def single_image_features(image, feature_parameter):
#1) Define an empty list to receive features
img_features = []
#2) Apply color conversion if other than 'RGB'
feature_image = convert_color(image, feature_parameter.color_space)
#3) Compute spatial features if flag is set
if feature_parameter.spatial_feat == True:
spatial_features = bin_spatial(feature_image,
size=feature_parameter.spatial_size)
#4) Append features to list
img_features.append(spatial_features)
#5) Compute histogram features if flag is set
if feature_parameter.hist_feat == True:
hist_features, rhist, ghist, bhist, bin_centers = color_hist(
feature_image, nbins=feature_parameter.hist_bins)
#6) Append features to list
img_features.append(hist_features)
#7) Compute HOG features if flag is set
if feature_parameter.hog_feat == True:
hog_features = get_hog_features(feature_image,
feature_parameter.orient,
feature_parameter.pix_per_cell,
feature_parameter.cell_per_block,
feature_parameter.hog_channel)
#8) Append features to list
img_features.append(hog_features)
#9) Return concatenated array of features
return np.concatenate(img_features)
def extract_features(img, orient, pix_per_cell, cell_per_block, spatial_size,
hist_bins):
ch1 = img[:, :, 0]
ch2 = img[:, :, 1]
ch3 = img[:, :, 2]
# Get hog features for each color
hog1 = get_hog_features(ch1,
orient,
pix_per_cell,
cell_per_block,
vis=False,
feature_vec=False)
hog2 = get_hog_features(ch2,
orient,
pix_per_cell,
cell_per_block,
vis=False,
feature_vec=False)
hog3 = get_hog_features(ch3,
orient,
pix_per_cell,
cell_per_block,
vis=False,
feature_vec=False)
# print("HOG 1 {} 2 {} 3 {} ".format(hog1.shape,hog2.shape,hog3.shape));
# print("RAVELD HOG 1 {} 2 {} 3 {} ".format(hog1.ravel().shape,hog2.ravel().shape,hog3.ravel().shape));
hog_features = np.hstack((hog1.ravel(), hog2.ravel(), hog3.ravel()))
# print(" hog_features {}".format(hog_features))
# get the spatial fetures
spatial_features = bin_spatial(img, size=spatial_size)
hist_features = color_hist(img, nbins=hist_bins)
# print(" spatial={} hist={} hog={}\n".format(spatial_features.shape,hist_features.shape,hog_features.shape))
x = np.concatenate((spatial_features.ravel(), hist_features.ravel(),
hog_features.ravel()))
return x
return test_features
def show_hog_features(imgname, title):
img = mpimg.imread(imgname)
print(img.shape)
plt.title("Image")
plt.imshow(img)
plt.show()
img = convert_color(img, conv='RGB2YCrCb')
plt.title("Color Converted")
plt.imshow(img)
plt.show()
print(img.shape)
orient = 9
pix_per_cell = 8
cell_per_block = 2
ch1 = img[:, :, 0]
ch2 = img[:, :, 1]
ch3 = img[:, :, 2]
features1, hog_image1 = get_hog_features(ch1,
orient,
pix_per_cell,
cell_per_block,
vis=True)
features2, hog_image2 = get_hog_features(ch2,
orient,
pix_per_cell,
cell_per_block,
vis=True)
features3, hog_image3 = get_hog_features(ch3,
orient,
pix_per_cell,
cell_per_block,
vis=True)
plt.title(title)
plt.imshow(hog_image1)
plt.show()
plt.imshow(hog_image2)
plt.show()
plt.imshow(hog_image3)
plt.show()
def test_code():
ind = np.random.randint(0,len(cars))
image = mpimg.imread(cars[ind])
features, hog_image = get_hog_features(image, 9, 8,4,True,True)
fig = plt.figure()
plt.subplot(121)
plt.imshow(image, cmap = 'gray' )
plt.title('Example Car Image')
plt.subplot(122)
plot.imshow(hog_image, cmap = 'gray' )
plot.title('HOG Visulization')
def extract_features(imgs,
cspace='RGB',
orient=9,
pix_per_cell=8,
cell_per_block=2,
hog_channel=0):
# Create a list to append feature vectors to
features = []
# Iterate through the list of images
for file in imgs:
# Read in each one by one
image = mpimg.imread(file)
# apply color conversion if other than 'RGB'
feature_image = convert_color(image, cspace)
hog_features = get_hog_features(feature_image, orient, pix_per_cell,
cell_per_block, hog_channel)
# Append the new feature vector to the features list
features.append(hog_features)
# Return list of feature vectors
return features
def find_cars(img, ystart, ystop, scale, svc, X_scaler, orient, pix_per_cell,
cell_per_block, spatial_size, hist_bins):
draw_img = np.copy(img)
img = img.astype(np.float32) / 255
img_tosearch = img[ystart:ystop, :, :]
ctrans_tosearch = convert_color(img_tosearch, conv='RGB2YCrCb')
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) - 1
nyblocks = (ch1.shape[0] // pix_per_cell) - 1
nfeat_per_block = orient * cell_per_block**2
# 64 was the orginal sampling rate, with 8 cells and 8 pix per cell
window = 64
nblocks_per_window = (window // pix_per_cell) - 1
cells_per_step = 3 # Instead of overlap, define how many cells to step
nxsteps = (nxblocks - nblocks_per_window) // cells_per_step
nysteps = (nyblocks - nblocks_per_window) // cells_per_step
# Compute individual channel HOG features for the entire image
hog1 = get_hog_features(ch1,
orient,
pix_per_cell,
cell_per_block,
vis=False,
feature_vec=False)
hog2 = get_hog_features(ch2,
orient,
pix_per_cell,
cell_per_block,
vis=False,
feature_vec=False)
hog3 = get_hog_features(ch3,
orient,
pix_per_cell,
cell_per_block,
vis=False,
feature_vec=False)
bbox = np.empty((1, 2, 2))
print("hog1", hog1.shape, "hog2", hog2.shape, "hog3", hog3.shape)
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
# print("ypos",ypos,"xpos",xpos)
# print("hog_feat",nblocks_per_window, nblocks_per_window)
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))
# print("hog_feat1=",hog_feat1.shape,"hog_feat2=",hog_feat2.shape," hog_feat3=",hog_feat3.shape)
xleft = xpos * pix_per_cell
ytop = ypos * pix_per_cell
# Extract the image patch1
subimg = cv2.resize(
ctrans_tosearch[ytop:ytop + window, xleft:xleft + window],
(64, 64))
# print("subimg shape=",subimg.shape)
# Get color features
spatial_features = bin_spatial(subimg, size=spatial_size)
hist_features = color_hist(subimg, nbins=hist_bins)
# print("find_cars spatial={} hist={} hog={}\n".format(spatial_features.shape,hist_features.shape,hog_features.shape))
# Scale features and make a prediction
x = np.concatenate(
(spatial_features.ravel(), hist_features.ravel(),
hog_features.ravel()))
# print("find_cars x=",x.shape)
test_features = X_scaler.transform(x).reshape(1, -1)
#test_features = X_scaler.transform(np.hstack((shape_feat, hist_feat)).reshape(1, -1))
test_prediction = svc.predict(test_features)
#print("TEst prediction ",test_prediction,xb,yb)
if (test_prediction == 1):
xbox_left = np.int(xleft * scale)
ytop_draw = np.int(ytop * scale)
win_draw = np.int(window * scale)
cv2.rectangle(
draw_img, (xbox_left, ytop_draw + ystart),
(xbox_left + win_draw, ytop_draw + win_draw + ystart),
(0, 0, 255), 6)
v = np.array([[int(xbox_left),
int(ytop_draw + ystart)],
[
int(xbox_left + win_draw),
int(ytop_draw + win_draw + ystart)
]]).reshape((1, 2, 2))
# print("PREDICTION! V shape",v.shape, "box=",bbox.shape)
bbox = np.vstack((bbox, v))
#plt.imshow(draw_img)
#plt.show()
#plt.pause(.001)
bbox = np.delete(bbox, (0), 0) # Remove the first element that I put on
print("find_cars returning bbox shape:", bbox.shape, " values=", bbox)
return draw_img, bbox