Python cv2.CHAIN_APPROX_TC89_KCOS Examples
The following are 11
code examples of cv2.CHAIN_APPROX_TC89_KCOS().
You can vote up the ones you like or vote down the ones you don't like,
and go to the original project or source file by following the links above each example.
You may also want to check out all available functions/classes of the module
cv2
, or try the search function
.
.
Example #1
| Source File: ChickenVision.py From ChickenVision with MIT License | 6 votes |
|
def findTargets(frame, mask):
# Finds contours
_, contours, _ = cv2.findContours(mask, cv2.RETR_TREE, cv2.CHAIN_APPROX_TC89_KCOS)
# Take each frame
# Gets the shape of video
screenHeight, screenWidth, _ = frame.shape
# Gets center of height and width
centerX = (screenWidth / 2) - .5
centerY = (screenHeight / 2) - .5
# Copies frame and stores it in image
image = frame.copy()
# Processes the contours, takes in (contours, output_image, (centerOfImage)
if len(contours) != 0:
image = findTape(contours, image, centerX, centerY)
else:
# pushes that it deosn't see vision target to network tables
networkTable.putBoolean("tapeDetected", False)
# Shows the contours overlayed on the original video
return image
# Finds the balls from the masked image and displays them on original stream + network tables
Example #2
| Source File: ChickenVision.py From ChickenVision with MIT License | 6 votes |
|
def findCargo(frame, mask):
# Finds contours
_, contours, _ = cv2.findContours(mask, cv2.RETR_TREE, cv2.CHAIN_APPROX_TC89_KCOS)
# Take each frame
# Gets the shape of video
screenHeight, screenWidth, _ = frame.shape
# Gets center of height and width
centerX = (screenWidth / 2) - .5
centerY = (screenHeight / 2) - .5
# Copies frame and stores it in image
image = frame.copy()
# Processes the contours, takes in (contours, output_image, (centerOfImage)
if len(contours) != 0:
image = findBall(contours, image, centerX, centerY)
else:
# pushes that it doesn't see cargo to network tables
networkTable.putBoolean("cargoDetected", False)
# Shows the contours overlayed on the original video
return image
# Draws Contours and finds center and yaw of orange ball
# centerX is center x coordinate of image
# centerY is center y coordinate of image
Example #3
| Source File: util.py From SpaceNet_Off_Nadir_Solutions with Apache License 2.0 | 5 votes |
|
def contours_hierarchy(mask):
# first, find contours with cv2: it's much faster than shapely
image, contours, hierarchy = cv2.findContours(
((mask == 1) * 255).astype(np.uint8),
cv2.RETR_CCOMP,
cv2.CHAIN_APPROX_TC89_KCOS) # cv2.CHAIN_APPROX_SIMPLE,#orig cv2.RETR_CCOMP, cv2.CHAIN_APPROX_TC89_KCOS
return contours, hierarchy
Example #4
| Source File: binarized_filter_result.py From DVCNN_Lane_Detection with Apache License 2.0 | 5 votes |
|
def binarized_whatlike_filtered_image(self, img):
"""
Do normalization and thresholding on the result of weighted hat-like filter image to extract line candidate
:param img: input image
:return: list of roi pair (top_roi, fv_roi) class which defined in imdb.py
"""
if img is None:
raise ValueError('Image data is invalid')
# intensity normalizing the image and thresholding thre image
image = img[:, :, 0]
inds = np.where(image[:, :] > 650)
norm_thresh_img = np.zeros(image.shape).astype(np.uint8)
norm_thresh_img[inds] = 255
# find connected component
image, contours, hierarchy = cv2.findContours(image=norm_thresh_img, mode=cv2.RETR_CCOMP,
method=cv2.CHAIN_APPROX_TC89_KCOS)
response_points = self.__find_response_points_in_contours(contours=contours, image=norm_thresh_img)
# find rotate rect of each contour and check if it fits the condition, if fits the condition then save the
# bounding rectangle of the contour
result = []
valid_contours = 0
for index, contour in enumerate(contours):
rotrect = cv2.minAreaRect(contour)
if self.__is_rrect_valid(rotrect):
# the contours is valid and can be saved
roi_contours = contour
roi_contours = np.reshape(roi_contours, newshape=(roi_contours.shape[0], roi_contours.shape[2]))
roi_index = valid_contours
valid_contours += 1
top_roi_db = imdb.Roidb(roi_index=roi_index, roi_contours=roi_contours,
roi_response_points=response_points[index]) # type:
fv_roi_db, roi_is_valid = self.__map_roi_to_front_view(roidb=top_roi_db)
if roi_is_valid:
result.append((top_roi_db, fv_roi_db))
return result, norm_thresh_img
Example #5
| Source File: extract_candidate.py From DVCNN_Lane_Detection with Apache License 2.0 | 5 votes |
|
def __extract_line_from_filtered_image(img):
"""
Do normalization and thresholding on the result of weighted hat-like filter image to extract line candidate
:param img:input image
:return:rotate rect list []
"""
image = img[:, :, 0]
inds = np.where(image[:, :] > 300)
norm_thresh_image = np.zeros(image.shape).astype(np.uint8)
norm_thresh_image[inds] = 255
# find connected component
image, contours, hierarchy = cv2.findContours(image=norm_thresh_image, mode=cv2.RETR_CCOMP,
method=cv2.CHAIN_APPROX_TC89_KCOS)
# find rotate rect of each contour and check if it fits the condition, if fits the condition then save the
# bounding rectangle of the contour
rotate_rect_list = []
bounding_rect_list = []
for i in range(len(contours)):
contour = contours[i]
rotrect = cv2.minAreaRect(contour)
if RoiExtractor.__is_rrect_valid(rotrect):
rotate_rect_list.append(rotrect)
bnd_rect = cv2.boundingRect(contour)
bounding_rect_list.append(bnd_rect)
result = {
'rotate_rect_list': rotate_rect_list,
'bounding_rect_list': bounding_rect_list
}
return result
Example #6
| Source File: TableRecognition.py From OTR with GNU General Public License v3.0 | 5 votes |
|
def __init__(self, img, **kwargs):
contours, hierarchy = _find_contours(img, cv2.RETR_TREE, cv2.CHAIN_APPROX_TC89_KCOS, **kwargs)
self.hierarchy = hierarchy
self.contours = contours
self.imgshape = img.shape
Example #7
| Source File: TableRecognition.py From OTR with GNU General Public License v3.0 | 5 votes |
|
def compute_missing_cell_contours(self, missing_cells_mask):
contx, _ = _find_contours(missing_cells_mask, cv2.RETR_LIST, cv2.CHAIN_APPROX_TC89_KCOS)
return contx
Example #8
| Source File: main.py From DE3-ROB1-CHESS with Creative Commons Attribution 4.0 International | 4 votes |
|
def extractPiece(tile, margin=0.05):
imgs = [tile]
w, h, _ = tile.shape
im_gray = cv2.cvtColor(tile, cv2.COLOR_BGR2GRAY)
imgs.append(cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR))
# im_gray = im_gray[(h*margin):(h*(1-margin)),
# (w*margin):(w*(1-margin))]
# imgs.append(cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR))
# im_gray = cv2.equalizeHist(im_gray)
im_gray = cv2.medianBlur(im_gray, 3)
imgs.append(cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR))
bright = np.mean(im_gray)
im_bw = im_gray
im_bw[np.where(im_gray < bright)] = 0
im_bw[np.where(im_gray >= bright)] = 255
imgs.append(cv2.cvtColor(im_bw, cv2.COLOR_GRAY2BGR))
if np.mean(im_bw) < 128:
im_bw = 255 - im_bw
imgs.append(cv2.cvtColor(im_bw, cv2.COLOR_GRAY2BGR))
#_, im_bw = cv2.threshold(im_gray, 50, 250, cv2.THRESH_BINARY | cv2.THRESH_OTSU)
#im_bw = cv2.Canny(im_bw, 0,255, apertureSize=5)
contours,hierarchy = cv2.findContours(im_bw.copy(), cv2.RETR_CCOMP, cv2.CHAIN_APPROX_TC89_KCOS)
hulls = [cv2.convexHull(c) for c in contours]
ids = ignoreContours(im_bw, hulls, max_area_percentage=0.75, min_area_percentage=0.2)
im_bw = cv2.cvtColor(im_bw, cv2.COLOR_GRAY2BGR)
tmp = im_bw.copy()
for i in ids:
c = np.squeeze(hulls[i], 1)
drawContour(tmp, c, randomColor(), thickness=1)
imgs.append(tmp)
return imgs
Example #9
| Source File: extractBoard.py From DE3-ROB1-CHESS with Creative Commons Attribution 4.0 International | 4 votes |
|
def extractBoards(img, w, h):
"""Extracts all boards from an image. This function applies perspective correction.
:param img: source image
:param w: output width
:param h: output height
:returns: a list the extracted board images
"""
im_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
## Doc ##
#writeDocumentationImage(im_gray, "gray")
## Doc ##
(thresh, im_bw) = cv2.threshold(im_gray, 128, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)
## Doc ##
#writeDocumentationImage(im_bw, "bw")
## Doc ##
contours,hierarchy = cv2.findContours(im_bw, cv2.RETR_CCOMP, cv2.CHAIN_APPROX_TC89_KCOS)
## Doc ##
#doc_im_contour = cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR)
#for c in contours:
# c = np.squeeze(c,1)
# drawContour(doc_im_contour, c, randomColor())
#writeDocumentationImage(doc_im_contour, "contours")
#doc_im_perspective = cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR)
#doc_im_contour = cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR)
## Doc ##
contour_ids = ignoreContours(im_bw, contours, hierarchy)
boards = []
for i in contour_ids:
c = contours[i]
c = np.squeeze(c,1)
## Doc ##
#color = randomColor()
#drawContour(doc_im_contour, c, color)
## Doc ##
perspective = getPerspective(img, c)
if perspective is not None:
b = extractPerspective(img, perspective, w, h)
boards.append(b)
## Doc ##
#drawPerspective(doc_im_perspective, perspective)
## Doc ##
## Doc ##
#writeDocumentationImage(boards[-1], "extracted")
#writeDocumentationImage(doc_im_contour, "contours_filtered")
#writeDocumentationImage(doc_im_perspective, "perspective")
## Doc ##
return boards
Example #10
| Source File: main.py From DE3-ROB1-CHESS with Creative Commons Attribution 4.0 International | 4 votes |
|
def extractPiece(tile, margin=0.05):
imgs = [tile]
w, h, _ = tile.shape
im_gray = cv2.cvtColor(tile, cv2.COLOR_BGR2GRAY)
imgs.append(cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR))
# im_gray = im_gray[(h*margin):(h*(1-margin)),
# (w*margin):(w*(1-margin))]
# imgs.append(cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR))
# im_gray = cv2.equalizeHist(im_gray)
im_gray = cv2.medianBlur(im_gray, 3)
imgs.append(cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR))
bright = np.mean(im_gray)
im_bw = im_gray
im_bw[np.where(im_gray < bright)] = 0
im_bw[np.where(im_gray >= bright)] = 255
imgs.append(cv2.cvtColor(im_bw, cv2.COLOR_GRAY2BGR))
if np.mean(im_bw) < 128:
im_bw = 255 - im_bw
imgs.append(cv2.cvtColor(im_bw, cv2.COLOR_GRAY2BGR))
#_, im_bw = cv2.threshold(im_gray, 50, 250, cv2.THRESH_BINARY | cv2.THRESH_OTSU)
#im_bw = cv2.Canny(im_bw, 0,255, apertureSize=5)
contours,hierarchy = cv2.findContours(im_bw.copy(), cv2.RETR_CCOMP, cv2.CHAIN_APPROX_TC89_KCOS)
hulls = [cv2.convexHull(c) for c in contours]
ids = ignoreContours(im_bw, hulls, max_area_percentage=0.75, min_area_percentage=0.2)
im_bw = cv2.cvtColor(im_bw, cv2.COLOR_GRAY2BGR)
tmp = im_bw.copy()
for i in ids:
c = np.squeeze(hulls[i], 1)
drawContour(tmp, c, randomColor(), thickness=1)
imgs.append(tmp)
return imgs
Example #11
| Source File: extract.py From DE3-ROB1-CHESS with Creative Commons Attribution 4.0 International | 4 votes |
|
def extractBoards(img, w, h):
"""Extracts all boards from an image. This function applies perspective correction.
:param img: source image
:param w: output width
:param h: output height
:returns: a list the extracted board images
"""
im_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
## Doc ##
#writeDocumentationImage(im_gray, "gray")
## Doc ##
(thresh, im_bw) = cv2.threshold(im_gray, 128, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)
## Doc ##
#writeDocumentationImage(im_bw, "bw")
## Doc ##
contours,hierarchy = cv2.findContours(im_bw, cv2.RETR_CCOMP, cv2.CHAIN_APPROX_TC89_KCOS)
## Doc ##
#doc_im_contour = cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR)
#for c in contours:
# c = np.squeeze(c,1)
# drawContour(doc_im_contour, c, randomColor())
#writeDocumentationImage(doc_im_contour, "contours")
#doc_im_perspective = cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR)
#doc_im_contour = cv2.cvtColor(im_gray, cv2.COLOR_GRAY2BGR)
## Doc ##
contour_ids = ignoreContours(im_bw, contours, hierarchy)
boards = []
for i in contour_ids:
c = contours[i]
c = np.squeeze(c,1)
## Doc ##
#color = randomColor()
#drawContour(doc_im_contour, c, color)
## Doc ##
perspective = getPerspective(img, c)
if perspective is not None:
b = extractPerspective(img, perspective, w, h)
boards.append(b)
## Doc ##
#drawPerspective(doc_im_perspective, perspective)
## Doc ##
## Doc ##
#writeDocumentationImage(boards[-1], "extracted")
#writeDocumentationImage(doc_im_contour, "contours_filtered")
#writeDocumentationImage(doc_im_perspective, "perspective")
## Doc ##
return boards
