Skip to content

Instantly share code, notes, and snippets.

@prl900
Last active April 27, 2020 10:06
Show Gist options
  • Save prl900/d175b9d9ec909ee76a2323b5e79ec05c to your computer and use it in GitHub Desktop.
Save prl900/d175b9d9ec909ee76a2323b5e79ec05c to your computer and use it in GitHub Desktop.
Display the source blob
Display the rendered blob
Raw
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 1.- Theoretical basis of the methodology for computing Fuel Moisture Content (FMC)\n",
"\n",
"#### In this work we try to match current reflectance values for each pixel with the ones corresponding to field locations where the FMC have been measured on the field.\n",
"\n",
"#### We represent reflectance values for the different spectral bands (plus a synthetic normalised index NDII). To find the similarities between these 6-dimensional vectors we use the angle $ \\theta $ from the geometrical definition of the dot product. From the definition we can find the angle span between two vectors as indicated in this figure:\n",
"\n",
"![title](https://upload.wikimedia.org/wikipedia/commons/7/76/Inner-product-angle.svg)\n",
"\n",
"#### For example, the angle between vectors $ a = [0,0,1] $ and $ b = [0,0.7,0.7] $ can be computed implementing this formula in numpy as follows:"
]
},
{
"cell_type": "code",
"execution_count": 154,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0.7853981633974483"
]
},
"execution_count": 154,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import numpy as np\n",
"\n",
"a = np.array([0,0,1])\n",
"b = np.array([0,0.7,0.7])\n",
"\n",
"θ = np.arccos((np.dot(a,b))/(np.linalg.norm(a)*np.linalg.norm(b)))\n",
"θ"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### This value corresponds to $ \\pi / 4 $, as we expect looking at the values of vectors a and b.\n",
"\n",
"#### The proposed method computes this operation matching each value of reflectance with the ones in a simulated radiative transfer Look Up Table (LUT) whose corresponding FMC values are known. To find the closest vectors in this LUT to a new modis pixel, we need to compute the angles between the vector defined by this pixel and each of the values in the LUT. \n",
"\n",
"#### The following example computes the angles between vector $a$ which represents the reflectance values of a Modis pixel and $b$ which represents the values of the LUT (5 vectors). In numpy we can perform this as a vectorised computation to maximise performance as follows: "
]
},
{
"cell_type": "code",
"execution_count": 179,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"array([0.78539816, 0.55859932, 1.01219701, 0.54902514, 0.5880026 ])"
]
},
"execution_count": 179,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"a = np.array([0,0,1])\n",
"b = np.array([[0,0.7,0.7],[0,0.5,0.8],[0,0.8,0.5],[0,0.52,0.85],[0,0.4,0.6]])\n",
"\n",
"θ = np.arccos(np.einsum('ij,j->i', b, a)/(np.einsum('i,i->', a, a)**.5*np.einsum('ij,ij->i',b, b)**.5))\n",
"θ"
]
},
{
"cell_type": "code",
"execution_count": 180,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"array([3, 4, 1])"
]
},
"execution_count": 180,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"idx = np.argpartition(θ, top_n)[:top_n]\n",
"\n",
"idx"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Now we compute the top 3 most similar entries (smaller $ \\theta $). In numpy there is a function which identifies the n-smallest values in an array. Note that the values it returns do not necessarily have to be in order."
]
},
{
"cell_type": "code",
"execution_count": 174,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"array([0.54902514, 0.5880026 , 0.55859932])"
]
},
"execution_count": 174,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"top_n = 3\n",
"\n",
"idx = np.argpartition(θ, top_n)[:top_n]\n",
"\n",
"θ[idx]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### The last step would require us to get the FMC values for these entries in the LUT and compute their median as our FMC estimation for this pixel. \n",
"\n",
"\n",
"## 2.- Validation of the methodology using the field data and provided LUT\n",
"\n",
"### 2.1- Computing FMC for field data\n",
"\n",
"#### The data for performing this experiment comes from a Matlab file provided by David Riano which can be found at the NCI:\n",
"\n",
"```\n",
"/g/data1a/xc0/user/Riano/FMC/NCI_scripts_v2/run_all_20180411/MODISFMC_collection6_20180411_allMARIANO_wwindows_wFMCaveQCOK3.mat\n",
"```\n",
"(I keep a copy of this in case it dissapears from that location)\n",
"\n",
"#### Let's start by loading the data containing the vegetation type, Modis reflectance values and FMC values for the field data. The data has been extracted from Matlab and converted into numpy format."
]
},
{
"cell_type": "code",
"execution_count": 181,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"(69622, 8)\n",
"(69622, 1)\n"
]
}
],
"source": [
"vegtype = np.load(\"vegtype.npy\").astype(np.uint8)\n",
"mask = ~np.isnan(vegtype) * (vegtype!=0)\n",
"vegtype = vegtype[mask]\n",
"\n",
"x = np.load(\"fmc_x.npy\").T / 10000\n",
"x = x[mask,:]\n",
"ndii = (x[:,1]-x[:,5])/(x[:,1]+x[:,5])\n",
"refs = np.concatenate((x, ndii[:,None]), axis=1)\n",
"print(refs.shape)\n",
"\n",
"fmc_obs = np.load(\"fmc_y.npy\")[:,None]\n",
"\n",
"fmc_obs = fmc_obs[mask,:]\n",
"print(fmc_obs.shape)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### There are almost 70,000 FMC measurements and their corresponding Modis reflectance values (7 bands + NDII)\n",
"\n",
"#### We are going to estimate the FCM values using the proposed methodology and compare with the observed values for this locations.\n",
"\n",
"#### We start by loading the LUT containing the simulated sprectral values for the three types of vegetation:"
]
},
{
"cell_type": "code",
"execution_count": 182,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(8708, 6)"
]
},
"execution_count": 182,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"lut = np.load(\"LUT.npy\")\n",
"#bands = [1,2,4,6,7] + NDII\n",
"lut = lut[:, [0,1,3,5,6,7]]\n",
"lut.shape"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### The LUT contains 8708 entries which can be separated by types of vegetation using the following function:"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def get_vegtype_idx(veg_type):\n",
" if veg_type == 1.:\n",
" return (0, 2563)\n",
" elif veg_type == 2.:\n",
" return (2563, 4226)\n",
" elif veg_type == 3.:\n",
" return (4226, 8708)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### This function splits the values in the FMC table by vegetation [1=shrub,2=forest,3=pasture] (!check this values with Marta although the correspondance is not relevant for computing FMC.)\n",
"\n",
"#### The following function identifies the top n closest reflectance vectors (their location or indexes) in the LUT to the input Modis pixel using the methodology previously described."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"def get_fmc_idxs(veg_type, ref_modis):\n",
" top_n = 40\n",
" idx = get_vegtype_idx(veg_type)\n",
"\n",
" # Select Veg type subset from LUT table\n",
" lut_veg = lut[idx[0]:idx[1], :]\n",
"\n",
" θ = np.arccos(np.einsum('ij,j->i', lut_veg, ref_modis)/(np.einsum('i,i->', ref_modis, ref_modis)**.5*np.einsum('ij,ij->i',lut_veg, lut_veg)**.5))\n",
"\n",
" idxs = np.argpartition(θ, top_n)[:top_n] + idx[0]\n",
"\n",
" return idxs"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### The resulting indexes are used to extract the FMC values for the reflectances in the LUT. We stpre this values in the `fmc` array and we make sure the size is consistent with the number of entries in the LUT"
]
},
{
"cell_type": "code",
"execution_count": 183,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"(8708,)\n"
]
}
],
"source": [
"fmc = np.load(\"FMC.npy\")\n",
"print(fmc.shape)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Now we compute the top 40 vectors in the LUT to using the field data Modis reflectance values:"
]
},
{
"cell_type": "code",
"execution_count": 184,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(69622, 40)"
]
},
"execution_count": 184,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"results = []\n",
"\n",
"for i in range(vegtype.shape[0]):\n",
" results.append(fmc[get_fmc_idxs(vegtype[i], refs[i,[0,1,3,5,6,7]])])\n",
" \n",
"results = np.array(results)\n",
"results.shape"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### This array contains the 40 FMC values corresponding to the closest vectors in the LUT. \n",
"\n",
"#### Let's see these values for the first point and its median which would be the final FMC value as per the methodology we are following:"
]
},
{
"cell_type": "code",
"execution_count": 187,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[126.9 61.8 61.8 61.8 90.5 90.5 90.5 90.5 90.5 90.5 90.5 90.5\n",
" 90.5 128.9 128.9 128.9 128.9 128.9 128.9 128.9 128.9 128.9 128.9 128.9\n",
" 128.9 128.9 128.9 90.5 90.5 90.5 126.9 61.8 61.8 126.9 90.5 61.8\n",
" 61.8 135.7 61.8 135.7]\n"
]
},
{
"data": {
"text/plain": [
"90.5"
]
},
"execution_count": 187,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"print(results[0,:])\n",
"np.median(results[0,:])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### 2.2.- Statistical analysis of the results\n",
"\n",
"#### First let's plot the estimated vs the observed FMC values to have an idea about the accuracy of the methodology."
]
},
{
"cell_type": "code",
"execution_count": 197,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[<matplotlib.lines.Line2D at 0x7f3d9fc3f1d0>,\n",
" <matplotlib.lines.Line2D at 0x7f3d9fc3f320>]"
]
},
"execution_count": 197,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 720x432 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"%matplotlib inline\n",
"\n",
"from matplotlib import pyplot as plt\n",
"\n",
"plt.figure(figsize=(10,6))\n",
"# every 300 points to avoid clutter\n",
"plt.plot(np.vstack((np.median(results, axis=1),fmc_obs[:,0])).T[::300,:])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Let's compute the Mean Square Error (MSE) at estimating FMC and compare it to the variance of the observed FMC, which will indicate the error of a model always predicting the mean FMC value."
]
},
{
"cell_type": "code",
"execution_count": 211,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"MSE for estimated FMC: 2282.87\n",
"Variance observed FMC: 1894.79\n"
]
}
],
"source": [
"print(f\"MSE for estimated FMC: {np.mean(np.square(np.median(results, axis=1)-fmc_obs[:,0])):0.2f}\")\n",
"print(f\"Variance observed FMC: {np.var(fmc_obs[:,0]):0.2f}\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 3.- Using a machine learning method to compute FMC\n",
"\n",
"#### We start by preparing the data. First we add the vegetation type as another input feature for the model, after we shuffle the order of the data and create the train and validation split datasets"
]
},
{
"cell_type": "code",
"execution_count": 203,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"((69622, 9), (69622,))"
]
},
"execution_count": 203,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"x = np.concatenate((refs,vegtype[:,None]),axis=1)\n",
"y = fmc_obs[:,0]\n",
"x.shape, y.shape"
]
},
{
"cell_type": "code",
"execution_count": 204,
"metadata": {},
"outputs": [],
"source": [
"np.random.seed(0)\n",
"idxs = np.arange(x.shape[0])\n",
"np.random.shuffle(idxs)\n",
"x = x[idxs, :]\n",
"y = y[idxs]\n",
"\n",
"x_train = x[:50000,:]\n",
"x_test = x[50000:,:]\n",
"y_train = y[:50000]\n",
"y_test = y[50000:]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Now we train a Random Forest with these data:"
]
},
{
"cell_type": "code",
"execution_count": 216,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/prl900/miniconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
" \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n"
]
},
{
"data": {
"text/plain": [
"RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=10,\n",
" max_features='auto', max_leaf_nodes=None,\n",
" min_impurity_decrease=0.0, min_impurity_split=None,\n",
" min_samples_leaf=1, min_samples_split=2,\n",
" min_weight_fraction_leaf=0.0, n_estimators=10,\n",
" n_jobs=None, oob_score=False, random_state=0, verbose=0,\n",
" warm_start=False)"
]
},
"execution_count": 216,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from sklearn.ensemble import RandomForestRegressor\n",
"\n",
"regr = RandomForestRegressor(max_depth=10, random_state=0)\n",
"regr.fit(x_train, y_train)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Let's plot the results over the validation partition:"
]
},
{
"cell_type": "code",
"execution_count": 217,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[<matplotlib.lines.Line2D at 0x7f3d58179eb8>,\n",
" <matplotlib.lines.Line2D at 0x7f3d58187048>]"
]
},
"execution_count": 217,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 720x432 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.figure(figsize=(10,6))\n",
"# every 100 points to avoid clutter\n",
"plt.plot(np.vstack((regr.predict(x_test),y_test)).T[::100,:])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Let's see the MSE values for each partition and the overall performance:"
]
},
{
"cell_type": "code",
"execution_count": 218,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"MSE for validation FMC: 1371.09\n",
"Variance validation FMC: 1895.43\n",
"-------------------------------------------\n",
"MSE for training FMC: 1167.62\n",
"Variance training FMC: 1894.52\n",
"-------------------------------------------\n",
"MSE overall FMC: 1224.97\n",
"Variance overall FMC: 1894.79\n"
]
}
],
"source": [
"print(f\"MSE for validation FMC: {np.mean(np.square(y_test-regr.predict(x_test))):0.2f}\")\n",
"print(f\"Variance validation FMC: {y_test.var():0.2f}\")\n",
"print('-------------------------------------------')\n",
"print(f\"MSE for training FMC: {np.mean(np.square(y_train-regr.predict(x_train))):0.2f}\")\n",
"print(f\"Variance training FMC: {y_train.var():0.2f}\")\n",
"print('-------------------------------------------')\n",
"print(f\"MSE overall FMC: {np.mean(np.square(y-regr.predict(x))):0.2f}\")\n",
"print(f\"Variance overall FMC: {y.var():0.2f}\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Let's measure the effect of the vegetation type input in the performance of the model. This is important to understand the need of this parameter."
]
},
{
"cell_type": "code",
"execution_count": 219,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/prl900/miniconda3/lib/python3.7/site-packages/sklearn/ensemble/forest.py:245: FutureWarning: The default value of n_estimators will change from 10 in version 0.20 to 100 in 0.22.\n",
" \"10 in version 0.20 to 100 in 0.22.\", FutureWarning)\n"
]
},
{
"data": {
"text/plain": [
"RandomForestRegressor(bootstrap=True, criterion='mse', max_depth=10,\n",
" max_features='auto', max_leaf_nodes=None,\n",
" min_impurity_decrease=0.0, min_impurity_split=None,\n",
" min_samples_leaf=1, min_samples_split=2,\n",
" min_weight_fraction_leaf=0.0, n_estimators=10,\n",
" n_jobs=None, oob_score=False, random_state=0, verbose=0,\n",
" warm_start=False)"
]
},
"execution_count": 219,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"regr = RandomForestRegressor(max_depth=10, random_state=0)\n",
"regr.fit(x_train[:,:-1], y_train)"
]
},
{
"cell_type": "code",
"execution_count": 220,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"MSE for validation FMC: 1547.39\n",
"Variance validation FMC: 1895.43\n",
"-------------------------------------------\n",
"MSE for training FMC: 1353.59\n",
"Variance training FMC: 1894.52\n",
"-------------------------------------------\n",
"MSE overall FMC: 1408.21\n",
"Variance overall FMC: 1894.79\n"
]
}
],
"source": [
"print(f\"MSE for validation FMC: {np.mean(np.square(y_test-regr.predict(x_test[:,:-1]))):0.2f}\")\n",
"print(f\"Variance validation FMC: {y_test.var():0.2f}\")\n",
"print('-------------------------------------------')\n",
"print(f\"MSE for training FMC: {np.mean(np.square(y_train-regr.predict(x_train[:,:-1]))):0.2f}\")\n",
"print(f\"Variance training FMC: {y_train.var():0.2f}\")\n",
"print('-------------------------------------------')\n",
"print(f\"MSE overall FMC: {np.mean(np.square(y-regr.predict(x[:,:-1]))):0.2f}\")\n",
"print(f\"Variance overall FMC: {y.var():0.2f}\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### So, comparing these results with the previous ones we can conclude that the vegetation type has a significant influence improving the performance of the model."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.3"
}
},
"nbformat": 4,
"nbformat_minor": 4
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment