bp_analysis.abc_tracker#

Functions#

`calc_laplacian`(image, config[, kernel])
`find_seeds`(src_image, config, **kwargs)
`get_n_rounds_dilation`(config)
`dilate`(config, args, *kwargs)
`dilate_laplacian`(config, seeds[, im, laplacian, mask, ...])
`dilate_contour`(config, seeds, im[, n_rounds, ...])
`remove_edge_touchers`(labeled_feats, tracked_image)
`remove_false_positives`(labeled_feats, laplacian, ...)	Remove features for which another round of dilation would add too many
`filter_close_neighbors`(labeled_feats, config, ...)	Remove features that are too close to another feature, on the grounds that
`filter_size`(tracked_image, config)
`remove_overlapping_false_pos`(tracked_image)	If, in a central frame, a feature gets too big and is flagged as a false
`id_files`(config[, dir, procs])
`scan_directory_for_data`(dir)
`fully_process_one_image`(→ bp_analysis.feature.TrackedImage)
`load_file`(im, config)
`id_image_first_half`(im, config[, mask])
`id_image_second_half`(config, features, seeds, ...)

bp_analysis.abc_tracker.dilate_laplacian(config, seeds, im=None, laplacian=None, mask=None, n_rounds=None)#

bp_analysis.abc_tracker.dilate_contour(config, seeds, im, n_rounds=None, _region_rounds_override=0)#

bp_analysis.abc_tracker.remove_edge_touchers(labeled_feats, tracked_image)#

bp_analysis.abc_tracker.remove_false_positives(labeled_feats, laplacian, config, im, seeds, tracked_image)#: Remove features for which another round of dilation would add too many pixels, on the idea that, after the rounds already done, true bright points should be mostly marked and the surrounding pixels should be dark lanes, while false positives are inside granules and surrounded by more bright pixels.

bp_analysis.abc_tracker.filter_close_neighbors(labeled_feats, config, tracked_image)#: Remove features that are too close to another feature, on the grounds that there may be room for confusion.

bp_analysis.abc_tracker.filter_size(tracked_image: bp_analysis.feature.TrackedImage, config)#

bp_analysis.abc_tracker.remove_overlapping_false_pos(tracked_image: bp_analysis.feature.TrackedImage)#: If, in a central frame, a feature gets too big and is flagged as a false positive, but in the surrounding frames the feature is good, then the good feature will be filled in in the central frame by the temporal smoothing, but there will already be a false positive feature there. So here we remove false positive features that overlap another feature.

bp_analysis.abc_tracker.fully_process_one_image(file, config) → bp_analysis.feature.TrackedImage#

bp_analysis.abc_tracker.id_image_second_half(config, features, seeds, tracked_image, im)#