bdzl2/ai_project_v1
6
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

增减数量超限报警功能

Browse Source
This commit is contained in:
martin 2026-04-15 16:12:24 +08:00
parent be99472837
commit 0b283f6b8c
4 changed files with 44 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
middleware/query_model.py
View File

@ -46,6 +46,7 @@ class ModelData:
so_path: str
repeat_dis: float
repeat_time: float
high_count_warn: float
func_description: Optional[str]
filter_indices: List[int]
class_indices: List[int]
@ -250,6 +251,7 @@ class ModelConfigDAO:
aml.py_func,
aml.repeat_dis,
aml.repeat_time,
aml.high_count_warn,
am.scope,
am.yolo_version,
am.PATH,
@ -572,6 +574,7 @@ WHERE
filter_indices=filter_indices,
repeat_dis=repeat_dis,
repeat_time=row.get('repeat_time'),
high_count_warn=row.get('high_count_warn'),
class_indices=row['cls_index'],
conf=conf,
classes=classes,

27
yolo/cv_multi_model_back_video.py
View File

@ -10,6 +10,7 @@ from dataclasses import dataclass
import json
import time
from sympy import false
from ultralytics import YOLO
import torch
@ -2365,7 +2366,7 @@ async def process_frames(detector: MultiYoloTrtDetectorTrackId, cancel_flag: asy
time_pr_start = time.time_ns()
detections, detections_list, model_para = await detector.predict(frame)
time_pr_end = time.time_ns()
print(f"time_pr_starttime_pr_start {(time_pr_end - time_pr_start) / 1000000}")
# print(f"time_pr_starttime_pr_start {(time_pr_end - time_pr_start) / 1000000}")
predict_state = True
if detections:
print("检测到任何目标")
@ -2635,8 +2636,8 @@ async def write_results_to_rtmp(task_id: str, output_url: str = None, input_fps:
# 如果开起侵限功能，就只显示侵限内的框
point_x = (x1 + x2) / 2
point_y = (y1 + y2) / 2
print(f"class_name--{class_name}")
print(f"model_class_names: {model_para[0]['model_class_names']}")
# print(f"class_name--{class_name}")
# print(f"model_class_names: {model_para[0]['model_class_names']}")
if class_name not in model_para[0]["model_class_names"]:
continue
@ -3298,7 +3299,7 @@ invade_cache_lock = Lock() # 用于保护共享变量的锁
async def send_frame_to_s3_mq(loop,upload_executor,task_id, mqtt, mqtt_topic, cancel_flag: asyncio.Event,
cv_frame_queue: asyncio.Queue,
event_queue: asyncio.Queue = None,
device_height: float = float(200), repeat_dis: float = -1, repeat_time: float = -1):
device_height: float = float(200), repeat_dis: float = -1, repeat_time: float = -1,high_count_warn: float = -1):
global stats
start_time = time.time()
# executor = ThreadPoolExecutor(max_workers=Config.MAX_WORKERS)
@ -3383,7 +3384,15 @@ async def send_frame_to_s3_mq(loop,upload_executor,task_id, mqtt, mqtt_topic, ca
should_report = True
print(f"target_pointtarget_point {len(target_point)}")
count_item = 0
des_location_result=[]
des_location_result=[]
high_count_warn_status=False
high_count_warn_num=0
if target_point is not None and 0 < high_count_warn < len(target_point):# 触发计数报警
high_count_warn_num=len(target_point)
high_count_warn_status=True
for item in target_point:
# # 跳过无效的track_id
# 检查是否应该上报该track_id
@ -3560,6 +3569,10 @@ async def send_frame_to_s3_mq(loop,upload_executor,task_id, mqtt, mqtt_topic, ca
"latitude": cam_latitude
},
"count_message":count_message,
"high_count_warn":{
"high_count_warn_status":high_count_warn_status,
"high_count_warn_num":high_count_warn_num
},
"des_location":des_location_result
}
await event_queue.put({
@ -3988,7 +4001,7 @@ async def start_rtmp_processing(video_url: str, task_id: str, model_configs: Lis
mqtt_sub_ip: str, mqtt_sub_port: int, mqtt_sub_topic: str,
output_rtmp_url: str,
invade_enable: bool, invade_file: str, camera_para_url: str,
device_height: float, repeat_dis: float, repeat_time: float):
device_height: float, repeat_dis: float, repeat_time: float,high_count_warn: float):
# 初始化资源
# await initialize_resources()
logger.info(f"拉流地址{video_url}")
@ -4153,7 +4166,7 @@ async def start_rtmp_processing(video_url: str, task_id: str, model_configs: Lis
upload_task = asyncio.create_task(
send_frame_to_s3_mq(loop, upload_executor, task_id, mqtt, mqtt_pub_topic,
cancel_flag, cv_frame_queue, event_queue, device_height, repeat_dis,
repeat_time),
repeat_time,high_count_warn),
name=f"send_frame_to_s3_mq_{_}"
)
upload_tasks.append(upload_task)

34
yolo/detect/multi_yolo_trt_detect_track_trt10_yolo11.py
View File

@ -470,13 +470,13 @@ class YoLo11TRT(object):
# 重塑为二维数组
pred = np.reshape(output[1:], (-1, num_values_per_detection))[:num, :]
# 调试：打印输出信息
if num > 0:
logger.info(f"检测到 {num} 个目标，每个目标 {num_values_per_detection} 个值")
logger.info(f"第一个检测框原始值: center_x={pred[0, 0]:.2f}, center_y={pred[0, 1]:.2f}, "
f"width={pred[0, 2]:.2f}, height={pred[0, 3]:.2f}, conf={pred[0, 4]:.4f}")
logger.info(f"原始图片尺寸: {origin_w}x{origin_h}")
logger.info(f"缩放比例: r_w={self.input_w / origin_w:.4f}, r_h={self.input_h / origin_h:.4f}")
# # 调试：打印输出信息
# if num > 0:
# logger.info(f"检测到 {num} 个目标，每个目标 {num_values_per_detection} 个值")
# logger.info(f"第一个检测框原始值: center_x={pred[0, 0]:.2f}, center_y={pred[0, 1]:.2f}, "
# f"width={pred[0, 2]:.2f}, height={pred[0, 3]:.2f}, conf={pred[0, 4]:.4f}")
# logger.info(f"原始图片尺寸: {origin_w}x{origin_h}")
# logger.info(f"缩放比例: r_w={self.input_w / origin_w:.4f}, r_h={self.input_h / origin_h:.4f}")
# 执行NMS
boxes = self.non_max_suppression(pred, origin_h, origin_w, conf_thres=CONF_THRESH, nms_thres=IOU_THRESHOLD)
@ -521,20 +521,20 @@ class YoLo11TRT(object):
return np.array([])
# 调试：打印转换前的坐标
if len(boxes) > 0:
logger.info(f"NMS前第一个框: center_x={boxes[0, 0]:.2f}, center_y={boxes[0, 1]:.2f}, "
f"width={boxes[0, 2]:.2f}, height={boxes[0, 3]:.2f}, conf={boxes[0, 4]:.4f}")
# if len(boxes) > 0:
# logger.info(f"NMS前第一个框: center_x={boxes[0, 0]:.2f}, center_y={boxes[0, 1]:.2f}, "
# f"width={boxes[0, 2]:.2f}, height={boxes[0, 3]:.2f}, conf={boxes[0, 4]:.4f}")
# 关键修复：使用正确的xywh2xyxy函数
boxes[:, :4] = self.xywh2xyxy(origin_h, origin_w, boxes[:, :4])
# 调试：打印转换后的坐标
if len(boxes) > 0:
logger.info(f"坐标转换后第一个框: x1={boxes[0, 0]:.2f}, y1={boxes[0, 1]:.2f}, "
f"x2={boxes[0, 2]:.2f}, y2={boxes[0, 3]:.2f}")
logger.info(f"框尺寸: {boxes[0, 2] - boxes[0, 0]:.1f}x{boxes[0, 3] - boxes[0, 1]:.1f}")
logger.info(
f"框占图片比例: 宽度={100 * (boxes[0, 2] - boxes[0, 0]) / origin_w:.1f}%, 高度={100 * (boxes[0, 3] - boxes[0, 1]) / origin_h:.1f}%")
# # 调试：打印转换后的坐标
# if len(boxes) > 0:
# logger.info(f"坐标转换后第一个框: x1={boxes[0, 0]:.2f}, y1={boxes[0, 1]:.2f}, "
# f"x2={boxes[0, 2]:.2f}, y2={boxes[0, 3]:.2f}")
# logger.info(f"框尺寸: {boxes[0, 2] - boxes[0, 0]:.1f}x{boxes[0, 3] - boxes[0, 1]:.1f}")
# logger.info(
# f"框占图片比例: 宽度={100 * (boxes[0, 2] - boxes[0, 0]) / origin_w:.1f}%, 高度={100 * (boxes[0, 3] - boxes[0, 1]) / origin_h:.1f}%")
# 裁剪坐标到图像边界内
boxes[:, 0] = np.clip(boxes[:, 0], 0, origin_w - 1)

6
yolo_api.py
View File

@ -664,6 +664,8 @@ async def run_back_Multi_Detect_async(request, request_json, stop_event: asyncio
print("去重的距离:", config.repeat_dis)
repeat_dis = config.repeat_dis
repeat_time = config.repeat_time
high_count_warn = config.high_count_warn
print(f"config.high_count_warn {config.high_count_warn}")
model_configs.append(
{
@ -760,7 +762,7 @@ async def run_back_Multi_Detect_async(request, request_json, stop_event: asyncio
mqtt_sub_ip, mqtt_sub_port, mqtt_sub_topic,
push_url,
invade_enable, invade_file, camera_para_url,
device_height, repeat_dis, repeat_time
device_height, repeat_dis, repeat_time,high_count_warn
)
except Exception as e:
logger.error(f"处理流程异常: {e}")
@ -1846,5 +1848,5 @@ if __name__ == "__main__":
print("正在安装psutil库...")
subprocess.check_call([sys.executable, "-m", "pip", "install", "psutil"])
app.run(host="0.0.0.0", port=12315, debug=False, access_log=True)
app.run(host="192.168.110.103", port=12315, debug=False, access_log=True)
# app.run(host="0.0.0.0", workers=3, port=12315)