def __init__(self, k=1, debug=False):
if get_opencv_version() == 3:
self.knn = cv2.ml.KNearest_create()
else:
self.knn = cv2.KNearest()
self.k = k
self.debug = debug
def classify(self, features):
if FEATURE_DATATYPE != numpy.float32:
features = numpy.asarray(features, dtype=numpy.float32)
if get_opencv_version() == 3:
retval, result_classes, neigh_resp, dists = self.knn.findNearest(features, k=1)
else:
retval, result_classes, neigh_resp, dists = self.knn.find_nearest(features, k=1)
return result_classes
def train(self, features, classes):
if FEATURE_DATATYPE != numpy.float32:
features = numpy.asarray(features, dtype=numpy.float32)
if CLASS_DATATYPE != numpy.float32:
classes = numpy.asarray(classes, dtype=numpy.float32)
features, classes = Classifier._filter_unclassified(features, classes)
if get_opencv_version() == 3:
self.knn.train(features, cv2.ml.ROW_SAMPLE, classes)
else:
self.knn.train(features, classes)
def _segment(self, image):
self.image = image
image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
image = cv2.adaptiveThreshold(image, maxValue=255, adaptiveMethod=cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
thresholdType=cv2.THRESH_BINARY, blockSize=self.block_size, C=self.c)
if get_opencv_version() == 3:
_, contours, hierarchy = cv2.findContours(image, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
else:
contours, hierarchy = cv2.findContours(image, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
segments = segments_to_numpy([cv2.boundingRect(c) for c in contours])
self.contours, self.hierarchy = contours, hierarchy # store, may be needed for debugging
return segments