Vehicle Detection.py

# coding: utf-8

# ## Import Packages

# In[1]:


import matplotlib.image as mpimg
import matplotlib.pyplot as plt
import numpy as np
import cv2
import glob
import time
from sklearn.svm import LinearSVC
from sklearn.preprocessing import StandardScaler
from skimage.feature import hog
get_ipython().run_line_magic('matplotlib', 'inline')


# Reminder to self: when we use PNG together with `mpimg.imread`, the intensity of the pixels is between 0 and 1 (requiring a normalization) and the channels in the RGB order, if the image is JPEG and we use that same method, the intensity is between 0 and 255 ( in which case no normalization is required). For PNG images you can use `cv2.imread` which already leaves the intensity between 0 and 255 (however with the channels in the order BGR) or use `mpimg.imread` and do normalization.

# ## Define Detection Functions
# These are largely derived from the "Object Detection" Lesson

# In[2]:


# Define a function to return HOG features and visualization
def get_hog_features(img, orient, pix_per_cell, cell_per_block, 
                        vis=False, feature_vec=True):
    # Call with two outputs if vis==True
    if vis == True:
        features, hog_image = hog(img, orientations=orient, 
                                  pixels_per_cell=(pix_per_cell, pix_per_cell),
                                  block_norm= 'L2-Hys',
                                  cells_per_block=(cell_per_block, cell_per_block), 
                                  transform_sqrt=True, 
                                  visualise=vis, feature_vector=feature_vec)
        return features, hog_image
    # Otherwise call with one output
    else:      
        features = hog(img, orientations=orient, 
                       pixels_per_cell=(pix_per_cell, pix_per_cell),
                       cells_per_block=(cell_per_block, cell_per_block), 
                       block_norm= 'L2-Hys',
                       transform_sqrt=True, 
                       visualise=vis, feature_vector=feature_vec)
        return features

# Define a function to compute binned color features  
def bin_spatial(img, size=(32, 32)):
    # Use cv2.resize().ravel() to create the feature vector
    features = cv2.resize(img, size).ravel() 
    # Return the feature vector
    return features

# Define a function to compute color histogram features 
# NEED TO CHANGE bins_range if reading .png files with mpimg!
def color_hist(img, nbins=32, bins_range=(0, 256)):
    # Compute the histogram of the color channels separately
    channel1_hist = np.histogram(img[:,:,0], bins=nbins, range=bins_range)
    channel2_hist = np.histogram(img[:,:,1], bins=nbins, range=bins_range)
    channel3_hist = np.histogram(img[:,:,2], bins=nbins, range=bins_range)
    # Concatenate the histograms into a single feature vector
    hist_features = np.concatenate((channel1_hist[0], channel2_hist[0], channel3_hist[0]))
    # Return the individual histograms, bin_centers and feature vector
    return hist_features

# Define a function to convert color spaces
def convert_color(img, color_space='RGB'):
    if color_space != 'RGB':
        if color_space == 'HSV':
            return cv2.cvtColor(img, cv2.COLOR_RGB2HSV)
        elif color_space == 'LUV':
            return cv2.cvtColor(img, cv2.COLOR_RGB2LUV)
        elif color_space == 'HLS':
            return cv2.cvtColor(img, cv2.COLOR_RGB2HLS)
        elif color_space == 'YUV':
            return cv2.cvtColor(img, cv2.COLOR_RGB2YUV)
        elif color_space == 'YCrCb':
            return cv2.cvtColor(img, cv2.COLOR_RGB2YCrCb)
    else: return np.copy(img)      

# Define a function to extract features from a list of images
# Have this function call bin_spatial() and color_hist()
def extract_features(imgs, color_space='RGB', spatial_size=(32, 32),
                        hist_bins=32, orient=9, 
                        pix_per_cell=8, cell_per_block=2, hog_channel=0,
                        spatial_feat=True, hist_feat=True, hog_feat=True):
    # Create a list to append feature vectors to
    features = []
    # Iterate through the list of images
    for file in imgs:
        file_features = []
        # Read in each one by one
        image = mpimg.imread(file)
        # apply color conversion if other than 'RGB'
        feature_image = convert_color(image, color_space=color_space)      

        if spatial_feat == True:
            spatial_features = bin_spatial(feature_image, size=spatial_size)
            file_features.append(spatial_features)
        if hist_feat == True:
            # Apply color_hist()
            hist_features = color_hist(feature_image, nbins=hist_bins)
            file_features.append(hist_features)
        if hog_feat == True:
        # Call get_hog_features() with vis=False, feature_vec=True
            if hog_channel == 'ALL':
                hog_features = []
                for channel in range(feature_image.shape[2]):
                    hog_features.append(get_hog_features(feature_image[:,:,channel], 
                                        orient, pix_per_cell, cell_per_block, 
                                        vis=False, feature_vec=True))
                hog_features = np.ravel(hog_features)        
            else:
                hog_features = get_hog_features(feature_image[:,:,hog_channel], orient, 
                            pix_per_cell, cell_per_block, vis=False, feature_vec=True)
            # Append the new feature vector to the features list
            file_features.append(hog_features)
        features.append(np.concatenate(file_features))
    # Return list of feature vectors
    return features
    
# Define a function that takes an image,
# start and stop positions in both x and y, 
# window size (x and y dimensions),  
# and overlap fraction (for both x and y)
def slide_window(img, x_start_stop=[None, None], y_start_stop=[None, None], 
                    xy_window=(64, 64), xy_overlap=(0.5, 0.5)):
    # If x and/or y start/stop positions not defined, set to image size
    if x_start_stop[0] == None:
        x_start_stop[0] = 0
    if x_start_stop[1] == None:
        x_start_stop[1] = img.shape[1]
    if y_start_stop[0] == None:
        y_start_stop[0] = 0
    if y_start_stop[1] == None:
        y_start_stop[1] = img.shape[0]
    # Compute the span of the region to be searched    
    xspan = x_start_stop[1] - x_start_stop[0]
    yspan = y_start_stop[1] - y_start_stop[0]
    # Compute the number of pixels per step in x/y
    nx_pix_per_step = np.int(xy_window[0]*(1 - xy_overlap[0]))
    ny_pix_per_step = np.int(xy_window[1]*(1 - xy_overlap[1]))
    # Compute the number of windows in x/y
    nx_buffer = np.int(xy_window[0]*(xy_overlap[0]))
    ny_buffer = np.int(xy_window[1]*(xy_overlap[1]))
    nx_windows = np.int((xspan-nx_buffer)/nx_pix_per_step) 
    ny_windows = np.int((yspan-ny_buffer)/ny_pix_per_step) 
    # Initialize a list to append window positions to
    window_list = []
    # Loop through finding x and y window positions
    # Note: you could vectorize this step, but in practice
    # you'll be considering windows one by one with your
    # classifier, so looping makes sense
    for ys in range(ny_windows):
        for xs in range(nx_windows):
            # Calculate window position
            startx = xs*nx_pix_per_step + x_start_stop[0]
            endx = startx + xy_window[0]
            starty = ys*ny_pix_per_step + y_start_stop[0]
            endy = starty + xy_window[1]
            
            # Append window position to list
            window_list.append(((startx, starty), (endx, endy)))
    # Return the list of windows
    return window_list

# Define a function to draw bounding boxes
def draw_boxes(img, bboxes, color=(0, 0, 255), thick=6):
    # Make a copy of the image
    imcopy = np.copy(img)
    # Iterate through the bounding boxes
    for bbox in bboxes:
        # Draw a rectangle given bbox coordinates
        cv2.rectangle(imcopy, bbox[0], bbox[1], color, thick)
    # Return the image copy with boxes drawn
    return imcopy

# Define a function to extract features from a single image window
# This function is very similar to extract_features()
# just for a single image rather than list of images
def single_img_features(img, color_space='RGB', spatial_size=(32, 32),
                        hist_bins=32, orient=9, 
                        pix_per_cell=8, cell_per_block=2, hog_channel=0,
                        spatial_feat=True, hist_feat=True, hog_feat=True):    
    #1) Define an empty list to receive features
    img_features = []
    #2) Apply color conversion if other than 'RGB'
    feature_image = convert_color(image, color_space=color_space) 
    #3) Compute spatial features if flag is set
    if spatial_feat == True:
        spatial_features = bin_spatial(feature_image, size=spatial_size)
        #4) Append features to list
        img_features.append(spatial_features)
    #5) Compute histogram features if flag is set
    if hist_feat == True:
        hist_features = color_hist(feature_image, nbins=hist_bins)
        #6) Append features to list
        img_features.append(hist_features)
    #7) Compute HOG features if flag is set
    if hog_feat == True:
        if hog_channel == 'ALL':
            hog_features = []
            for channel in range(feature_image.shape[2]):
                hog_features.extend(get_hog_features(feature_image[:,:,channel], 
                                    orient, pix_per_cell, cell_per_block, 
                                    vis=False, feature_vec=True))      
        else:
            hog_features = get_hog_features(feature_image[:,:,hog_channel], orient, 
                        pix_per_cell, cell_per_block, vis=False, feature_vec=True)
        #8) Append features to list
        img_features.append(hog_features)

    #9) Return concatenated array of features
    return np.concatenate(img_features)

# Define a function you will pass an image 
# and the list of windows to be searched (output of slide_windows())
def search_windows(img, windows, clf, scaler, color_space='RGB', 
                    spatial_size=(32, 32), hist_bins=32, 
                    hist_range=(0, 256), orient=9, 
                    pix_per_cell=8, cell_per_block=2, 
                    hog_channel=0, spatial_feat=True, 
                    hist_feat=True, hog_feat=True):

    #1) Create an empty list to receive positive detection windows
    on_windows = []
    #2) Iterate over all windows in the list
    for window in windows:
        #3) Extract the test window from original image
        test_img = cv2.resize(img[window[0][1]:window[1][1], window[0][0]:window[1][0]], (64, 64))      
        #4) Extract features for that window using single_img_features()
        features = single_img_features(test_img, color_space=color_space, 
                            spatial_size=spatial_size, hist_bins=hist_bins, 
                            orient=orient, pix_per_cell=pix_per_cell, 
                            cell_per_block=cell_per_block, 
                            hog_channel=hog_channel, spatial_feat=spatial_feat, 
                            hist_feat=hist_feat, hog_feat=hog_feat)
        #5) Scale extracted features to be fed to classifier
        test_features = scaler.transform(np.array(features).reshape(1, -1))
        #6) Predict using your classifier
        prediction = clf.predict(test_features)
        #7) If positive (prediction == 1) then save the window
        if prediction == 1:
            on_windows.append(window)
    #8) Return windows for positive detections
    return on_windows


# ## Data Exploration

# In[3]:


# Read in training data sets
car_files = glob.glob('./vehicles/*/*.png')
noncar_files = glob.glob('./non-vehicles/*/*.png')
num_car_files = len(car_files)
num_noncar_files = len(noncar_files)
print('{} Car images in data set'.format(num_car_files))
print('{} Non-car images in data set'.format(num_noncar_files))


# In[4]:


car_subsample = []
noncar_subsample = []
# for ind in range(3):
#     rand_ind_c = np.random.randint(0, num_car_files)
#     rand_ind_n = np.random.randint(0, num_noncar_files)
# The above lines were originally used, but to preserve predictability I froze the random indices
for rand_ind_c, rand_ind_n in ((1208,5594),(2641,8582),(4217,2820),(5757,4035)):
    car_subsample.append(car_files[rand_ind_c])
    noncar_subsample.append(noncar_files[rand_ind_n])
    
    random_car_img = mpimg.imread(car_files[rand_ind_c])
    random_noncar_img = mpimg.imread(noncar_files[rand_ind_n])

    f, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 5))
    f.tight_layout()

    ax1.imshow(random_car_img)
    ax1.set_title('Random Car', fontsize=20)

    ax2.imshow(random_noncar_img)
    ax2.set_title('Random Noncar', fontsize=20)

    plt.savefig('./output_images/random_images_{}_{}'.format(rand_ind_c, rand_ind_n))


# ## Feature Extraction
# ### Color Exploration

# In[5]:


# Define a function that takes a list of images (same type), adds up their histograms for each color channel
# and prints out the images as well as the histograms

def aggregate_histogram(image_list, color_space, nbins=32, bins_range=(0, 256)):
    hist_vals = np.zeros((3,nbins))
    for image in image_list:
        orig_img = mpimg.imread(image)*255 # convert from 0-1 scaled
        img = convert_color(orig_img, color_space=color_space)
        hist_vals[0] = hist_vals[0] + np.histogram(img[:,:,0], bins=nbins, range=bins_range)[0]
        hist_vals[1] = hist_vals[1] + np.histogram(img[:,:,1], bins=nbins, range=bins_range)[0]
        hist_vals[2] = hist_vals[2] + np.histogram(img[:,:,2], bins=nbins, range=bins_range)[0]

    c_up = [i for i, c in enumerate(color_space) if c.isupper()]
    bin_edges = np.histogram(img[:,:,0], bins=nbins, range=bins_range)[1]
    bin_centers = (bin_edges[1:]  + bin_edges[0:len(bin_edges)-1])/2
    fig = plt.figure(figsize=(20,4))
    plt.subplot(131)
    plt.bar(bin_centers, hist_vals[0])
    plt.xlim(0, 256)
    plt.ylim(0, 500*len(image_list))
    plt.title('{} Histogram'.format(color_space[:c_up[1]]))
    plt.subplot(132)
    plt.bar(bin_centers, hist_vals[1])
    plt.xlim(0, 256)
    plt.ylim(0, 500*len(image_list))
    plt.title('{} Histogram'.format(color_space[c_up[1]:c_up[2]]))
    plt.subplot(133)
    plt.bar(bin_centers, hist_vals[2])
    plt.xlim(0, 256)
    plt.ylim(0, 500*len(image_list))
    plt.title('{} Histogram'.format(color_space[c_up[2]:]))
    return bin_centers, hist_vals

color_spaces = ['RGB','HSV','YCrCb','LUV', 'HLS', 'YUV']

import random
car_samples = [random.choice(car_files) for i in range(100)]
noncar_samples = [random.choice(noncar_files) for i in range(100)]

for color in color_spaces:
    car_centers, car_values = aggregate_histogram(car_samples, color)
    plt.suptitle('Car histograms - {}'.format(color))
#     plt.savefig('./output_images/car_histograms_{}'.format(color)) // freeze for output consistency
    noncar_centers, noncar_values = aggregate_histogram(noncar_samples, color)
    plt.suptitle('Noncar histograms - {}'.format(color))
#     plt.savefig('./output_images/noncar_histograms_{}'.format(color))



# ### HOG Parameters

# In[6]:


# Function that accepts an image and a color space and HOG images of all channels

def hog_plotter(image_name, color_space, orient=8, pix_per_cell=8, cell_per_block=2):
    img = mpimg.imread(image_name)
    color_image = convert_color(img, color_space=color_space)
    _, hog_image1 = get_hog_features(color_image[:,:,0], orient, pix_per_cell, cell_per_block, vis=True, feature_vec=True)
    _, hog_image2 = get_hog_features(color_image[:,:,1], orient, pix_per_cell, cell_per_block, vis=True, feature_vec=True)
    _, hog_image3 = get_hog_features(color_image[:,:,2], orient, pix_per_cell, cell_per_block, vis=True, feature_vec=True)
    fig = plt.figure(figsize=(20,3))
    plt.subplot(141)
    plt.imshow(color_image)
    plt.subplot(142)
    plt.imshow(hog_image1)
    plt.subplot(143)
    plt.imshow(hog_image2)
    plt.subplot(144)
    plt.imshow(hog_image3)


# In[7]:


for ind in range(len(car_subsample)):
    car_image_name = car_subsample[ind]
    hog_plotter(car_image_name, 'RGB')
    plt.suptitle('Car Hog - RGB')
    plt.savefig('./output_images/car_{}_hog_RGB'.format(car_image_name.split('/')[-1][:-4]))
    noncar_image_name = noncar_subsample[ind]
    hog_plotter(noncar_image_name, 'RGB')
    plt.suptitle('Noncar Hog - RGB')
    plt.savefig('./output_images/noncar_{}_hog_RGB'.format(noncar_image_name.split('/')[-1][:-4]))
    hog_plotter(car_image_name, 'HLS')
    plt.suptitle('Car Hog - HLS')
    plt.savefig('./output_images/car_{}_hog_HLS'.format(car_image_name.split('/')[-1][:-4]))
    hog_plotter(noncar_image_name, 'HLS')
    plt.suptitle('Noncar Hog - HLS')
    plt.savefig('./output_images/noncar_{}_hog_HLS'.format(noncar_image_name.split('/')[-1][:-4]))


# #### Trying fitting more pixels per cell

# In[30]:


for ind in range(len(car_subsample)):
    car_image_name = car_subsample[ind]
    noncar_image_name = noncar_subsample[ind]
    hog_plotter(car_image_name, 'RGB', pix_per_cell=16)
    plt.suptitle('Car Hog - RGB, 16 pixels/cell')
    plt.savefig('./output_images/car_{}_hog_RGB_pix{}'.format(car_image_name.split('/')[-1][:-4], 16))
    hog_plotter(noncar_image_name, 'RGB', pix_per_cell=16)
    plt.suptitle('Noncar Hog - RGB, 16 pixels/cell')
    plt.savefig('./output_images/noncar_{}_hog_RGB_pix{}'.format(noncar_image_name.split('/')[-1][:-4], 16))


# #### Less Cells per block

# In[8]:


for ind in range(len(car_subsample)):
    car_image_name = car_subsample[ind]
    noncar_image_name = noncar_subsample[ind]
    hog_plotter(car_image_name, 'RGB', pix_per_cell=16, cell_per_block=1)
    plt.suptitle('Car Hog - RGB, 1 cell/block')
    plt.savefig('./output_images/car_{}_hog_RGB_pix16_blk1'.format(car_image_name.split('/')[-1][:-4], 16))
    hog_plotter(noncar_image_name, 'RGB', pix_per_cell=16, cell_per_block=1)
    plt.suptitle('Noncar Hog - RGB, 1 cell/block')
    plt.savefig('./output_images/noncar_{}_hog_RGB_pix16_blk1'.format(noncar_image_name.split('/')[-1][:-4], 16))


# ### Training the linear classifier

# In[10]:


from sklearn.model_selection import train_test_split

hist_bins = 32
orient = 9
pix_per_cell = 16
cell_per_block = 4
color_space = 'YCrCb'
spatial_size = (32,32)
spatial_feat = False


car_features = extract_features(car_files, color_space=color_space, hist_bins=hist_bins, spatial_size=spatial_size,
                                orient=orient, pix_per_cell=pix_per_cell, cell_per_block=cell_per_block, hog_channel='ALL',
                                spatial_feat=spatial_feat, hist_feat=True, hog_feat = True)

notcar_features = extract_features(noncar_files, color_space=color_space, hist_bins=hist_bins,
                                orient=orient, pix_per_cell=pix_per_cell, cell_per_block=cell_per_block, hog_channel='ALL',
                                spatial_feat=spatial_feat, hist_feat=True, hog_feat = True)

# Create an array stack of feature vectors
X = np.vstack((car_features, notcar_features)).astype(np.float64)

# Define the labels vector
y = np.hstack((np.ones(len(car_features)), np.zeros(len(notcar_features))))

# Split up data into randomized training and test sets
rand_state = np.random.randint(0, 100)
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=rand_state)
    
# Fit a per-column scaler only on the training data
X_scaler = StandardScaler().fit(X_train)
# Apply the scaler to X_train and X_test
X_train = X_scaler.transform(X_train)
X_test = X_scaler.transform(X_test)

print('Feature vector length:', len(X_train[0]))
# Use a linear SVC 
svc = LinearSVC()
# Check the training time for the SVC
t=time.time()
svc.fit(X_train, y_train)
t2 = time.time()
print(round(t2-t, 2), 'Seconds to train SVC...')
# Check the score of the SVC
print('Test Accuracy of SVC = ', round(svc.score(X_test, y_test), 4))
# Check the prediction time for a single sample
t=time.time()
n_predict = 10
print('My SVC predicts: ', svc.predict(X_test[0:n_predict]))
print('For these',n_predict, 'labels: ', y_test[0:n_predict])
t2 = time.time()
print(round(t2-t, 5), 'Seconds to predict', n_predict,'labels with SVC')


# ## Sliding Window Search

# In[119]:


get_ipython().run_line_magic('matplotlib', 'notebook')
fig = plt.figure(figsize=(15,12))
plt.imshow(img)


# In[11]:



# Define a single function that can extract features using hog sub-sampling and make predictions
def find_cars(img, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler, orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat):
#     draw_img = np.copy(img)
    img = img.astype(np.float32)/255

    img_tosearch = img[ystart:ystop,xstart:xstop,:]
    ctrans_tosearch = convert_color(img_tosearch, color_space=color_space)
    if scale != 1:
        imshape = ctrans_tosearch.shape
        ctrans_tosearch = cv2.resize(ctrans_tosearch, (np.int(imshape[1]/scale), np.int(imshape[0]/scale)))
        
    ch1 = ctrans_tosearch[:,:,0]
    ch2 = ctrans_tosearch[:,:,1]
    ch3 = ctrans_tosearch[:,:,2]

    # Define blocks and steps as above
    nxblocks = (ch1.shape[1] // pix_per_cell) - cell_per_block + 1
    nyblocks = (ch1.shape[0] // pix_per_cell) - cell_per_block + 1 
    nfeat_per_block = orient*cell_per_block**2
    
    # 64 was the orginal sampling rate, with 8 cells and 8 pix per cell
    window = pix_per_cell*cell_per_block
    nblocks_per_window = (window // pix_per_cell) - cell_per_block + 1
    cells_per_step = 2  # Instead of overlap, define how many cells to step
    nxsteps = (nxblocks - nblocks_per_window) // cells_per_step + 1
    nysteps = (nyblocks - nblocks_per_window) // cells_per_step + 1
    
    # Compute individual channel HOG features for the entire image
    hog1 = get_hog_features(ch1, orient, pix_per_cell, cell_per_block, feature_vec=False)
    hog2 = get_hog_features(ch2, orient, pix_per_cell, cell_per_block, feature_vec=False)
    hog3 = get_hog_features(ch3, orient, pix_per_cell, cell_per_block, feature_vec=False)
    
    bboxes = []
    for xb in range(nxsteps):
        for yb in range(nysteps):
            ypos = yb*cells_per_step
            xpos = xb*cells_per_step
            # Extract HOG for this patch
            hog_feat1 = hog1[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel() 
            hog_feat2 = hog2[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel() 
            hog_feat3 = hog3[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel() 
            hog_features = np.hstack((hog_feat1, hog_feat2, hog_feat3))

            xleft = xpos*pix_per_cell
            ytop = ypos*pix_per_cell

            # Extract the image patch
            subimg = cv2.resize(ctrans_tosearch[ytop:ytop+window, xleft:xleft+window], (64,64))
          
            # Get color features
            spatial_features = bin_spatial(subimg, size=spatial_size)
            hist_features = color_hist(subimg, nbins=hist_bins)

            # Scale features and make a prediction
            if spatial_feat:
                test_features = X_scaler.transform(np.hstack((hist_features, spatial_features, hog_features)).reshape(1, -1))
            else:
                test_features = X_scaler.transform(np.hstack((hist_features, hog_features)).reshape(1, -1))
            #test_features = X_scaler.transform(np.hstack((shape_feat, hist_feat)).reshape(1, -1))    
            test_prediction = svc.predict(test_features)
            
            if test_prediction == 1:
                xbox_left = np.int(xleft*scale)
                ytop_draw = np.int(ytop*scale)
                win_draw = np.int(window*scale)
                bboxes.append(((xbox_left + xstart, ytop_draw+ystart),(xbox_left+win_draw+xstart,ytop_draw+win_draw+ystart)))
#                 cv2.rectangle(draw_img,(xbox_left + xstart, ytop_draw+ystart),(xbox_left+win_draw+xstart,ytop_draw+win_draw+ystart),(0,0,255),6) 
                
    return bboxes


# ### Playing around with windows and scale

# In[12]:


def tiered_window_search(image):
    bboxes = []
    
    xstart = 500
    xstop = 850
    ystart = 400
    ystop = 475
    scale = 0.5

    bboxes = bboxes + find_cars(image, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    xstart = 0
    xstop = 1200
    ystart = 400
    ystop = 550
    scale = 1
    
    bboxes = bboxes + find_cars(image, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    ystart = 400
    ystop = 650
    scale = 1.5
    
    bboxes = bboxes + find_cars(image, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    ystart = 450
    ystop = 700
    scale = 2
    
    bboxes = bboxes + find_cars(image, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    return bboxes


# In[13]:


get_ipython().run_line_magic('matplotlib', 'inline')
test_files = glob.glob('./test_images/*.jpg')

for filename in test_files:
    img = mpimg.imread(filename)
    draw_img = np.copy(img)

    bboxes = tiered_window_search(img)
    
    for bbx in bboxes:
        cv2.rectangle(draw_img,bbx[0],bbx[1],(0,0,255),6)
    fig = plt.figure(figsize=(15,9))
    plt.imshow(draw_img)
    plt.savefig('output_images/test_bboxes_{}'.format(filename.split('/')[-1]))


# ## Video Generation

# In[14]:


# Import everything needed to edit/save/watch video clips
from moviepy.editor import VideoFileClip
from IPython.display import HTML
import os


# In[15]:


def tiered_window_search_video(img):
    draw_img = np.copy(img)
    bboxes = []
    
    xstart = 500
    xstop = 850
    ystart = 400
    ystop = 475
    scale = 0.5

    bboxes = bboxes + find_cars(img, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    xstart = 0
    xstop = 1200
    ystart = 400
    ystop = 550
    scale = 1
    
    bboxes = bboxes + find_cars(img, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    ystart = 400
    ystop = 650
    scale = 1.5
    
    bboxes = bboxes + find_cars(img, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    ystart = 450
    ystop = 700
    scale = 2
    
    bboxes = bboxes + find_cars(img, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    for bbx in bboxes:
        cv2.rectangle(draw_img,bbx[0],bbx[1],(0,0,255),6)
    return draw_img


# In[19]:


out_video_dir = "output_video_images"
if not os.path.exists(out_video_dir):
    os.mkdir(out_video_dir)
video_output = out_video_dir + '/vehicles_rough2.mp4'
## To speed up the testing process you may want to try your pipeline on a shorter subclip of the video
## To do so add .subclip(start_second,end_second) to the end of the line below
## Where start_second and end_second are integer values representing the start and end of the subclip
## You may also uncomment the following line for a subclip of the first 5 seconds
# clip1 = VideoFileClip("test_videos/solidWhiteRight.mp4").subclip(0,5)
video_clip = VideoFileClip("project_video.mp4")
clip = video_clip.fl_image(tiered_window_search_video) #NOTE: this function expects color images!!
get_ipython().run_line_magic('time', 'clip.write_videofile(video_output, audio=False)')


# In[20]:


HTML("""
<video width="960" height="540" controls>
  <source src="{0}">
</video>
""".format(video_output))


# In[29]:


from scipy.ndimage.measurements import label

# Define a function that generates a heatmap from a list of bboxes
def add_heat(heatmap, bbox_list):
    # Iterate through list of bboxes
    for box in bbox_list:
        # Add += 1 for all pixels inside each bbox
        # Assuming each "box" takes the form ((x1, y1), (x2, y2))
        heatmap[box[0][1]:box[1][1], box[0][0]:box[1][0]] += 1

    # Return updated heatmap
    return heatmap
    
# Define a function that thresholds heatmap to filter out noise
def apply_threshold(heatmap, threshold):
    # Zero out pixels below the threshold
    heatmap[heatmap <= threshold] = 0
    # Return thresholded map
    return heatmap

from scipy.ndimage.measurements import label

# Define a function that draws bboxes around an image with distinct features
def draw_labeled_bboxes(img, labels):
    labels = label(img)
    # Iterate through all detected cars
    for car_number in range(1, labels[1]+1):
        # Find pixels with each car_number label value
        nonzero = (labels[0] == car_number).nonzero()
        # Identify x and y values of those pixels
        nonzeroy = np.array(nonzero[0])
        nonzerox = np.array(nonzero[1])
        # Define a bounding box based on min/max x and y
        bbox = ((np.min(nonzerox), np.min(nonzeroy)), (np.max(nonzerox), np.max(nonzeroy)))
        # Draw the box on the image
        cv2.rectangle(img, bbox[0], bbox[1], (0,255,), 20)
    # Return the image
    return img

def tiered_window_search_smoothed(img):
    
    global recent_heat_maps, frame_ind
    
    draw_img = np.copy(img)
    bboxes = []
    
    xstart = 500
    xstop = 850
    ystart = 400
    ystop = 475
    scale = 0.5

    bboxes = bboxes + find_cars(img, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    xstart = 0
    xstop = 1200
    ystart = 400
    ystop = 550
    scale = 1
    
    bboxes = bboxes + find_cars(img, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    ystart = 400
    ystop = 650
    scale = 1.5
    
    bboxes = bboxes + find_cars(img, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    ystart = 450
    ystop = 700
    scale = 2
    
    bboxes = bboxes + find_cars(img, color_space, xstart, xstop, ystart, ystop, scale, svc, X_scaler,
                        orient, pix_per_cell, cell_per_block, spatial_size, hist_bins, spatial_feat)
    
    if (frame_ind == 0):
        recent_heat_map = np.zeros_like(img[:,:,0]).astype(np.float)

    heat_map = add_heat(recent_heat_map, bboxes)
    heat_map = apply_threshold(heat_map,1)
    labels = label(heat_map)

    draw_img = draw_labeled_bboxes(draw_img, labels)
    
    recent_heat_map = apply_threshold(recent_heat_map, 1)
    frame_ind = 0
        
    return draw_img


# In[30]:


frame_ind = 0

out_video_dir = "output_video_images"
if not os.path.exists(out_video_dir):
    os.mkdir(out_video_dir)
video_output = out_video_dir + '/vehicles_filtered.mp4'
## To speed up the testing process you may want to try your pipeline on a shorter subclip of the video
## To do so add .subclip(start_second,end_second) to the end of the line below
## Where start_second and end_second are integer values representing the start and end of the subclip
## You may also uncomment the following line for a subclip of the first 5 seconds
# clip1 = VideoFileClip("test_videos/solidWhiteRight.mp4").subclip(0,5)
video_clip = VideoFileClip("project_video.mp4")
clip = video_clip.fl_image(tiered_window_search_smoothed) #NOTE: this function expects color images!!
get_ipython().run_line_magic('time', 'clip.write_videofile(video_output, audio=False)')


# In[31]:


HTML("""
<video width="960" height="540" controls>
  <source src="{0}">
</video>
""".format(video_output))