File size: 8,761 Bytes
8387086 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 | """
Phase 2 - YOLOv8 Detection + Bounding Box Visualisation
Runs real YOLOv8 detections on warehouse images, draws annotated
bounding boxes, saves results, and optionally uploads to GCS/S3.
Run:
python phase2_detect.py --input data/sample_images/
python phase2_detect.py --input data/sample_images/warehouse_01.jpg
python phase2_detect.py --input data/sample_images/ --upload gcs
python phase2_detect.py --input data/sample_images/ --upload aws
"""
import argparse
import json
from pathlib import Path
import numpy as np
import cv2
from loguru import logger
from datetime import datetime
# βββ Config βββββββββββββββββββββββββββββββββββββββββββββββββββββββ
OUTPUT_DIR = Path("output/annotated")
REPORT_DIR = Path("output/reports")
MODEL_NAME = "yolov8n.pt" # nano = fastest on CPU; swap to yolov8s.pt for better accuracy
CONF_THRESHOLD = 0.35
IMAGE_EXTENSIONS = {".jpg", ".jpeg", ".png", ".bmp", ".webp"}
# Warehouse-relevant COCO classes (YOLOv8 default model)
WAREHOUSE_LABELS = {
"person": ("worker", (0, 200, 0)), # green
"truck": ("vehicle", (0, 100, 255)), # orange
"car": ("vehicle", (0, 100, 255)),
"motorcycle":("vehicle", (0, 100, 255)),
"bicycle": ("vehicle", (0, 100, 255)),
"backpack": ("parcel", (255, 200, 0)), # cyan
"suitcase": ("parcel", (255, 200, 0)),
"bottle": ("item", (200, 200, 200)),
"chair": ("obstacle", (0, 0, 220)), # red
"couch": ("obstacle", (0, 0, 220)),
"box": ("pallet", (255, 100, 0)),
"laptop": ("equipment",(180, 0, 180)),
"tv": ("equipment",(180, 0, 180)),
}
def load_model():
"""Load YOLOv8 model (downloads automatically on first run ~6MB)."""
from ultralytics import YOLO
logger.info(f"Loading YOLOv8 model: {MODEL_NAME}")
model = YOLO(MODEL_NAME)
logger.success("Model loaded β")
return model
def detect_image(model, image_path: Path) -> dict:
"""
Run YOLOv8 on a single image.
Returns:
dict with image array, detections list, and metadata
"""
image = cv2.imread(str(image_path))
if image is None:
logger.warning(f"Could not load: {image_path}")
return None
h, w = image.shape[:2]
results = model(image, conf=CONF_THRESHOLD, verbose=False)
detections = []
for result in results:
for box in result.boxes:
cls_name = result.names[int(box.cls)]
label_info = WAREHOUSE_LABELS.get(cls_name)
warehouse_label = label_info[0] if label_info else cls_name
colour = label_info[1] if label_info else (200, 200, 200)
conf = float(box.conf)
x1, y1, x2, y2 = [int(v) for v in box.xyxy[0].tolist()]
detections.append({
"original_class": cls_name,
"warehouse_label": warehouse_label,
"confidence": round(conf, 3),
"bbox": [x1, y1, x2, y2],
"colour": colour,
})
logger.info(f" {image_path.name}: {len(detections)} object(s) detected")
return {
"image": image,
"path": image_path,
"detections": detections,
"image_size": (w, h),
}
def draw_annotations(result: dict) -> np.ndarray:
"""
Draw bounding boxes and labels on the image.
Returns annotated image as np.ndarray.
"""
image = result["image"].copy()
for det in result["detections"]:
x1, y1, x2, y2 = det["bbox"]
colour = det["colour"]
label = f"{det['warehouse_label']} {det['confidence']:.0%}"
# Bounding box
cv2.rectangle(image, (x1, y1), (x2, y2), colour, 2)
# Label background
(tw, th), _ = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX, 0.55, 1)
cv2.rectangle(image, (x1, y1 - th - 8), (x1 + tw + 4, y1), colour, -1)
# Label text
cv2.putText(
image, label,
(x1 + 2, y1 - 4),
cv2.FONT_HERSHEY_SIMPLEX, 0.55,
(255, 255, 255), 1, cv2.LINE_AA,
)
# Summary overlay (top-left)
summary = f"Objects: {len(result['detections'])} | Model: {MODEL_NAME}"
cv2.putText(image, summary, (10, 28),
cv2.FONT_HERSHEY_SIMPLEX, 0.7, (0, 0, 0), 3, cv2.LINE_AA)
cv2.putText(image, summary, (10, 28),
cv2.FONT_HERSHEY_SIMPLEX, 0.7, (255, 255, 255), 1, cv2.LINE_AA)
return image
def save_annotated(annotated: np.ndarray, original_path: Path) -> Path:
"""Save annotated image to output directory."""
OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
out_path = OUTPUT_DIR / f"annotated_{original_path.name}"
cv2.imwrite(str(out_path), annotated)
logger.success(f" Saved annotated: {out_path}")
return out_path
def save_report(results: list, output_path: Path) -> Path:
"""Save JSON detection report."""
REPORT_DIR.mkdir(parents=True, exist_ok=True)
report = {
"generated_at": datetime.utcnow().isoformat(),
"model": MODEL_NAME,
"confidence_threshold": CONF_THRESHOLD,
"images_processed": len(results),
"total_detections": sum(len(r["detections"]) for r in results),
"results": [
{
"image": r["path"].name,
"image_size": r["image_size"],
"detections": [
{k: v for k, v in d.items() if k != "colour"}
for d in r["detections"]
],
}
for r in results
],
}
with open(output_path, "w") as f:
json.dump(report, f, indent=2)
logger.success(f"Report saved: {output_path}")
return output_path
def upload_to_cloud(files: list[Path], provider: str):
"""Upload annotated images and report to GCS or S3."""
if provider == "gcs":
from cloud_infra.setup_gcs import upload_image
for f in files:
upload_image(f, gcs_folder="output/phase2/")
elif provider == "aws":
from cloud_infra.setup_aws import upload_image
for f in files:
upload_image(f, s3_folder="output/phase2/")
logger.success(f"Uploaded {len(files)} file(s) to {provider.upper()}")
def run_pipeline(input_path: Path, upload: str = None):
"""Main Phase 2 pipeline."""
# 1. Collect image paths
if input_path.is_file():
image_paths = [input_path]
elif input_path.is_dir():
image_paths = sorted([
p for p in input_path.iterdir()
if p.suffix.lower() in IMAGE_EXTENSIONS
])
else:
logger.error(f"Input not found: {input_path}")
return
if not image_paths:
logger.error("No images found. Run: python data/download_samples.py first")
return
logger.info(f"Processing {len(image_paths)} image(s) from {input_path}")
# 2. Load model
model = load_model()
# 3. Detect + annotate
all_results = []
annotated_paths = []
for img_path in image_paths:
result = detect_image(model, img_path)
if result is None:
continue
annotated = draw_annotations(result)
out_path = save_annotated(annotated, img_path)
all_results.append(result)
annotated_paths.append(out_path)
# 4. Save report
report_path = REPORT_DIR / f"phase2_report_{datetime.now().strftime('%Y%m%d_%H%M%S')}.json"
save_report(all_results, report_path)
annotated_paths.append(report_path)
# 5. Summary
total = sum(len(r["detections"]) for r in all_results)
logger.success(f"\n{'='*50}")
logger.success(f" Phase 2 Complete!")
logger.success(f" Images processed : {len(all_results)}")
logger.success(f" Total detections : {total}")
logger.success(f" Annotated images : {OUTPUT_DIR}/")
logger.success(f" Report : {report_path}")
logger.success(f"{'='*50}")
# 6. Optional cloud upload
if upload:
upload_to_cloud(annotated_paths, upload)
return all_results
if __name__ == "__main__":
parser = argparse.ArgumentParser(description="Phase 2: YOLOv8 Warehouse Detection")
parser.add_argument("--input", type=str, default="data/sample_images/", help="Image or folder")
parser.add_argument("--upload", type=str, choices=["gcs", "aws"], default=None, help="Upload to cloud")
parser.add_argument("--conf", type=float, default=CONF_THRESHOLD, help="Confidence threshold")
args = parser.parse_args()
CONF_THRESHOLD = args.conf
run_pipeline(Path(args.input), upload=args.upload)
|