Spaces:
Sleeping
Sleeping
File size: 13,076 Bytes
98a79a7 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 | """
Crowd Density Heatmap Generator
Implements REQ-4, REQ-5: Generate and visualize crowd density zones
"""
import cv2
import numpy as np
from typing import List, Dict, Tuple
import time
import logging
logger = logging.getLogger(__name__)
class HeatmapGenerator:
"""
Generate crowd density heatmaps
Satisfies SRS Requirements:
- REQ-4: Generate localized density zones
- REQ-5: Apply color map representing crowd concentration
"""
def __init__(self, config: Dict):
"""
Initialize heatmap generator
Args:
config: Configuration dictionary
"""
self.config = config
self.enabled = config['heatmap']['enabled']
self.kernel_size = config['heatmap']['kernel_size']
self.alpha = config['heatmap']['alpha']
self.colormap_name = config['heatmap']['colormap']
# Adaptive heatmap settings
self.adaptive = config['heatmap'].get('adaptive', True)
self.min_kernel_size = config['heatmap'].get('min_kernel_size', 30)
self.max_kernel_size = config['heatmap'].get('max_kernel_size', 150)
self.blur_strength = config['heatmap'].get('blur_strength', 0.6)
# Map colormap name to OpenCV constant
colormap_dict = {
'jet': cv2.COLORMAP_JET,
'hot': cv2.COLORMAP_HOT,
'viridis': cv2.COLORMAP_VIRIDIS,
'plasma': cv2.COLORMAP_PLASMA,
'rainbow': cv2.COLORMAP_RAINBOW,
'cool': cv2.COLORMAP_COOL
}
self.colormap = colormap_dict.get(self.colormap_name, cv2.COLORMAP_JET)
# Performance tracking
self.generation_times = []
logger.info(f"Heatmap Generator initialized: Enabled={self.enabled}")
logger.info(f"Kernel size: {self.kernel_size}, Alpha: {self.alpha}, Colormap: {self.colormap_name}")
logger.info(f"Adaptive mode: {self.adaptive}, Range: {self.min_kernel_size}-{self.max_kernel_size}")
def generate_heatmap(self, frame: np.ndarray, detections: List[Dict]) -> Tuple[np.ndarray, float]:
"""
Generate crowd density heatmap with ADAPTIVE kernel sizing
Args:
frame: Input frame
detections: List of detections with center points and bbox
Returns:
heatmap_overlay: Frame with heatmap overlay
generation_time: Time taken to generate heatmap
Implements:
- REQ-4: Generate localized density zones
- REQ-5: Apply color map for crowd concentration
- ADAPTIVE: Auto-adjusts kernel size based on detection box dimensions
"""
start_time = time.time()
# Validate inputs
if frame is None or frame.size == 0:
logger.error("Invalid frame provided to heatmap generator")
return np.zeros((480, 640, 3), dtype=np.uint8), 0.0
# Only check if there are detections
if not detections or len(detections) == 0:
generation_time = time.time() - start_time
return frame.copy(), generation_time
try:
h, w = frame.shape[:2]
# Validate frame dimensions
if h <= 0 or w <= 0:
logger.error(f"Invalid frame dimensions: {h}x{w}")
return frame.copy(), 0.0
# Create empty density map (REQ-4: localized density zones)
density_map = np.zeros((h, w), dtype=np.float32)
# Calculate adaptive kernel size with validation
if self.adaptive and len(detections) > 0:
total_size = 0
valid_detections = 0
for det in detections:
try:
bbox = det.get('bbox', [])
if len(bbox) != 4:
continue
x1, y1, x2, y2 = bbox
box_width = max(0, x2 - x1)
box_height = max(0, y2 - y1)
if box_width > 0 and box_height > 0:
total_size += (box_width + box_height) / 2
valid_detections += 1
except (KeyError, TypeError, ValueError) as e:
logger.debug(f"Skipping invalid detection: {e}")
continue
if valid_detections > 0:
avg_box_size = total_size / valid_detections
# Scale kernel size based on average object size
kernel_radius = int(np.clip(avg_box_size * 0.8,
self.min_kernel_size,
self.max_kernel_size))
kernel_radius = max(15, kernel_radius)
logger.debug(f"Adaptive kernel: avg={avg_box_size:.1f}, radius={kernel_radius}")
else:
kernel_radius = self.kernel_size
else:
kernel_radius = self.kernel_size
# Add Gaussian blobs at each detection center with validation
for det in detections:
try:
center = det.get('center', [])
bbox = det.get('bbox', [])
if len(center) != 2 or len(bbox) != 4:
continue
cx, cy = center
# Validate center coordinates
if not (0 <= cx < w and 0 <= cy < h):
logger.debug(f"Skipping out-of-bounds detection at ({cx}, {cy})")
continue
# Get detection-specific size for better adaptation
if self.adaptive:
x1, y1, x2, y2 = bbox
det_width = max(0, x2 - x1)
det_height = max(0, y2 - y1)
det_size = (det_width + det_height) / 2
if det_size <= 0:
det_kernel = kernel_radius
else:
det_kernel = int(np.clip(det_size * 0.8,
self.min_kernel_size,
self.max_kernel_size))
det_kernel = max(15, det_kernel)
else:
det_kernel = kernel_radius
# Calculate ROI bounds with proper clamping
y_min = max(0, cy - det_kernel)
y_max = min(h, cy + det_kernel)
x_min = max(0, cx - det_kernel)
x_max = min(w, cx + det_kernel)
# Validate ROI dimensions
kernel_height = y_max - y_min
kernel_width = x_max - x_min
if kernel_height <= 0 or kernel_width <= 0:
continue
# Create 2D Gaussian with bounds checking
y_range = np.arange(y_min, y_max) - cy
x_range = np.arange(x_min, x_max) - cx
if len(y_range) == 0 or len(x_range) == 0:
continue
x_grid, y_grid = np.meshgrid(x_range, y_range)
# Gaussian formula with adaptive sigma
det_sigma = det_kernel * self.blur_strength
gaussian = np.exp(-(x_grid**2 + y_grid**2) / (2 * det_sigma**2))
# Use confidence as intensity multiplier for better visualization
intensity = det.get('confidence', 1.0)
# Add to density map with bounds safety
try:
density_map[y_min:y_max, x_min:x_max] += gaussian.astype(np.float32) * intensity
except (ValueError, IndexError) as e:
logger.debug(f"Skipping gaussian placement: {e}")
continue
except (KeyError, TypeError, ValueError, IndexError) as e:
logger.debug(f"Error processing detection for heatmap: {e}")
continue
# Normalize density map to 0-255
if density_map.max() > 0:
density_map = (density_map / density_map.max() * 255).astype(np.uint8)
else:
density_map = density_map.astype(np.uint8)
# Apply single Gaussian blur for smooth appearance (removed double blur)
blur_size = max(11, min(21, kernel_radius // 4)) # Adaptive blur size
if blur_size % 2 == 0:
blur_size += 1 # Must be odd
density_map = cv2.GaussianBlur(density_map, (blur_size, blur_size), 0)
# Apply colormap (REQ-5: color map representing concentration)
heatmap_colored = cv2.applyColorMap(density_map, self.colormap)
# Overlay heatmap on original frame
heatmap_overlay = cv2.addWeighted(
frame,
1 - self.alpha,
heatmap_colored,
self.alpha,
0
)
generation_time = time.time() - start_time
self.generation_times.append(generation_time)
# Check performance constraint (SRS: ≤ 1.5s per frame)
if generation_time > self.config['constraints']['max_heatmap_delay']:
logger.warning(
f"Heatmap generation exceeded constraint: "
f"{generation_time:.3f}s > {self.config['constraints']['max_heatmap_delay']}s"
)
return heatmap_overlay, generation_time
except Exception as e:
logger.error(f"Heatmap generation error: {e}")
import traceback
logger.error(traceback.format_exc())
return frame.copy(), time.time() - start_time
def generate_density_grid(self, frame: np.ndarray, detections: List[Dict],
grid_size: int = 50) -> np.ndarray:
"""
Generate grid-based density visualization (alternative method)
Args:
frame: Input frame
detections: List of detections
grid_size: Size of each grid cell
Returns:
grid_overlay: Frame with grid density overlay
"""
h, w = frame.shape[:2]
overlay = frame.copy()
# Create grid
grid_h = h // grid_size + 1
grid_w = w // grid_size + 1
density_grid = np.zeros((grid_h, grid_w), dtype=int)
# Count detections in each grid cell
for det in detections:
cx, cy = det['center']
grid_x = min(cx // grid_size, grid_w - 1)
grid_y = min(cy // grid_size, grid_h - 1)
density_grid[grid_y, grid_x] += 1
# Draw grid with color intensity based on density
max_density = density_grid.max() if density_grid.max() > 0 else 1
for gy in range(grid_h):
for gx in range(grid_w):
if density_grid[gy, gx] > 0:
x1 = gx * grid_size
y1 = gy * grid_size
x2 = min(x1 + grid_size, w)
y2 = min(y1 + grid_size, h)
# Color intensity based on density
intensity = int(255 * (density_grid[gy, gx] / max_density))
color = (0, intensity, 255 - intensity) # Blue to red
# Draw semi-transparent rectangle
sub_img = overlay[y1:y2, x1:x2]
rect = np.full_like(sub_img, color, dtype=np.uint8)
overlay[y1:y2, x1:x2] = cv2.addWeighted(sub_img, 0.7, rect, 0.3, 0)
return overlay
def get_statistics(self) -> Dict:
"""Get heatmap generation statistics"""
if not self.generation_times:
return {
'avg_generation_time': 0.0,
'total_heatmaps': 0
}
return {
'avg_generation_time': np.mean(self.generation_times[-100:]),
'total_heatmaps': len(self.generation_times),
'max_generation_time': max(self.generation_times),
'min_generation_time': min(self.generation_times)
}
|