orisys-tech/orisys-sdk
3
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fixed new sample cannot detect boxes due to filter

Browse Source
This commit is contained in:
unknown
2025-12-29 15:06:06 +08:00
parent 4f8165b04b
commit 55228413d4
10 changed files with 180 additions and 92 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/orisys/Stitcher.py
+11 -5
View File
@@ -8,9 +8,14 @@ class Stitcher:
def __init__(self,isstitch=False, n=[3,3], index=np.array([1,2,3,6,7,8,11,12,13])):
# Settings
# Settings
self.isstitch = isstitch
# Find Boxes Related
self.min_area_frac = 0.01
self.max_area_frac = 0.05
self.max_aspect_ratio = 100
# stitch related
self.stitch_valid = False
self.boxes = self.units = None
@@ -21,6 +26,7 @@ class Stitcher:
self.nominal_ov_pos=100
self.margin=10
self.bound=self.nominal_ov_pos - self.margin
# for sensor 1
self.pos_default = [(260, 56), (577, 62), (897, 69), (1214, 75), (1530, 81), (256, 391), (570, 397), (889, 403), (1208, 409), (1523, 416), (248, 724), (563, 730), (882, 738), (1201, 743), (1516, 751)]
@@ -542,7 +548,7 @@ class Stitcher:
def __filter_boxes_by_size(self, boxes, img_shape, min_area_frac=0.018, max_area_frac=0.035):
def __filter_boxes_by_size(self, boxes, img_shape):
"""
Remove boxes that are too small or too large relative to the image.
- boxes: list of (x,y,w,h)
@@ -552,12 +558,12 @@ class Stitcher:
"""
h, w = img_shape
img_area = h * w
min_area = img_area * float(min_area_frac)
max_area = img_area * float(max_area_frac)
min_area = img_area * float(self.min_area_frac)
max_area = img_area * float(self.max_area_frac)
filtered = []
for (x,y,ww,hh) in boxes:
area = ww * hh
if area < min_area or area > max_area or (abs(ww - hh) > 20):
if area < min_area or area > max_area or (abs(ww - hh) > self.max_aspect_ratio):
continue
filtered.append((x,y,ww,hh))
return filtered
src/orisys/configs/prototype.json
+13
View File
@@ -9,6 +9,19 @@
"resolution_stitch": [400, 400],
"index": [1,2,3,6,7,8,11,12,13],
"pos_default": [[260, 56], [577, 62], [897, 69], [1214, 75], [1530, 81], [256, 391], [570, 397], [889, 403], [1208, 409], [1523, 416], [248, 724], [563, 730], [882, 738], [1201, 743], [1516, 751]]
},
"stitcher": {
"min_area_frac": 0.018,
"max_area_frac": 0.035,
"max_aspect_ratio": 20
},
"contact": {
"min_contour_area": 3000,
"min_magnitude": 0.02,
"expansion_threshold": 0.01,
"downsample_factor": 1,
"kernel_size": [5,5],
"denoise_sigma": 15
}
}
src/orisys/configs/sample_1.json
+23
View File
@@ -0,0 +1,23 @@
{
"name": "prototype",
"version": "1.0",
"camera": {
"resolution_raw": [1920, 1080],
"fps": 60,
"units": 15,
"units_used": 9,
"resolution_stitch": [400, 400],
"index": [1,2,3,6,7,8,11,12,13],
"pos_default": [[260, 56], [577, 62], [897, 69], [1214, 75], [1530, 81], [256, 391], [570, 397], [889, 403], [1208, 409], [1523, 416], [248, 724], [563, 730], [882, 738], [1201, 743], [1516, 751]]
},
"contact": {
"min_contour_area": 3000,
"min_magnitude": 1.0,
"expansion_threshold": 0.01,
"downsample_factor": 1,
"kernel_size": [5,5],
"denoise_sigma": 15
}
}
src/orisys/get_depth.py
+58 -32
View File
@@ -12,6 +12,41 @@ except ImportError:
# 缓存形态学操作的核，避免重复创建
_morphology_kernels = {}
def _create_foe_result(expansion_mask, contraction_mask, divergence_map,
centroid=None, centroid_found=False, contour=None,
all_contours=None):
"""
创建标准化的FOE结果字典，统一所有函数的返回格式。
Args:
expansion_mask: 扩张区域掩码
contraction_mask: 收缩区域掩码
divergence_map: 散度图
centroid: 质心坐标，可以是单个元组 (x, y) 或列表 [(x1, y1), (x2, y2), ...]
centroid_found: 是否找到质心
contour: 轮廓，可以是单个轮廓或轮廓列表
all_contours: 所有轮廓（仅在return_all_contours=True时使用）
Returns:
dict: 标准化的FOE结果字典
"""
result = {
'expansion_mask': expansion_mask,
'contraction_mask': contraction_mask,
'divergence_map': divergence_map,
'centroid': centroid if centroid is not None else [0,0],
'centroid_found': centroid_found,
}
# 可选字段
if contour is not None:
result['contour'] = contour
if all_contours is not None:
result['_all_contours'] = all_contours
return result
def compute_foe(flow_field, min_magnitude=1.0, expansion_threshold=1, downsample_factor=1, return_all_contours=False, kernel_size=(5,5), denoise_sigma=10):
"""
计算光流场的扩张焦点(Focus of Expansion)。
@@ -73,20 +108,15 @@ def compute_foe(flow_field, min_magnitude=1.0, expansion_threshold=1, downsample
divergence = cv2.GaussianBlur(divergence, (0, 0), denoise_sigma)
# 优化：使用原地操作
np.where(divergence > 0.02, divergence, 0, out=divergence)
np.where(divergence > min_magnitude, divergence, 0, out=divergence)
# 创建掩码（使用原地操作优化）
valid_mask = magnitude > min_magnitude
expansion_mask = (valid_mask & (divergence > expansion_threshold))
contraction_mask = (valid_mask & (divergence < -expansion_threshold))
result = {
'expansion_mask': expansion_mask,
'contraction_mask': contraction_mask,
'divergence_map': divergence,
'foe_points': [0, 0],
'is_foe_found': False
}
# 初始化结果字典
result = _create_foe_result(expansion_mask, contraction_mask, divergence)
# 寻找扩张区域轮廓
# 优化：缓存形态学操作的核
@@ -100,16 +130,16 @@ def compute_foe(flow_field, min_magnitude=1.0, expansion_threshold=1, downsample
contours, _ = cv2.findContours(expansion_binary_closed, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
if contours:
result['is_foe_found'] = True
result['centroid_found'] = True
if return_all_contours:
result['_all_contours'] = contours
largest_contour = max(contours, key=cv2.contourArea)
result['expansion_region'] = largest_contour
result['contour'] = largest_contour
M = cv2.moments(largest_contour)
if M["m00"] > 0:
result['foe_point'] = (int(M["m10"] / M["m00"]), int(M["m01"] / M["m00"]))
result['centroid'] = (int(M["m10"] / M["m00"]), int(M["m01"] / M["m00"]))
return result
@@ -206,7 +236,7 @@ def compute_foe_gpu(flow_field, min_magnitude=1.0, expansion_threshold=1, downsa
# 优化：在GPU上完成掩码操作，减少数据传输
# CuPy的where不支持out参数，使用赋值代替
divergence_gpu = cp.where(divergence_gpu > 0.02, divergence_gpu, 0)
divergence_gpu = cp.where(divergence_gpu > min_magnitude, divergence_gpu, 0)
# 在GPU上创建掩码
valid_mask_gpu = magnitude_gpu > min_magnitude
@@ -219,13 +249,8 @@ def compute_foe_gpu(flow_field, min_magnitude=1.0, expansion_threshold=1, downsa
expansion_mask = cp.asnumpy(expansion_mask_gpu)
contraction_mask = cp.asnumpy(contraction_mask_gpu)
result = {
'expansion_mask': expansion_mask,
'contraction_mask': contraction_mask,
'divergence_map': divergence,
'foe_points': [0, 0],
'is_foe_found': False
}
# 初始化结果字典
result = _create_foe_result(expansion_mask, contraction_mask, divergence)
# 寻找扩张区域轮廓（在CPU上进行，因为OpenCV的轮廓检测需要CPU）
# 优化：缓存形态学操作的核
@@ -239,16 +264,16 @@ def compute_foe_gpu(flow_field, min_magnitude=1.0, expansion_threshold=1, downsa
contours, _ = cv2.findContours(expansion_binary_closed, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
if contours:
result['is_foe_found'] = True
result['centroid_found'] = True
if return_all_contours:
result['_all_contours'] = contours
largest_contour = max(contours, key=cv2.contourArea)
result['expansion_region'] = largest_contour
result['contour'] = largest_contour
M = cv2.moments(largest_contour)
if M["m00"] > 0:
result['foe_point'] = (int(M["m10"] / M["m00"]), int(M["m01"] / M["m00"]))
result['centroid'] = (int(M["m10"] / M["m00"]), int(M["m01"] / M["m00"]))
return result
@@ -275,7 +300,7 @@ def compute_multiple_foe(flow_field, min_magnitude=1.0, expansion_threshold=0.00
- 'divergence_map': 散度图
"""
# 调用 compute_foe 获取基础数据和所有轮廓
foe_result = compute_foe(
foe_result = compute_foe_gpu(
flow_field,
min_magnitude=min_magnitude,
expansion_threshold=expansion_threshold,
@@ -286,13 +311,12 @@ def compute_multiple_foe(flow_field, min_magnitude=1.0, expansion_threshold=0.00
)
# 构建多焦点结果，复用基础数据
result = {
'expansion_mask': foe_result['expansion_mask'],
'contraction_mask': foe_result['contraction_mask'],
'divergence_map': foe_result['divergence_map'],
'foe_points': [],
'expansion_regions': []
}
result = _create_foe_result(
expansion_mask=foe_result['expansion_mask'],
contraction_mask=foe_result['contraction_mask'],
divergence_map=foe_result['divergence_map'],
centroid_found=foe_result['centroid_found']
)
# 处理所有轮廓，过滤并计算焦点
if '_all_contours' in foe_result:
@@ -304,7 +328,9 @@ def compute_multiple_foe(flow_field, min_magnitude=1.0, expansion_threshold=0.00
if M["m00"] > 0:
cx = int(M["m10"] / M["m00"])
cy = int(M["m01"] / M["m00"])
result['foe_points'].append((cx, cy))
result['expansion_regions'].append(contour)
result['centroid'].append((cx, cy))
if 'contour' not in result:
result['contour'] = []
result['contour'].append(contour)
return result
src/orisys/sensor.py
+53 -21
View File
@@ -45,6 +45,8 @@ class Sensor:
FNORMAL = auto()
FSHEARX = auto()
FSHEARY = auto()
CENTROID = auto()
CONTOUR = auto()
DEPTH = auto()
# =========================================================================
@@ -91,15 +93,22 @@ class Sensor:
self.vupdate = False
self.stitch_calibration = []
# Data variables
# Data Variables
self.img = None
self.img_size_raw = None
self.img_filtered = None
self.flow = None
self.vnormal = None
self.shear = None
self.depth = None
self.vshear = None
# self.depth = None
self.n_boxes = None
self.fnormal = self.fshearx = self.fsheary = self.ftor = 0.0
# Contact Data Variables
self.centroid = None
self.centroid_found = False
self.contour = None
self.depth_map = None
# Timing
self.last_log_time = time.perf_counter()
@@ -116,7 +125,11 @@ class Sensor:
self._stop_event = None
# Load configuration
self.config_cam = None
self.config_stitcher = None
self.config_contact = None
self._load_config()
# Initialize components
self._stitcher = Stitcher(isstitch=isstitch, index=self.config_index)
@@ -167,7 +180,8 @@ class Sensor:
if log_timing: t0 = time.perf_counter()
img = np.ascontiguousarray(self._stitcher.stitch_rt_optimized())
if log_timing: timings['stitch'] = (time.perf_counter() - t0) * 1000
self.img_size_raw = img.shape
# Resize
img = cv2.resize(img, tuple(self.config_resolution_stitch), interpolation=cv2.INTER_CUBIC)
@@ -222,7 +236,7 @@ class Sensor:
# Get components
self.vnormal = self.hhd.get_normal()
self.shear = self.hhd.get_shear()
self.vshear = self.hhd.get_shear()
# # Depth computation
# if log_timing: t0 = time.perf_counter()
@@ -244,21 +258,31 @@ class Sensor:
return self.flow
def get_contact(self, multiple_foe=False, min_contour_area=3000, min_magnitude=1.0, expansion_threshold=0.01, downsample_factor=1, kernel_size=(5,5), denoise_sigma=15):
def compute_contact(self, multiple_foe=False, min_contour_area=3000, min_magnitude=1.0, expansion_threshold=0.01, downsample_factor=1, kernel_size=(5,5), denoise_sigma=15):
if multiple_foe:
return compute_multiple_foe(self.vnormal, min_contour_area=min_contour_area, min_magnitude=min_magnitude, expansion_threshold=expansion_threshold, downsample_factor=downsample_factor, kernel_size=kernel_size, denoise_sigma=denoise_sigma)
else:
if self.cuda:
return compute_foe_gpu(self.vnormal, min_magnitude=min_magnitude, expansion_threshold=expansion_threshold, downsample_factor=downsample_factor, kernel_size=kernel_size, denoise_sigma=denoise_sigma)
else:
return compute_foe(self.vnormal, min_magnitude=min_magnitude, expansion_threshold=expansion_threshold, downsample_factor=downsample_factor, kernel_size=kernel_size, denoise_sigma=denoise_sigma)
result = compute_multiple_foe(self.vnormal, min_contour_area=min_contour_area, min_magnitude=min_magnitude, expansion_threshold=expansion_threshold, downsample_factor=downsample_factor, kernel_size=kernel_size, denoise_sigma=denoise_sigma)
else:
if self.cuda:
result = compute_foe_gpu(self.vnormal, min_magnitude=min_magnitude, expansion_threshold=expansion_threshold, downsample_factor=downsample_factor, kernel_size=kernel_size, denoise_sigma=denoise_sigma)
else:
result = compute_foe(self.vnormal, min_magnitude=min_magnitude, expansion_threshold=expansion_threshold, downsample_factor=downsample_factor, kernel_size=kernel_size, denoise_sigma=denoise_sigma)
self.centroid = result['centroid']
self.centroid_found = result['centroid_found']
if 'contour' in result:
self.contour = result['contour']
self.depth_map = result['divergence_map']
return result
# def get_depth(self, denoise_sigma=15):
# return getFilteredDensity(self.vnormal, use_cuda=self.cuda)
# self.depth = getFilteredDensity(self.vnormal, use_cuda=self.cuda)
def get_depth(self, denoise_sigma=15):
return getFilteredDensity(self.vnormal, use_cuda=self.cuda)
self.depth = getFilteredDensity(self.vnormal, use_cuda=self.cuda)
def update(self):
"""
Convenience method: get_img + compute_deformation in one call.
@@ -270,7 +294,7 @@ class Sensor:
self.compute_deformation()
return {
'img': self.img, 'flow': self.flow,
'normal': self.vnormal, 'shear': self.shear, 'depth': self.depth,
'normal': self.vnormal, 'shear': self.vshear, 'depth': self.depth,
'fps': self.fps, 'fnormal': self.fnormal,
'fshearx': self.fshearx, 'fsheary': self.fsheary, 'ftor': self.ftor,
}
@@ -291,13 +315,15 @@ class Sensor:
mapping = {
self.info.VRAW: lambda: self.flow if self.flow is not None else np.array([]),
self.info.VNORMAL: lambda: self.vnormal if self.vnormal is not None else np.array([]),
self.info.VSHEAR: lambda: self.shear if self.shear is not None else np.array([]),
self.info.VSHEAR: lambda: self.vshear if self.vshear is not None else np.array([]),
self.info.FNORMAL: lambda: self.fnormal,
self.info.FSHEARX: lambda: self.fshearx,
self.info.FSHEARY: lambda: self.fsheary,
self.info.FPS: lambda: self.fps,
self.info.TIMESTAMP: lambda: time.strftime('%H:%M:%S'),
self.info.DEPTH: lambda: self.depth if self.depth is not None else np.array([]),
self.info.DEPTH: lambda: self.depth_map if self.depth_map is not None else np.array([]),
self.info.CENTROID: lambda: self.centroid if self.centroid is not None else np.array([]),
self.info.CONTOUR: lambda: self.contour if self.contour is not None else np.array([]),
}
return tuple(mapping.get(arg, lambda: None)() for arg in args if isinstance(arg, self.info))
@@ -431,6 +457,12 @@ class Sensor:
self.config_index = cam["index"]
self.config_pos_default = cam["pos_default"]
config_stitcher = self.config["stitcher"]
config_contact = self.config["contact"]
if self.verbose:
print(f"Config: {self.config_name}")
print(f" Resolution: {self.config_resolution_raw} → {self.config_resolution_stitch}")
@@ -579,7 +611,7 @@ class Sensor:
while not self._stop_event.is_set():
try:
img, flow = self._flow_queue.get(timeout=0.1)
contact_result = self.get_contact(flow)
contact_result = self.compute_contact(flow)
self._queue_put(self._result_queue, contact_result)
except Empty:
continue
src/orisys/util.py
+11 -13
View File
@@ -132,37 +132,35 @@ def draw_fps(image, fps):
return image
def draw_foe(image, foe_result, point_color=(0, 255, 0), region_color=(0, 255, 255)):
def draw_contact(image, contour, centroid, point_color=(0, 255, 0), region_color=(0, 255, 255)):
"""
在图像上绘制扩张焦点和扩张区域。
Args:
image: 背景图像
foe_result: compute_foe 函数的返回结果
point_color: 焦点点的颜色 (BGR)
region_color: 扩张区域轮廓的颜色 (BGR)
contour: 轮廓
centroid: 触碰点
point_color: 触碰点的颜色 (BGR)
region_color: 轮廓的颜色 (BGR)
Returns:
np.ndarray: 绘制了焦点和区域的图像
np.ndarray: 绘制了触碰点和轮廓的图像
"""
image = image.copy()
# 绘制扩张区域轮廓
if 'expansion_region' in foe_result:
cv2.drawContours(image, [foe_result['expansion_region']], -1, region_color, 2)
if contour is not None:
cv2.drawContours(image, [contour], -1, region_color, 2)
# 绘制焦点点
if 'foe_point' in foe_result:
cx, cy = foe_result['foe_point']
# 绘制触碰点
if centroid is not None:
cx, cy = centroid
cv2.circle(image, (cx, cy), 8, point_color, -1)
cv2.circle(image, (cx, cy), 10, point_color, 2)
cv2.putText(image, "FOE", (cx + 15, cy),
cv2.FONT_HERSHEY_SIMPLEX, 0.6, point_color, 2)
return image
def draw_multiple_foe(image, foe_result, point_color=(0, 255, 0),
region_color=(0, 255, 255), label_colors=None):
"""