john
/
asteroid-automator
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

More functionality for CVImage class 

TODO: standardize the chainable outputs from CVImage. Will we return a new CVImage or a numpy array? I'm leaning towards the second one, so CVImage's return values can be used as numpy/cv2 inputs, and the ImagePipeline class will handle multiple modification calls.
This commit is contained in:
John McCardle 2021-12-26 21:11:14 -05:00
parent ce51ea9a1c
commit fd1767cfe1
1 changed files with 36 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

46
imagepipeline.py
View File

@ -2,26 +2,26 @@ import cv2
import numpy as np import numpy as np
class CVImage: class CVImage:
def __init__(self, label="", img=None, iscolor=False): def __init__(self, label="", img=None, color=False, **kwargs):
self.label = label self.label = label
self.image = img self.image = img
self.iscolor = iscolor self.iscolor = color
if kwargs:
kwargs["color"] = color
self.load(**kwargs)
def load(self, filename, color=False, label=None): def load(self, filename, color=False, label=None):
self.image = cv2.imread(filename, int(color)) self.image = cv2.imread(filename, int(color))
if label: self.label = label if label: self.label = label
return self
def copy(self): def copy(self):
return np.copy(self.image) return np.copy(self.image)
## def snip(self, point, width_height):
## return self.image[self.point[0]:self.point[0]+self.width_height[0],
## self.point[1]:self.point[1]+self.width_height[1]]
def snip(self, rect): def snip(self, rect):
assert all((len(rect)==2, len(rect[0])==2, len(rect[1])==2)) #((x,y),(w,h)) assert all((len(rect)==2, len(rect[0])==2, len(rect[1])==2)) #((x,y),(w,h))
return self.image[self.rect[0][0]:self.rect[0][1], return self.image[rect[0][0]:rect[1][0],
self.rect[1][0]:self.rect[1][1]] rect[0][1]:rect[1][1]]
def mask(self, rect, mask_color=None, nonmask_color=None): def mask(self, rect, mask_color=None, nonmask_color=None):
assert all((len(rect)==2, len(rect[0])==2, len(rect[1])==2)) #((x,y),(w,h)) assert all((len(rect)==2, len(rect[0])==2, len(rect[1])==2)) #((x,y),(w,h))
@ -38,16 +38,40 @@ class CVImage:
keypoints, descriptions = sift.detectAndCompute(self.image, None) keypoints, descriptions = sift.detectAndCompute(self.image, None)
return pointcluster.cluster_set([k.pt for k in keypoints], cluster_radius) return pointcluster.cluster_set([k.pt for k in keypoints], cluster_radius)
def set_blob_params(self, minThreshold = 10, maxThreshold = 200, def blob_params(self, minThreshold = 10, maxThreshold = 200,
minArea = None, maxArea = None, minArea = None, maxArea = None,
minCircularity = None, maxCircularity = None, minCircularity = None, maxCircularity = None,
minConvexity = None, maxConvexity = None, minConvexity = None, maxConvexity = None,
minInertiaRatio = None, maxInertiaRatio = None): minInertiaRatio = None, maxInertiaRatio = None):
p = cv2.SimpleBlobDetector_Params() p = cv2.SimpleBlobDetector_Params()
p.minThreshold = minThreshold
p.maxThreshold = maxThreshold
if minArea or maxArea:
p.filterByArea = True
if minArea: p.minArea = minArea
if maxArea: p.maxArea = maxArea
if minConvexity or maxConvexity:
p.filterByConvexity = True
if minConvexity: p.minConvexity = minConvexity
if maxConvexity: p.maxConvexity = maxConvexity
if minInertiaRatio or maxInertiaRatio:
p.filterByInertiaRatio = True
if minInertiaRatio: p.minInertiaRatio = minInertiaRatio
if maxInertiaRatio: p.maxInertiaRatio = maxInertiaRatio
if minCircularity or maxCircularity:
p.filterByCircularity = True
if minCircularity: p.minCircularity = minCircularity
if maxCircularity: p.maxCircularity = maxCircularity
return p
def blob_detect(self, params=None, invert=False):
if params is None: params = self.blob_params()
detector = cv2.SimpleBlobDetector_create(params)
return detector.detect(cv2.bitwise_not(self.image) if invert else self.image)
def blob_detect(self, params=None): def show(self, delay=0):
pass cv2.imshow(self.label, self.image)
cv2.waitKey(delay)
class ImagePipeline: class ImagePipeline: