Skip to content

Add object detection, tracking, mobile notifications, and search to any security camera.

clean

roryclear80b45621736b
diff --git a/detection/yolov9.py b/detection/yolov9.py
index ea88945..623cf02 100644
--- a/detection/yolov9.py
+++ b/detection/yolov9.py
@@ -462,72 +462,6 @@ def postprocess(output, max_det=300, conf_threshold=0.25, iou_threshold=0.45):
no_overlap_mask = high_iou_mask.sum(axis=1) == 0
return boxes * no_overlap_mask.unsqueeze(-1)
-def draw_bounding_boxes(orig_img_path, predictions, class_labels):
- color_dict = {
- label: tuple((((i+1) * 50) % 256, ((i+1) * 100) % 256, ((i+1) * 150) % 256))
- for i, label in enumerate(class_labels)
- }
- font = cv2.FONT_HERSHEY_SIMPLEX
-
- def is_bright_color(color):
- r, g, b = color
- brightness = (r * 299 + g * 587 + b * 114) / 1000
- return brightness > 127
-
- orig_img = (
- cv2.imread(orig_img_path)
- if not isinstance(orig_img_path, np.ndarray)
- else cv2.imdecode(orig_img_path, 1)
- )
-
- height, width, _ = orig_img.shape
- box_thickness = int((height + width) / 400)
- font_scale = (height + width) / 2500
- object_count = defaultdict(int)
-
- for pred in predictions:
- if len(pred) == 7: # todo
- x1, y1, x2, y2, conf, class_id, _ = pred
- else:
- x1, y1, x2, y2, conf, class_id = pred
- if conf == 0:
- continue
-
- x1, y1, x2, y2, class_id = map(int, (x1, y1, x2, y2, class_id))
- color = color_dict[class_labels[class_id]]
-
- cv2.rectangle(orig_img, (x1, y1), (x2, y2), color, box_thickness)
-
- label = f"{class_labels[class_id]} {conf:.2f}"
- text_size, _ = cv2.getTextSize(label, font, font_scale, 1)
- label_y, bg_y = (
- (y1 - 4, y1 - text_size[1] - 4)
- if y1 - text_size[1] - 4 > 0
- else (y1 + text_size[1], y1)
- )
-
- cv2.rectangle(
- orig_img,
- (x1, bg_y),
- (x1 + text_size[0], bg_y + text_size[1]),
- color,
- -1,
- )
-
- font_color = (0, 0, 0) if is_bright_color(color) else (255, 255, 255)
- cv2.putText(
- orig_img,
- label,
- (x1, label_y),
- font,
- font_scale,
- font_color,
- 1,
- cv2.LINE_AA,
- )
-
- object_count[class_labels[class_id]] += 1
- return orig_img
SIZES = {"t": [16, 64, 96, 24, 128, 256, 224, 160, 48, 144, 192, 80, 32, 16, 3, 96, 32, 64, 128, 64, 64, 128,"t"],
"s": [32, 128, 192, 48, 256, 512, 448, 320, 96, 288, 384, 128, 64, 32, 3, 192, 64, 64, 128, 128, 128, 256, "s"],
diff --git a/test/test_mot.py b/test/test_mot.py
index 5af1d69..85e4004 100644
--- a/test/test_mot.py
+++ b/test/test_mot.py
@@ -1,6 +1,7 @@
from detection.yolov9 import YOLOv9, safe_load, load_state_dict, Sequential, Silence, Conv, RepNCSPELAN4, AConv,\
-ADown, CBLinear, CBFuse, SPPELAN, Upsample, Concat, DDetect, postprocess, fetch, draw_bounding_boxes
+ADown, CBLinear, CBFuse, SPPELAN, Upsample, Concat, DDetect, postprocess, fetch
from detection.rfdetr import RFDETR, detr_to_yolo
+from utils.helpers import draw_bounding_boxes
import cv2
from tinygrad import Tensor, TinyJit
from tinygrad.dtype import dtypes
diff --git a/utils/helpers.py b/utils/helpers.py
index 2a502a5..4c6569b 100644
--- a/utils/helpers.py
+++ b/utils/helpers.py
@@ -2,7 +2,7 @@ import uuid
import http
from pathlib import Path
import shutil
-from collections import deque
+from collections import deque, defaultdict
import time
import os
import subprocess
@@ -11,6 +11,8 @@ import json
import struct
from datetime import datetime
import threading
+import cv2
+import numpy as np
BASE_DIR = Path(__file__).parent.parent / "data"
def send_notif(session_token: str, text=None):
@@ -47,6 +49,73 @@ def send_notif(session_token: str, text=None):
conn.close()
+def draw_bounding_boxes(orig_img_path, predictions, class_labels):
+ color_dict = {
+ label: tuple((((i+1) * 50) % 256, ((i+1) * 100) % 256, ((i+1) * 150) % 256))
+ for i, label in enumerate(class_labels)
+ }
+ font = cv2.FONT_HERSHEY_SIMPLEX
+
+ def is_bright_color(color):
+ r, g, b = color
+ brightness = (r * 299 + g * 587 + b * 114) / 1000
+ return brightness > 127
+
+ orig_img = (
+ cv2.imread(orig_img_path)
+ if not isinstance(orig_img_path, np.ndarray)
+ else cv2.imdecode(orig_img_path, 1)
+ )
+
+ height, width, _ = orig_img.shape
+ box_thickness = int((height + width) / 400)
+ font_scale = (height + width) / 2500
+ object_count = defaultdict(int)
+
+ for pred in predictions:
+ if len(pred) == 7: # todo
+ x1, y1, x2, y2, conf, class_id, _ = pred
+ else:
+ x1, y1, x2, y2, conf, class_id = pred
+ if conf == 0:
+ continue
+
+ x1, y1, x2, y2, class_id = map(int, (x1, y1, x2, y2, class_id))
+ color = color_dict[class_labels[class_id]]
+
+ cv2.rectangle(orig_img, (x1, y1), (x2, y2), color, box_thickness)
+
+ label = f"{class_labels[class_id]} {conf:.2f}"
+ text_size, _ = cv2.getTextSize(label, font, font_scale, 1)
+ label_y, bg_y = (
+ (y1 - 4, y1 - text_size[1] - 4)
+ if y1 - text_size[1] - 4 > 0
+ else (y1 + text_size[1], y1)
+ )
+
+ cv2.rectangle(
+ orig_img,
+ (x1, bg_y),
+ (x1 + text_size[0], bg_y + text_size[1]),
+ color,
+ -1,
+ )
+
+ font_color = (0, 0, 0) if is_bright_color(color) else (255, 255, 255)
+ cv2.putText(
+ orig_img,
+ label,
+ (x1, label_y),
+ font,
+ font_scale,
+ font_color,
+ 1,
+ cv2.LINE_AA,
+ )
+
+ object_count[class_labels[class_id]] += 1
+ return orig_img
+
def export_clip(stream_dir, output_path: Path, live=False, length=5, end=0, start=None):
segments = sorted(stream_dir.glob("*.ts"), key=os.path.getmtime)
recent_segments = deque()