Visualize show spectra

docs/visualize_show_spectra.md

@@ -0,0 +1,27 @@
+## Show Spectra for Selected Pixels (Developing)
+
+`ShowSpectra` is a class that is capable of handling user's selecting and clicking behavior by showing spectra of user selected pixels and storing selected coordinates as well as spectra.
+
+*class* **plantcv.visualize.ShowSpectra(spectral_data, figsize=(12,6))**
+- To initialize the ShowSpectra class, the only required parameter is `spectral_data`, which is of type `__main__.Spectral_data`.
+- Another optional parameter is the desired figure size `figsize`, by default `figsize=(12,6)`.
+
+### Attributes
+**spectral_data** (`__main__.Spectral_data`, required): input hyperspectral image.
+
+**spectra** (`list`): spectra for all selected pixels.
+
+**points** (`list`): list of coordinates of selected pixels.
+
+```python
+
+from plantcv import plantcv as pcv
+
+show_spectra = pcv.visualize.ShowSpectra(spectral_data=array)
+
+```
+
+Check out this video for a sample usage:
+<iframe src="https://player.vimeo.com/video/522535625" width="640" height="360" frameborder="0" allow="autoplay; fullscreen; picture-in-picture" allowfullscreen></iframe>
+
+**Source Code:** [Here](https://github.com/danforthcenter/plantcv/blob/master/plantcv/plantcv/visualize/show_spectra.py)

plantcv/plantcv/visualize/__init__.py

@@ -9,6 +9,7 @@
 from plantcv.plantcv.visualize.obj_sizes import obj_sizes
 from plantcv.plantcv.visualize.obj_size_ecdf import obj_size_ecdf
 from plantcv.plantcv.visualize.hyper_histogram import hyper_histogram
+from plantcv.plantcv.visualize.show_spectra import ShowSpectra
 
 __all__ = ["pseudocolor", "colorize_masks", "histogram", "clustered_contours", "colorspaces", "auto_threshold_methods",
-           "overlay_two_imgs", "colorize_label_img", "obj_size_ecdf", "obj_sizes", "hyper_histogram"]
+           "overlay_two_imgs", "colorize_label_img", "obj_size_ecdf", "obj_sizes", "hyper_histogram", "ShowSpectra"]

plantcv/plantcv/visualize/show_spectra.py

@@ -0,0 +1,154 @@
+# Show spectral or spectra of mouse selected pixel(s) for a given hyperspectral image
+
+from scipy.spatial import distance
+import numpy as np
+import matplotlib.pyplot as plt
+import matplotlib.patches as patches
+from matplotlib.widgets import Slider
+import cv2
+
+
+def _find_closest(pt, pts):
+    """ Given coordinates of a point and a list of coordinates of a bunch of points, find the point that has the smallest Euclidean to the given point
+
+    :param pt: (tuple) coordinates of a point
+    :param pts: (a list of tuples) coordinates of a list of points
+    :return: index of the closest point and the coordinates of that point
+    """
+    if pt in pts:
+        return pt
+    dists = distance.cdist([pt], pts, 'euclidean')
+    idx = np.argmin(dists)
+    return idx, pts[idx]
+
+
+class ShowSpectra(object):
+    """
+    An interactive visualization tool that shows spectral (spectra) for selected pixel(s).
+    """
+
+    def __init__(self, spectral_data, figsize=(12, 8)):
+        """
+        Initialization
+        :param spectral_data: hyperspectral image data
+        :param figsize: desired figure size, (12,8) by default
+        """
+        print("Warning: this tool is under development and is expected to have updates frequently, please check the documentation page to make sure you are using the correct version!")
+
+        # initialize the pseudocolor rgb data (convert from BGR to RGB)
+        self.pseudo_rgb = cv2.cvtColor(spectral_data.pseudo_rgb, cv2.COLOR_BGR2RGB)
+
+        self.fig, self.axes = plt.subplots(1, 2, figsize=figsize)
+        self.axes[0].imshow(self.pseudo_rgb)
+        self.axes[0].set_title("Please click on interested pixels\n Right click to remove")
+
+        self.axes[1].set_xlabel("wavelength (nm)")
+        self.axes[1].set_ylabel("reflectance")
+        self.axes[1].set_title("Spectra")
+        self.axes[1].set_ylim([0, 1])
+
+        # adjust the main plot to make room for the sliders
+        plt.subplots_adjust(left=0.25, bottom=0.25)
+
+        # make a horizontal slider to control the radius
+        axradius = self.fig.add_axes([0.15, 0.035, 0.3, 0.035], facecolor='lightgoldenrodyellow')  # [left, bottom, width, height]
+        self.radius_slider = Slider(
+            ax=axradius,
+            label="radius",
+            valmin=0,
+            valmax=500,
+            valinit=1,
+            orientation="horizontal"
+        )
+
+        # Set useblit=True on most backends for enhanced performance.
+        # cursor = Cursor(axes[0], horizOn=True, vertOn=True, useblit=True, color='red', linewidth=2)
+
+        self.points = []
+        self.spectra = []
+        self.events = []
+
+        self.array_data = spectral_data.array_data
+        self.wvs = [k for k in spectral_data.wavelength_dict.keys()]
+
+        self.spectral_std = None
+        self.spectral_mean = None
+
+        # initialize radius
+        self.r = 0
+        self.x = None
+        self.y = None
+        self.rectangle = None
+
+        self.fig.canvas.mpl_connect('button_press_event', self.onclick)
+        self.radius_slider.on_changed(self.update)
+
+    def spectra_roi(self):
+        """Pull out the spectra inside a square ROI
+        """
+        r = int(self.r)
+        y = int(self.y)
+        x = int(self.x)
+        kernel_ = np.ones((2 * r + 1, 2 * r + 1)) / (4 * r * r + 4 * r + 1)
+
+        square = self.array_data[(y - r):(y + r + 1), (x - r):(x + r + 1), :]
+
+        num_bands = square.shape[2]
+
+        kernel = np.repeat(kernel_[:, :, np.newaxis], num_bands, axis=2)
+
+        multiplied = np.multiply(square, kernel)
+        multiplied_spectra = np.reshape(multiplied, (-1, num_bands))
+
+        self.spectral_mean = multiplied_spectra.sum(axis=0)
+        self.spectral_std = multiplied_spectra.std(axis=0)
+        # only prepare the patch for plotting if r>1 (not only one pixel, but a square of pixels)
+        if r > 0:
+            self.rectangle = patches.Rectangle((x - r, y - r), 2 * r, 2 * r, edgecolor="red", fill=False)
+
+    def onclick(self, event):
+        self.events.append(event)
+        if str(event.inaxes._subplotspec) == 'GridSpec(1, 2)[0:1, 0:1]':
+            if event.button == 1:
+                self.x, self.y = event.xdata, event.ydata
+                self.axes[0].plot(event.xdata, event.ydata, 'x', c='red')
+                self.spectra_roi()
+                if self.r > 1:
+                    self.axes[0].add_patch(self.rectangle)
+
+                self.axes[1].errorbar(self.wvs, self.spectral_mean, xerr=self.spectral_std / 2)
+                self.points.append((event.xdata, event.ydata))
+            else:
+                idx_remove, _ = _find_closest((event.xdata, event.ydata), self.points)
+                # remove the last added point
+                # idx_remove = -1
+
+                # remove the closest point to the user right clicked one
+                self.points.pop(idx_remove)
+                if len(self.points) > 0:
+                    self.x, self.y = self.points[-1]
+                ax0plots = self.axes[0].lines
+                ax0patches = self.axes[0].patches
+                ax1plots = self.axes[1].lines
+                self.axes[0].lines.remove(ax0plots[idx_remove])
+                if len(ax0patches) > 0:
+                    self.axes[0].patches.remove(ax0patches[idx_remove])
+                self.axes[1].lines.remove(ax1plots[idx_remove])
+        self.fig.canvas.draw()
+
+    def update(self, val):
+        self.r = self.radius_slider.val
+        self.spectra_roi()
+
+        # remove old plots
+        idx_remove = -1
+        ax0patches = self.axes[0].patches
+        if len(ax0patches) > 0:
+            self.axes[0].patches.remove(ax0patches[idx_remove])
+        ax1plots = self.axes[1].lines
+        self.axes[1].lines.remove(ax1plots[idx_remove])
+
+        # add new plots
+        self.axes[0].add_patch(self.rectangle)
+        self.axes[1].errorbar(self.wvs, self.spectral_mean, xerr=self.spectral_std / 2)
+        self.fig.canvas.draw_idle()

tests/tests.py

@@ -6492,6 +6492,68 @@ def test_plantcv_visualize_overlay_two_imgs_bad_alpha():
     with pytest.raises(RuntimeError):
         _ = pcv.visualize.overlay_two_imgs(img1=img1, img2=img2, alpha=alpha)
 
+def test_plantcv_visualize_show_spectra():
+    # read hypersectral data
+    spectral_filename = os.path.join(HYPERSPECTRAL_TEST_DATA, HYPERSPECTRAL_DATA)
+    array_data        = pcv.hyperspectral.read_data(filename=spectral_filename)
+
+    # initialization
+    show_spectra = pcv.visualize.ShowSpectra(array_data, figsize=(12, 6))
+    assert len(show_spectra.events) == 0
+
+    # create mock events
+    e1 = matplotlib.backend_bases.MouseEvent(name="button_press_event", canvas=show_spectra.fig.canvas, x=0, y=0,button=1)
+    e1.inaxes = show_spectra.axes[0]
+    e1.inaxes._subplotspec = show_spectra.axes[0]._subplotspec
+    e1.xdata = 0
+    e1.ydata = 0
+
+    e2 = matplotlib.backend_bases.MouseEvent(name="button_press_event", canvas=show_spectra.fig.canvas, x=0, y=0, button=3)
+    e2.inaxes = show_spectra.axes[0]
+    e2.inaxes._subplotspec = show_spectra.axes[0]._subplotspec
+    e2.xdata = 0
+    e2.ydata = 0
+
+    e1_ = matplotlib.backend_bases.MouseEvent(name="button_press_event", canvas=show_spectra.fig.canvas, x=0, y=0,button=1)
+    e1_.inaxes = show_spectra.axes[0]
+    e1_.inaxes._subplotspec = show_spectra.axes[0]._subplotspec
+    e1_.xdata = 0
+    e1_.ydata = 0
+
+    e3 = matplotlib.backend_bases.MouseEvent(name="button_press_event", canvas=show_spectra.fig.canvas, x=1, y=0, button=3)
+    e3.inaxes = show_spectra.axes[0]
+    e3.inaxes._subplotspec = show_spectra.axes[0]._subplotspec
+    e3.xdata = 1
+    e3.ydata = 0
+
+    show_spectra.onclick(e1)
+    show_spectra.onclick(e2)
+    show_spectra.onclick(e1_)
+    show_spectra.onclick(e1_)
+    show_spectra.onclick(e3)
+
+    assert len(show_spectra.events) == 5
+
+    # test for updating
+    # initialization
+    array_d = array_data.array_data
+    array_data.array_data = np.concatenate((array_d, array_d, array_d, array_d, array_d), axis=0)
+    show_spectra = pcv.visualize.ShowSpectra(array_data, figsize=(12, 6))
+    e1 = matplotlib.backend_bases.MouseEvent(name="button_press_event", canvas=show_spectra.fig.canvas, x=1, y=1,button=1)
+    e1.inaxes = show_spectra.axes[0]
+    e1.inaxes._subplotspec = show_spectra.axes[0]._subplotspec
+    e1.xdata = 2
+    e1.ydata = 2
+    show_spectra.onclick(e1)
+    val = 2
+    show_spectra.radius_slider.val = val
+    show_spectra.update(val)
+    show_spectra.onclick(e1)
+    show_spectra.update(val)
+    show_spectra.onclick(e2)
+    assert len(show_spectra.axes[0].patches) > 0
+    show_spectra.onclick(e1)
+    assert len(show_spectra.axes[0].patches) > 0
 
 def test_plantcv_visualize_overlay_two_imgs_size_mismatch():
     pcv.params.debug = None