def compute_MCD_weft(weftsPickled, target_path):
weft_points_list = floatPointList()
for pickled_path in weftsPickled:
weft_points_list.extend(pickle.load(open(pickled_path, "rb" )))
x_vals = [fp.x for fp in weft_points_list]
y_vals = [fp.y for fp in weft_points_list]
mean_hor_dist = weft_points_list.getMedianWeftDist()
min_x = min(x_vals) + 1.5 * mean_hor_dist
max_x = max(x_vals) - 1.5 * mean_hor_dist
min_y = min(y_vals) + 1.5 * mean_hor_dist
max_y = max(y_vals) - 1.5 * mean_hor_dist
inner_points = floatPointList()
for pt in weft_points_list:
if min_x < pt.x < max_x and min_y < pt.y < max_y:
inner_points.append(pt)
X = np.zeros([len(inner_points), 3])
for idx, pt in enumerate(inner_points):
X[idx,0] = pt.area
X[idx,1] = pt.right_dist
X[idx,2] = pt.left_dist
Y = X[~(X<=0).any(axis=1)]
robust_cov = MinCovDet(support_fraction=0.8).fit(Y)
pickle.dump(robust_cov, open(target_path, "wb"))
def detectYarns(hor_name,
ver_name,
target_name,
image=None,
write_images=False):
hor = cv2.imread(hor_name, cv2.IMREAD_GRAYSCALE)
ver = cv2.imread(ver_name, cv2.IMREAD_GRAYSCALE)
labels_hor = label(hor)
regions_hor = regionprops(labels_hor)
labels_ver = label(ver)
regions_ver = regionprops(labels_ver)
warp_points_list = floatPointList()
weft_points_list = floatPointList()
#Die float-points sind um pwh pixel verschoben, da das erkannte Bild um genau diese Anzahl Pixel kleiner ist als das original
#Nicht bei fcn!!
for props in regions_hor:
y0, x0 = props.centroid
temp = floatPoint(x=x0,
y=y0,
label="weft_float",
area=props.area,
convArea=props.convex_area)
weft_points_list.append(temp)
for props in regions_ver:
y0, x0 = props.centroid
temp = floatPoint(x=x0,
y=y0,
label="warp_float",
area=props.area,
convArea=props.convex_area)
warp_points_list.append(temp)
for myFloatPoint in warp_points_list:
myFloatPoint.find_warp_neighbors(warp_points_list)
for myFloatPoint in weft_points_list:
myFloatPoint.find_weft_neighbors(weft_points_list)
weft_points_list.calcDistances()
warp_points_list.calcDistances()
pickle.dump(weft_points_list, open(target_name + '__weft.p', "wb"))
pickle.dump(warp_points_list, open(target_name + '__warp.p', "wb"))
float_points_list = floatPointList(warp_points_list + weft_points_list)
if write_images == True:
float_points_list.showPoints(image)
plt.savefig(target_name + '.png', bbox_inches='tight', pad_inches=0)
plt.close("all")
def correctYarns(file_name_weft,
file_name_warp,
target_name,
image=None,
write_images=False):
weft_points_list = pickle.load(open(file_name_weft, "rb"))
warp_points_list = pickle.load(open(file_name_warp, "rb"))
check_consistency_warp(warp_points_list)
check_consistency_weft(weft_points_list)
mean_ver_dist = warp_points_list.getMedianWarpDist()
mean_hor_dist = weft_points_list.getMedianWeftDist()
patch_single_warp_connections(warp_points_list, mean_ver_dist)
patch_single_weft_connections(
weft_points_list,
mean_hor_dist,
)
patchDoubleWarps(warp_points_list, mean_ver_dist)
patchDoubleWefts(weft_points_list, mean_hor_dist)
float_points_list = floatPointList(warp_points_list + weft_points_list)
float_points_list.calcDistances()
if write_images == True:
float_points_list.showPoints(image)
plt.savefig(target_name + '.png')
plt.close("all")
pickle.dump(warp_points_list, open(target_name + '__warp.p', "wb"))
pickle.dump(weft_points_list, open(target_name + '__weft.p', "wb"))
def detectDefects_warps(file_name,
target_name,
robust_cov,
image=None,
write_images=False,
threshold=30):
#robust_cov = pickle.load(open(MCD_path, "rb" ))
float_points_list = pickle.load(open(file_name, "rb"))
for pt in float_points_list:
pt.faulty = False
#update distances, since they were changed
float_points_list.calcDistances()
x_vals = [fp.x for fp in float_points_list]
y_vals = [fp.y for fp in float_points_list]
mean_ver_dist = float_points_list.getMedianWarpDist()
min_x = min(x_vals) + 0.5 * mean_ver_dist
max_x = max(x_vals) - 0.5 * mean_ver_dist
min_y = min(y_vals) + 1.2 * mean_ver_dist
max_y = max(y_vals) - 1.2 * mean_ver_dist
inner_points = floatPointList()
for pt in float_points_list:
if min_x < pt.x < max_x and min_y < pt.y < max_y:
inner_points.append(pt)
#TODO: http://scikit-image.org/docs/dev/api/skimage.measure.html
X = np.zeros([len(inner_points), 3])
for idx, pt in enumerate(inner_points):
X[idx, 0] = pt.area
X[idx, 1] = pt.lower_dist
X[idx, 2] = pt.upper_dist
fault_count, faulty_points = calc_error(X, inner_points, robust_cov,
threshold)
faults00 = [] # up left
faults01 = [] # up right
faults10 = [] # down left
faults11 = [] # down right
num_points00 = 0
num_points01 = 0
num_points10 = 0
num_points11 = 0
mean_x = (max_x + min_x) / 2
mean_y = (max_y + min_y) / 2
overlap = 50
for point in inner_points:
if point.x < mean_x + overlap and point.y < mean_y + overlap:
num_points00 += 1
if point.x > mean_x - overlap and point.y < mean_y + overlap:
num_points01 += 1
if point.x < mean_x + overlap and point.y > mean_y - overlap:
num_points10 += 1
if point.x > mean_x - overlap and point.y > mean_y - overlap:
num_points11 += 1
for point in faulty_points:
if point.x < mean_x + overlap and point.y < mean_y + overlap:
faults00.append(point)
if point.x > mean_x - overlap and point.y < mean_y + overlap:
faults01.append(point)
if point.x < mean_x + overlap and point.y > mean_y - overlap:
faults10.append(point)
if point.x > mean_x - overlap and point.y > mean_y - overlap:
faults11.append(point)
# num_inner_points = inner_points.__len__()
# density = fault_count / num_inner_points * 100
# density, max_d = calculate_density(faulty_points)
density00 = faults00.__len__() / num_points00 * 100
density01 = faults01.__len__() / num_points01 * 100
density10 = faults10.__len__() / num_points10 * 100
density11 = faults11.__len__() / num_points11 * 100
#print(file_name + " fault count von: " + str(fault_count))
#print(file_name + " Faults per 100 flt points: " + str(density))
if write_images == True:
float_points_list.showPoints(image)
plt.savefig(target_name + '.png')
plt.close("all")
#return fault_count
return density00, density01, density10, density11