y - text_h - 10), (x + text_w + 10, y + 10), (0, 0, 0), -1)
|
||||
cv2.putText(frame, text, (x, y), font, font_scale, color, thickness)
|
||||
|
||||
if sub_text:
|
||||
cv2.putText(frame, sub_text, (x, y + 40), font, 0.7, (200, 200, 200), 2)
|
||||
|
||||
def process_frame(self, frame, camera_id: int, timestamp: float) -> Dict[str, Any]:
|
||||
h, w = frame.shape[:2]
|
||||
self.width, self.height = w, h
|
||||
|
||||
self.current_frame_idx += 1
|
||||
|
||||
current_time_sec = timestamp
|
||||
|
||||
# ========= 检测推理（黑包+人）=========
|
||||
detect_results = self.detector(frame)
|
||||
|
||||
# 初始化检测结果存储
|
||||
dets_xyxy = []
|
||||
dets_roles = []
|
||||
dets_for_tracker = []
|
||||
current_frame_alerts = []
|
||||
|
||||
# 解析检测结果（黑包cls_id=0，人员cls_id=1）
|
||||
if detect_results:
|
||||
for det in detect_results:
|
||||
x1, y1, x2, y2, conf, cls_id = det
|
||||
dets_xyxy.append([x1, y1, x2, y2])
|
||||
dets_for_tracker.append([x1, y1, x2, y2, conf])
|
||||
if cls_id == 0:
|
||||
dets_roles.append("black_bag")
|
||||
elif cls_id == 1:
|
||||
dets_roles.append("person")
|
||||
|
||||
# 跟踪器更新
|
||||
dets = np.array(dets_for_tracker, dtype=np.float32) if len(dets_for_tracker) else np.empty((0, 5))
|
||||
tracks = self.tracker.update(
|
||||
dets,
|
||||
[self.height, self.width],
|
||||
[self.height, self.width]
|
||||
)
|
||||
|
||||
# ========= 单帧统计初始化 =========
|
||||
self.current_person_count = 0
|
||||
current_blackbag_count = 0
|
||||
|
||||
# ========= 跟踪结果绘制与统计 =========
|
||||
for t in tracks:
|
||||
tid = t.track_id
|
||||
|
||||
# IoU匹配跟踪ID和类别
|
||||
REVALIDATE_FRAME_INTERVAL = 10
|
||||
|
||||
|
||||
#if tid not in self.track_role:
|
||||
if (self.current_frame_idx % REVALIDATE_FRAME_INTERVAL == 0) or (tid not in self.track_role):
|
||||
best_iou = 0
|
||||
best_role = "unknown"
|
||||
t_box = list(map(float, t.tlbr))
|
||||
for i, box in enumerate(dets_xyxy):
|
||||
iou_val = self.compute_iou(t_box, box)
|
||||
if iou_val > best_iou:
|
||||
best_iou = iou_val
|
||||
best_role = dets_roles[i]
|
||||
self.track_role[tid] = best_role if best_iou > 0.1 else "unknown"
|
||||
|
||||
role = self.track_role.get(tid, "unknown")
|
||||
x1, y1, x2, y2 = map(int, t.tlbr)
|
||||
color = (255, 255, 255)
|
||||
label = "Unknown"
|
||||
|
||||
# 人员检测（cls_id=1）
|
||||
if role == "person":
|
||||
self.current_person_count += 1
|
||||
color = (255, 0, 255) # 紫色框
|
||||
label = "Person"
|
||||
# 黑包检测（cls_id=0）
|
||||
elif role == "black_bag":
|
||||
current_blackbag_count += 1
|
||||
color = (0, 128, 0) # 绿色框
|
||||
label = "Black Bag"
|
||||
|
||||
# 绘制检测框和标签
|
||||
cv2.rectangle(frame, (x1, y1), (x2, y2), color, 2)
|
||||
cv2.putText(frame, label, (x1, y1 - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.6, color, 2)
|
||||
|
||||
# ==========================================
|
||||
# 黑包状态更新
|
||||
# ==========================================
|
||||
if current_blackbag_count > 0:
|
||||
self.blackbag_detection_frames += 1
|
||||
self.blackbag_missing_frames = 0
|
||||
if self.blackbag_detection_frames >= self.frame_thresh_blackbag:
|
||||
self.blackbag_alert_active = True
|
||||
else:
|
||||
self.blackbag_missing_frames += 1
|
||||
if self.blackbag_missing_frames >= self.frame_buffer_blackbag:
|
||||
self.blackbag_detection_frames = 0
|
||||
self.blackbag_alert_active = False
|
||||
|
||||
# ==========================================
|
||||
# 警告信息收集
|
||||
# ==========================================
|
||||
if self.blackbag_alert_active:
|
||||
duration_seconds = self.blackbag_detection_frames / self.fps
|
||||
current_frame_alerts.append(
|
||||
{
|
||||
'time': current_time_sec,
|
||||
'action': 'Black Bag',
|
||||
'details': f"Detected for {duration_seconds:.1f}s"
|
||||
}
|
||||
)
|
||||
self.draw_alert(frame, "Black Bag Alert", (0, 0, 255), sub_text=f"Detected for {duration_seconds:.1f}s")
|
||||
|
||||
# ==========================================
|
||||
# 绘制信息
|
||||
# ==========================================
|
||||
# 实时统计
|
||||
debug_info = f"Person: {self.current_person_count} | BlackBag: {current_blackbag_count}"
|
||||
cv2.putText(frame, debug_info, (20, 40), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (0, 255, 0), 2)
|
||||
|
||||
# 警告信息
|
||||
alert_y_start = 150
|
||||
for i, alert in enumerate(current_frame_alerts):
|
||||
action = alert['action']
|
||||
details = alert.get('details', '')
|
||||
color = (0, 0, 255) # 红色警告
|
||||
main_text = f"{action} ({details})"
|
||||
y_pos = alert_y_start + i * 50
|
||||
cv2.rectangle(frame, (20, y_pos - 40), (900, y_pos + 10), (0, 0, 0), -1)
|
||||
cv2.putText(frame, main_text, (30, y_pos), cv2.FONT_HERSHEY_SIMPLEX, 1.0, color, 2)
|
||||
|
||||
return {
|
||||
"image": frame,
|
||||
|
||||
"alerts":current_frame_alerts
|
||||
}
|
||||
|
||||
|
||||
# ========================= 帧处理线程 =========================
|
||||
class FrameProcessorWorker(BaseFrameProcessorWorker):
|
||||
"""轨迹检测帧处理线程"""
|
||||
|
||||
# 子类配置
|
||||
DETECTOR_FACTORY = lambda params: AbDetector(params)
|
||||
POST_TYPE = 2
|
||||
TARGET_FPS = RTSP_TARGET_FPS
|
||||
@@ -7,6 +7,8 @@ import time
|
||||
import queue
|
||||
import requests
|
||||
|
||||
from biz.base_frame_processor import BaseFrameProcessorWorker
|
||||
|
||||
# -------------------------- Kadian 检测相关导入 --------------------------
|
||||
from algorithm.common.npu_yolo_onnx_person_car_phone import YOLOv8_ONNX # 主检测模型（人/车/后备箱/手机）
|
||||
from common.constants import ALERT_PUSH_URL
|
||||
@@ -19,7 +21,7 @@ from yolox.tracker.byte_tracker import BYTETracker
|
||||
detector_model_path = 'YOLO_Weight/prisoner_model.onnx'
|
||||
|
||||
# 输入尺寸
|
||||
input_size = 1280
|
||||
input_size = 640
|
||||
|
||||
RTSP_TARGET_FPS = 10.0
|
||||
|
||||
@@ -28,9 +30,11 @@ ALERT_PUSH_INTERVAL = 5.0 # 相同action 5秒内仅推送一次
|
||||
|
||||
|
||||
class ZoulangDetector:
|
||||
def __init__(self):
|
||||
# 模型加载
|
||||
def __init__(self, params=None):
|
||||
# 摄像头额外参数
|
||||
self.params = params if params is not None else {}
|
||||
|
||||
# 模型加载
|
||||
self.police_prisoner_detector = YOLOv8_ONNX(detector_model_path, conf_threshold=0.5, iou_threshold=0.45,
|
||||
input_size=input_size)
|
||||
|
||||
@@ -41,8 +45,6 @@ class ZoulangDetector:
|
||||
match_thresh = 0.8
|
||||
mot20 = False
|
||||
|
||||
|
||||
|
||||
self.police_prisoner_track_role = {}
|
||||
|
||||
self.fps = RTSP_TARGET_FPS
|
||||
@@ -334,114 +336,11 @@ class ZoulangDetector:
|
||||
}
|
||||
|
||||
|
||||
|
||||
# ========================= 帧处理线程 =========================
|
||||
class FrameProcessorWorker(threading.Thread):
|
||||
def __init__(self,
|
||||
raw_frame_queue: "queue.Queue[Dict[str, Any]]",
|
||||
ws_send_queue: "queue.Queue[Dict[str, Any]]",
|
||||
stop_event: threading.Event):
|
||||
super().__init__(daemon=True)
|
||||
self.raw_queue = raw_frame_queue
|
||||
self.ws_queue = ws_send_queue
|
||||
self.stop_event = stop_event
|
||||
class FrameProcessorWorker(BaseFrameProcessorWorker):
|
||||
"""轨迹检测帧处理线程"""
|
||||
|
||||
self.last_ts: Dict[int, float] = {}
|
||||
|
||||
# 每个摄像头一个独立的 Kadian 检测器实例
|
||||
self.kadian_detectors: Dict[int, ZoulangDetector] = {}
|
||||
|
||||
# 新增：维护每个摄像头每个action的最后推送时间 {camera_id: {action: last_push_time}}
|
||||
self.last_alert_push_time: Dict[int, Dict[str, float]] = {}
|
||||
|
||||
|
||||
|
||||
def _encode_image_to_base64(self, image) -> str:
|
||||
ok, buf = cv2.imencode(".jpg", image)
|
||||
if not ok:
|
||||
raise RuntimeError("Failed to encode image to JPEG")
|
||||
return base64.b64encode(buf.tobytes()).decode("ascii")
|
||||
|
||||
def run(self):
|
||||
target_interval = 1.0 / RTSP_TARGET_FPS
|
||||
while not self.stop_event.is_set():
|
||||
try:
|
||||
item = self.raw_queue.get(timeout=0.5)
|
||||
except queue.Empty:
|
||||
continue
|
||||
|
||||
cam_id = item["camera_id"]
|
||||
ts = item["timestamp"]
|
||||
frame = item["frame"]
|
||||
|
||||
# 抽帧控制
|
||||
if ts - self.last_ts.get(cam_id, 0) < target_interval:
|
||||
self.raw_queue.task_done()
|
||||
continue
|
||||
self.last_ts[cam_id] = ts
|
||||
|
||||
# 获取检测器实例
|
||||
if cam_id not in self.kadian_detectors:
|
||||
self.kadian_detectors[cam_id] = ZoulangDetector()
|
||||
detector = self.kadian_detectors[cam_id]
|
||||
|
||||
# 执行检测
|
||||
result = detector.process_frame(frame.copy(), cam_id, ts)
|
||||
|
||||
result_img = result["image"]
|
||||
result_type = result["alerts"]
|
||||
|
||||
# ========= 核心修改：过滤5秒内重复的action =========
|
||||
# 初始化当前摄像头的推送时间记录
|
||||
if cam_id not in self.last_alert_push_time:
|
||||
self.last_alert_push_time[cam_id] = {}
|
||||
|
||||
# 筛选出符合推送条件的action（5秒内未推送过）
|
||||
push_actions = []
|
||||
current_time = time.time()
|
||||
for alert in result_type:
|
||||
action = alert['action']
|
||||
last_push = self.last_alert_push_time[cam_id].get(action, 0)
|
||||
# 检查是否超过推送间隔
|
||||
if current_time - last_push >= ALERT_PUSH_INTERVAL:
|
||||
push_actions.append(action)
|
||||
# 更新该action的最后推送时间
|
||||
self.last_alert_push_time[cam_id][action] = current_time
|
||||
|
||||
# 通过 WebSocket 发送帧结果
|
||||
try:
|
||||
img_b64 = self._encode_image_to_base64(result_img)
|
||||
except Exception as e:
|
||||
print(f"[ERROR] Encode image failed: {e}")
|
||||
img_b64 = None
|
||||
|
||||
if img_b64 is not None:
|
||||
# 将abnormal_actions对象数组转换为字符串数组
|
||||
#action_names = [action_info['action'] for action_info in push_actions]
|
||||
|
||||
msg = {
|
||||
"msg_type": "frame",
|
||||
"camera_id": item["camera_index"],
|
||||
"timestamp": ts,
|
||||
#"result_type": action_names,
|
||||
"result_type": push_actions,
|
||||
"image_base64": img_b64,
|
||||
}
|
||||
try:
|
||||
self.ws_queue.put(msg, timeout=1.0)
|
||||
if push_actions and len(push_actions) > 0:
|
||||
# 发送POST请求
|
||||
post_msg = msg.copy()
|
||||
post_msg['type'] = 2
|
||||
try:
|
||||
response = requests.post(ALERT_PUSH_URL, json=post_msg, timeout=5.0)
|
||||
if response.status_code == 200:
|
||||
print(f"[INFO] POST alert sent successfully for actions: {push_actions}")
|
||||
else:
|
||||
print(f"[WARN] POST alert failed with status: {response.status_code}")
|
||||
except Exception as e:
|
||||
print(f"[ERROR] POST alert request failed: {e}")
|
||||
except queue.Full:
|
||||
print("[WARN] ws_send_queue full, drop frame message")
|
||||
|
||||
self.raw_queue.task_done()
|
||||
# 子类配置
|
||||
DETECTOR_FACTORY = lambda params: ZoulangDetector(params)
|
||||
POST_TYPE = 2
|
||||
TARGET_FPS = RTSP_TARGET_FPS
|
||||
|
||||
Reference in New Issue
Block a user