Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import numpy as np
|
3 |
+
import laspy
|
4 |
+
from sklearn.cluster import DBSCAN
|
5 |
+
from sklearn.metrics import accuracy_score
|
6 |
+
from scipy.spatial import ConvexHull
|
7 |
+
from skimage.measure import profile_line
|
8 |
+
|
9 |
+
st.title("Tree Analysis App")
|
10 |
+
|
11 |
+
# Upload LAS file
|
12 |
+
uploaded_file = st.file_uploader("Upload a LAS file", type=["las", "laz"])
|
13 |
+
|
14 |
+
if uploaded_file is not None:
|
15 |
+
# Load the LAS file
|
16 |
+
las_file = laspy.read(uploaded_file)
|
17 |
+
height_filter = np.logical_and(las_file.z > 1, las_file.z < 30)
|
18 |
+
las_file = las_file[height_filter]
|
19 |
+
# Extract the x and y coordinates from the LAS file
|
20 |
+
x = las_file.x
|
21 |
+
y = las_file.y
|
22 |
+
|
23 |
+
# Combine the x and y coordinates into a feature matrix
|
24 |
+
feature_matrix = np.column_stack((x, y))
|
25 |
+
|
26 |
+
# Segment the trees using DBSCAN clustering with a specified distance threshold (e.g., 2 meters)
|
27 |
+
tree_labels = DBSCAN(eps=2, min_samples=10).fit_predict(feature_matrix)
|
28 |
+
|
29 |
+
# Count the number of trees
|
30 |
+
num_trees = len(set(tree_labels)) - (1 if -1 in tree_labels else 0)
|
31 |
+
st.write(f"Number of trees: {num_trees}")
|
32 |
+
for i in range(num_trees):
|
33 |
+
indices = np.where(tree_labels == i)[0]
|
34 |
+
|
35 |
+
tree_x = x[indices]
|
36 |
+
tree_y = y[indices]
|
37 |
+
|
38 |
+
tree_mid_x = np.mean(tree_x)
|
39 |
+
tree_mid_y = np.mean(tree_y)
|
40 |
+
|
41 |
+
st.write(f"Tree {i+1} middle point: ({tree_mid_x:.3f}, {tree_mid_y:.3f})")
|
42 |
+
|
43 |
+
def calculate_tree_data(points):
|
44 |
+
height = np.max(points.z) - np.min(points.z)
|
45 |
+
xy_points = np.column_stack((points.X, points.Y))
|
46 |
+
hull = ConvexHull(xy_points)
|
47 |
+
crown_spread = np.sqrt(hull.area / np.pi)/10
|
48 |
+
z_trunk = np.percentile(points.z, 20) # assume trunk is the lowest 20% of the points
|
49 |
+
trunk_points = points[points.z < z_trunk]
|
50 |
+
dbh = 2 * np.mean(np.sqrt((trunk_points.X - np.mean(trunk_points.X)) ** 2 + (trunk_points.Y - np.mean(trunk_points.Y)) ** 2))
|
51 |
+
return height, crown_spread, dbh
|
52 |
+
|
53 |
+
tree_data = []
|
54 |
+
for tree_label in range(num_trees):
|
55 |
+
# Extract points for the current tree
|
56 |
+
tree_points = las_file.points[tree_labels == tree_label]
|
57 |
+
# Calculate tree data
|
58 |
+
data = calculate_tree_data(tree_points)
|
59 |
+
# Append data to list
|
60 |
+
tree_data.append(data)
|
61 |
+
|
62 |
+
# Print the data for each tree
|
63 |
+
for i, data in enumerate(tree_data):
|
64 |
+
st.write(f"Tree {i + 1} - Height: {data[0]:.3f} m, Crown Spread: {data[1]:.3f} m, DBH: {data[2]:.3f} mm")
|