Plantcv

[](https://doi.org/10.5281/zenodo.3122076)

PlantCV v3.3.0 adds new functionality and fixes several bugs and usability issues. Big thanks to HaleySchuhl, dschneiderch, JLJ90, huberma, and karnoldbio for work and guidance on the updates below.

**Summary of changes**:

* Added `plantcv.visualize` sub-package
* Moved `pseudocolor` and `histogram` into sub-package
* Added `colorize_masks` function to sub-package to make false-colored images from a set of binary masks (e.g. output masks from the naive Bayes classifier)
* Added `plantcv.opening` and `plantcv.closing` functions (removes salt and pepper noise)
* Added `plantcv.threshold.custom_range` function (threshold based on upper and lower values)
* Added `plantcv.within_frame` function (tests if object, in a binary image, is within the field of view)
* Added `plantcv.morphology` sub-package
* Added `skeletonize` function to sub-package (skeletonizes a binary image)
* Added `prune` function to sub-package (removes spurs from skeleton)
* Added `check_cycles` function to sub-package (checks for connected cycles in skeleton)
* Added `find_branch_pts` function to sub-package (finds branch points in skeleton)
* Added `find_tips` function to sub-package (finds tips in skeleton)
* Added `segment_skeleton` function to sub-package (segments a skeleton into component paths)
* Added `segment_sort` function to sub-package (sorts segments into primary and secondary groups)
* Added `segment_id` function to sub-package (plots/labels segment IDs)
* Added `segment_path_length` function to sub-package (calculates segment lengths)
* Added `segment_euclidean_length` function to sub-package (calculates segment Euclidean lengths)
* Added `segment_curvature` function to sub-package (calculates the ratio of path length to Euclidean length)
* Added `segment_angle` function to sub-package (calculates the overall angle of the segment)
* Added `segment_insertion_angle` function to sub-package (calculates the angle that a segment intersects another segment)
* Added `segment_tangent_angle` function to sub-package (calculates the angle between the tangents of the ends of each segment)
* Added `parallel` sub-package.
* The sub-package contains functions that were originally from the `plantcv-pipeline.py` script file
* Renamed `plantcv-pipeline.py` to `plantcv-workflow.py`
* Removed SQLite database and requirements. Data are now output in a JSON-formatted text file
* `plantcv.print_results` now outputs data in JSON format
* The `Outputs` class now stores data in a single dictionary
* Added `add_observation` method to the `Outputs` class. Allows user to add custom observations to the output
* Output observations are stored by a unique variable name along with a trait name, method, scale (units), data type, value, and label(s)
* Keep 1st generation sub contours when using 'largest' in `roi_objects`
* In `analyze_color`, color and color property scales now use the conventional scale for each type (e.g. hue is a value from 0-359 degrees while green is a value from 0-255)
* Added summary statistics for hue in `analyze_color`: median hue value, circular mean hue value, and the circular mean standard deviation of hue
* Removed the `bins` argument from `analyze_color`

[](https://doi.org/10.5281/zenodo.2587982)

PlantCV v3.2.0 adds new functionality and fixes several bugs and usability issues.

**Summary of changes**:

* Added functionality to the `plantcv.print_results` function. It now allows users to print the data returned, naming the .txt file whatever they would like.
* Added complete scripts at the bottom of each PlantCV tutorial in the documentation.
* Added new function `plantcv.roi.multi`, allowing users to specify parameters for a grid of regions of interest (ROIs) or supply a list of centers for ROIs if they are not in a grid arrangement.
Restructure the way `plantcv.analyze_*` functions return outputs. Each function now returns the images so the user can save them.
* Enhancements to the `plantcv.pseudocolor` function including the addition of an “image” background option, adding the option to turn off titles/axes and colorbar, and an auto-crop option.
Changed Matplotlib import (now imported globally in `plantcv.__init__.py`), fixing the non-fatal warning from setting the matplotlib backend multiple times.
* Added debug mode to `plantcv.analyze_color` function.
`plantcv.analyze_bound_horizontal` was previously determining line position differently than the rest of the functions in PlantCV. Instead of `line_position=0` signifying the bottom of the image, it will now signify the top of the image.
* Add a `line_thickness` graphics options to the `params` class so users can change the line thickness for the functions plotting lines onto images (analyze_object, all ROI functions, analyze_bound_horizontal, analyze_bound_vertical, acute_vertex, x_axis_pseudolandmark, y_axis_pseudolandmark, scale_features, roi_objects, object_composition).
* Add a link in the table of contents to the PlantCV Hyperspectral subproject documentation.
* Add an “image” background option to the `plantcv.auto_crop` function.
* Improved code testing coverage to 100%.
* Allow string arguments to be case insensitive.
* Added a new option to `roi_type` in `plantcv.roi_objects` which allows only the largest contour to to be kept.
* Standardize argument order and naming across functions.
* Updated functionality of the `plantcv.plot_hist` function, including adding an optional mask argument and allowing users to save histograms.

* All analysis functions now output visualization images rather than attempting to save them directly. This removes the `filename` input parameter and gives users the flexibility to save, plot, etc. what they want to where they want.
* A new function `pseudocolor` was added to give users the ability and flexibility to take grayscale images and colorize them with any `matplotlib` colormap, autocrop them to the plant (or other object), mask out background, etc. This function can be used with the changes made above to customize visualization of output images. Colorized images have a built-in color scale bar.
* Due to the changes above, the `plot_colorbar` function was removed as it's no longer used.
* Histograms are now plotted with `plotnine` instead of `matplotlib`.
* A Canny edge detection function was added (`canny_edge_detect`).
* A color standard card auto-detection method (`transform.find_color_card`) was added.

[](https://doi.org/10.5281/zenodo.2542072)

This release modifies the Travis CI build procedure to use dependencies installed from PyPI rather than `conda`. This should fix issues with deployed versions of PlantCV on PyPI and allow us to identify dependency incompatibilities faster. It also decreased build and testing types by more than half.

[](https://doi.org/10.5281/zenodo.2539838)

This release fixes an issue that prevents PlantCV from being installed with OpenCV v4, which is not currently supported. The documentation and docstrings were thoroughly revised for correctness, completeness, and consistency. The installation instructions were updated to include methods for installing PlantCV from PyPI and Bioconda.

[](https://doi.org/10.5281/zenodo.2532362)

PlantCV v3.0.3 fixes some minor issues with PyPI deployment.