Vision API

This file presents the different functions and Enums Image Processing available for image processing.

These functions are divided in subsections:

  • Pure Image Processing are used to deal with the thresholding, contours detection, ..

  • Workspaces wise section contains functions to extract workspace and deal with the relative position in the workspace

  • The section Show allows to display images

  • Image Editing contains lot of function which can compress images, add text to image, …

Pure image processing

image_functions.threshold_hsv(list_min_hsv, list_max_hsv, reverse_hue=False, use_s_prime=False)

Take BGR image (OpenCV imread result) and return thresholded image according to values on HSV (Hue, Saturation, Value) Pixel will worth 1 if a pixel has a value between min_v and max_v for all channels

Parameters:

  • img (numpy.array) – image BGR if rgb_space = False

  • list_min_hsv (list[int]) – list corresponding to [min_value_H,min_value_S,min_value_V]

  • list_max_hsv (list[int]) – list corresponding to [max_value_H,max_value_S,max_value_V]

  • use_s_prime (bool) – True if you want to use S channel as S’ = S x V else classic

  • reverse_hue (bool) – Useful for Red color cause it is at both extremum

Returns:

threshold image

Return type:

numpy.array

image_functions.debug_threshold_color(color_hsv)

Return masked image to see the effect of color threshold

Parameters:

  • img (numpy.array) – OpenCV image

  • color_hsv (ColorHSV) – Color used for debug

Returns:

Masked image

Return type:

numpy.array

image_functions.morphological_transformations(morpho_type=MorphoType.CLOSE, kernel_shape=(5, 5), kernel_type=KernelType.ELLIPSE)

Take black & white image and apply morphological transformation

Parameters:

  • im_thresh (numpy.array) – Black & White Image

  • morpho_type (MorphoType) – CLOSE/OPEN/ERODE/DILATE => See on OpenCV/Google if you do not know these words

  • kernel_shape (tuple[float]) – tuple corresponding to the size of the kernel

  • kernel_type (KernelType) – RECT/ELLIPSE/CROSS => see on OpenCV

Returns:

image after processing

Return type:

numpy.array

image_functions.get_contour_barycenter()

Return barycenter of an OpenCV Contour

Parameters:

contour (OpenCV Contour)

Returns:

Barycenter X, Barycenter Y

Return type:

int, int

image_functions.get_contour_angle()

Return orientation of a contour according to the smallest side in order to be well oriented for gripper

Parameters:

contour (OpenCV Contour) – contour

Returns:

Angle in radians

Return type:

float

image_functions.biggest_contour_finder()
image_functions.biggest_contours_finder(nb_contours_max=3)

Function to find the biggest contour in an binary image

Parameters:

  • img (numpy.array) – Binary Image

  • nb_contours_max (int) – maximal number of contours which will be returned

Returns:

biggest contours found

Return type:

list[OpenCV Contour]

image_functions.draw_contours(contours)

Draw a list of contour on an image and return the drawing image

Parameters:

  • img (numpy.array) – Image

  • contours (list[OpenCV Contour]) – contours list

Returns:

Image with drawing

Return type:

numpy.array

image_functions.draw_barycenter(cx, cy, color=(255, 0, 255), marker_size=10, thickness=2)

Draw a barycenter marker on an image and return the drawing image

Parameters:

  • img (numpy.array) – Image

  • cx (int) – Barycenter X

  • cy (int) – Barycenter Y

  • color (list[uint8, utin8, uint8]) – (R, G, B) colors of the marker

  • marker_size (int) – size of the marker

  • thickness (int) – thickness of the marker

Returns:

Image with drawing

Return type:

numpy.array

image_functions.draw_angle(cx, cy, angle, color=(0, 0, 255), arrow_length=30, thickness=2)

Draw an arrow that represents the orientation of an object on an image and return the drawing image

Parameters:

  • img (numpy.array) – Image

  • cx (int) – Barycenter X

  • cy (int) – Barycenter Y

  • angle – angle to display

  • color (list[uint8, utin8, uint8]) – (R, G, B) colors of the marker

  • arrow_length (int) – length of the arrow marker

  • thickness (int) – thickness of the arrow marker

Returns:

Image with drawing

Return type:

numpy.array

Workspaces wise

image_functions.extract_img_workspace(workspace_ratio=1.0)

Extract working area from an image thanks to 4 Niryo’s markers

Parameters:

  • img (numpy.array) – OpenCV image which contain 4 Niryo’s markers

  • workspace_ratio (float) – Ratio between the width and the height of the area represented by the markers

Returns:

extracted and warped working area image

Return type:

numpy.array

image_functions.debug_markers(workspace_ratio=1.0)

Display detected markers on an image

Parameters:

  • img (numpy.array) – OpenCV image which contain Niryo’s markers

  • workspace_ratio (float) – Ratio between the width and the height of the area represented by the markers

Returns:

(status, annotated image)

Return type:

numpy.array

image_functions.relative_pos_from_pixels(x_pixels, y_pixels)

Transform a pixels position to a relative position

Parameters:

  • img (numpy.array) – Image where the object is detected

  • x_pixels (int) – coordinate X

  • y_pixels (int) – coordinate Y

Returns:

X relative, Y relative

Return type:

float, float

Show

image_functions.show_img_and_check_close(img)

Display an image and check whether the user want to close

Parameters:

  • window_name (str) – window’s name

  • img (numpy.array) – Image

Returns:

boolean indicating if the user wanted to leave

Return type:

bool

image_functions.show_img(img, wait_ms=1)

Display an image during a certain time

Parameters:

  • window_name (str) – window’s name

  • img (numpy.array) – Image

  • wait_ms (int) – Wait time in milliseconds

Returns:

value of the key pressed during the display

Return type:

int

image_functions.show_img_and_wait_close(img)

Display an image and wait that the user close it

Parameters:

  • window_name (str) – window’s name

  • img (numpy.array) – Image

Returns:

None

Return type:

None

Image Editing

image_functions.compress_image(quality=90)

Compress OpenCV image

Parameters:

  • img (numpy.array) – OpenCV Image

  • quality (int) – integer between 1 - 100. The higher it is, the less information will be lost, but the heavier the compressed image will be

Returns:

status & string representing compressed image

Return type:

bool, str

image_functions.uncompress_image()

Take a compressed img and return an OpenCV image

Parameters:

compressed_image (str) – compressed image

Returns:

OpenCV image

Return type:

numpy.array

image_functions.add_annotation_to_image(text, write_on_top=True)

Add Annotation to an image

Parameters:

  • img (numpy.array) – Image

  • text (str) – text string

  • write_on_top (bool) – if you write the text on top

Returns:

img with text written on it

Return type:

numpy.array

image_functions.undistort_image(mtx, dist)

Use camera intrinsics to undistort raw image

Parameters:

  • img (numpy.array) – Raw Image

  • mtx (list[list[float]]) – Camera Intrinsics matrix

  • dist (list[list[float]]) – Distortion Coefficient

Returns:

Undistorted image

Return type:

numpy.array

image_functions.resize_img(width=None, height=None, inter=3)

Resize an image. The user should precise only width or height if he wants to keep image’s ratio

Parameters:

  • img (numpy.array) – OpenCV Image

  • width (int) – Target Width

  • height (int) – Target Height

  • inter (long) – OpenCV interpolation flag

Returns:

resized image

Return type:

numpy.array

image_functions.concat_imgs(axis=1)

Concat multiple images along 1 axis

Parameters:

  • tuple_imgs (tuple[numpy.array]) – tuple of images

  • axis (int) – 0 means vertically and 1 means horizontally

Returns:

Concat image

Return type:

numpy.array

Enums Image Processing

Enums are used to pass specific parameters to functions.

List of enums:

class pyniryo.vision.enums.ColorHSV(*values)

MIN HSV, MAX HSV, Invert Hue (bool)

BLUE = ([90, 50, 85], [125, 255, 255], False)
RED = ([15, 80, 75], [170, 255, 255], True)
GREEN = ([40, 60, 75], [85, 255, 255], False)
ANY = ([0, 50, 100], [179, 255, 255], False)
class pyniryo.vision.enums.ColorHSVPrime(*values)

MIN HSV, MAX HSV, Invert Hue (bool)

BLUE = ([90, 70, 100], [115, 255, 255], False)
RED = ([15, 70, 100], [170, 255, 255], True)
GREEN = ([40, 70, 100], [85, 255, 255], False)
ANY = ([0, 70, 140], [179, 255, 255], False)
class pyniryo.vision.enums.ObjectType(*values)
SQUARE = 4
TRIANGLE = 3
CIRCLE = -1
ANY = 0
class pyniryo.vision.enums.MorphoType(*values)
ERODE = 0
DILATE = 1
OPEN = 2
CLOSE = 3
class pyniryo.vision.enums.KernelType(*values)
RECT = 0
ELLIPSE = 2
CROSS = 1