Crater Detection Pipeline

Pipeline and Image Objects

class pycda.CDA(detector='tiny', extractor='fast_circle', classifier='convolution')

Bases: object

CDA is the crater detection model pipeline. Its three main components are the detector, the extractor, and the classifier; with no keywords, will initialize with some basic models with general applicability and fast performance.

Attributes:

detector (str, detector object): The detector model used by

pipeline. Accepted text arguments are: ‘tiny’ : default, fast, performant detector ‘unet’ : similar model to tiny, more robust, slower ‘dummy’: does not make detections, used for testing.

extractor (str, extractor object): The extractor model used by

pipeline. Accepted text arguments are: ‘fast_circle’ : default, converts detections into circles ‘watershed’ : uses watershed segmentation to generate proposals ‘dummy’ : does not make extractions, used for testing

classifier (str, classifier object): The classifier model used

by pipeline. Accepted string arguments are: ‘convolution’ : default, uses a convolutional neural net model ‘none’ : use no classifier (None type also accepted) ‘dummy’ : assigns random likelihoods, used for testing.

get_prediction(input_image, verbose=False)

Performs a detection on input_image. Returns a prediction object.

predict(input_image, threshold=0.5, verbose=False)

Performs a detection on input_image. Returns a pandas dataframe with detections.

class pycda.CDAImage(image)

Bases: object

CDA image object; image stored as array; .show() method allows for easy viewing.

as_array()

Returns the image as a numpy array.

show(show_ticks=False)

Displays the input image by plotting raster

pycda.load_image(filename)

load an image from the input filename path. returns a CDAImage object.