Open Source CV Image Processing and Drawing Text Line Segments

1. Experiment Objective

Master the basic OpenCV image processing operations (grayscale conversion, binarization, edge detection) and drawing geometric annotations on images (lines, circles, rectangles, ellipses, polygons, text).

2. Experiment Principle

2.1 Keyword Explanation

2.2 Technical Architecture

This experiment is a purely offline Python script, divided into two categories of operations:

• Image processing (grayscale / binarization / edge detection): Performs global operations on pixel values to generate a new image matrix
• Drawing annotations (line / circle / rectangle / ellipse / polygon / text): Modifies pixel values in place on the original image ndarray

Data flow: Local image → cv2.imread() → cv2.cvtColor() color space conversion → cv2.threshold() / cv2.Canny() image processing → cv2.line() / cv2.circle() and other drawing APIs → window display.

3. Experiment Steps

3.1 Hardware Preparation

3.2 Startup Method

Image processing category:

​x
# Grayscale conversion
python3 ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_image_grayscale_conversion.py
​
# Binarization
python3 ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_image_binarization_processing.py
​
# Edge detection
python3 ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_image_edge_detection.py

Drawing annotation category:

xxxxxxxxxx
# Line
python3 ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_draw_line.py
​
# Circle
python3 ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_draw_circle.py
​
# Rectangle
python3 ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_draw_rectangle.py
​
# Ellipse
python3 ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_draw_ellipse.py
​
# Polygon
python3 ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_draw_polygon.py
​
# Text
python3 ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_draw_text.py

3.3 Operation Instructions

The image processing scripts display the processed result side by side next to the original image; the drawing scripts overlay annotations on the original image and display a single window. Press the q key to close all windows.

3.4 Stop Method

Press the q key to close all OpenCV windows and exit the program.

4. Experiment Phenomena

• Master image processing functions such as cv2.cvtColor(), cv2.threshold(), cv2.Canny()
• Master drawing functions such as cv2.line(), cv2.circle(), cv2.rectangle(), cv2.ellipse(), cv2.polylines(), cv2.putText()

Original image: The input image for image processing and drawing; all effects are operated on and overlaid based on this image.

Grayscale image: Uses cv2.cvtColor() to convert the BGR color image into a single-channel grayscale image, where each pixel retains only luminance information.

Binary image: Uses cv2.threshold() to split the grayscale image into pure black (0) and pure white (255) pixels based on a threshold.

Edge detection image: Before Canny edge detection, Gaussian blur is applied for noise reduction, then pixels with sharp luminance changes are extracted to form object contour lines.

Drawing a line: Uses cv2.line() to draw a line segment with a specified start and end point on the image; color and line width can be customized.

Drawing a circle: Uses cv2.circle() to draw a circle outline with a specified center and radius on the image.

Drawing a rectangle: Uses cv2.rectangle() to draw a rectangle border with specified top-left and bottom-right corner coordinates on the image.

Drawing an ellipse: Uses cv2.ellipse() to draw an ellipse arc with a specified center, axis lengths, and rotation angle on the image.

Drawing a polygon: Uses cv2.polylines() to draw a closed polygon formed by sequentially connecting multiple vertex coordinates on the image.

Drawing text: Uses cv2.putText() to overlay text annotations on the image; position, font, size, and color can be customized.

5. Program Analysis

Source code paths:

• Script directory: ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/
• Test image: ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/images/yahboom.png

5.1 Core Logic Analysis

All scripts share a unified load-process-display framework. The core code for image processing and drawing annotations is shown below respectively.

Grayscale + binarization (opencv_image_binarization_processing.py)

xxxxxxxxxx
# Source file: ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_image_binarization_processing.py
import cv2
​
if __name__ == '__main__':
    img = cv2.imread('~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/images/yahboom.png', cv2.IMREAD_UNCHANGED)
    if img is None:
        print("Failed to read the image. Please check whether the images/yahboom.png path is correct")
        exit()
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    ret, thresh1 = cv2.threshold(gray, 180, 255, cv2.THRESH_BINARY_INV)
​
    while True:
        cv2.imshow("yahboom", img)
        cv2.imshow('yahboom_gray', gray)
        cv2.imshow("yahboom_binary", thresh1)
        action = cv2.waitKey(10) & 0xFF
        if action == ord('q') or action == 113:
            break
​
    cv2.destroyAllWindows()

Key logic explanation:

• cv2.cvtColor(img, cv2.COLOR_BGR2GRAY): Converts BGR three channels into a single-channel grayscale, invoking the internal weighted formula
• cv2.threshold(src, thresh, maxval, type): Pixels with a value > thresh(180) are set to maxval(255), otherwise set to 0
• THRESH_BINARY_INV indicates inverted binarization (above the threshold becomes 0), making the dark logo the white foreground
• ret returns the threshold actually used (equal to the passed thresh; it differs only in OTSU adaptive mode)

Canny edge detection (opencv_image_edge_detection.py)

xxxxxxxxxx
# Source file: ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/opencv_image_edge_detection.py
import cv2
​
if __name__ == '__main__':
    img = cv2.imread('~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/images/yahboom.png', cv2.IMREAD_UNCHANGED)
    if img is None:
        print("Failed to read the image. Please check whether the images/yahboom.png path is correct")
        exit()
​
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
    imgG = cv2.GaussianBlur(gray, (3, 3), 0)
    dst = cv2.Canny(imgG, 50, 50)
​
    while True:
        cv2.imshow("yahboom", img)
        cv2.imshow('yahbeeom_gray', gray)
        cv2.imshow("yahboom_canny", dst)
        action = cv2.waitKey(10) & 0xFF
        if action == ord('q') or action == 113:
            break
​
    cv2.destroyAllWindows()

Key logic explanation:

• The image must be converted to grayscale before edge detection (Canny requires single-channel input)
• cv2.GaussianBlur(gray, (3,3), 0): 3×3 Gaussian kernel convolution for noise reduction, suppressing high-frequency noise to avoid false detections
• cv2.Canny(img, 50, 50): threshold1=50 (low threshold; below this value is definitely not an edge), threshold2=50 (high threshold; above this value is definitely an edge). Values in between are judged based on connectivity to strong edges
• Note: Setting both thresholds to the same value 50 limits the algorithm's effectiveness; in practice the low threshold is usually set to 1/2~1/3 of the high threshold

Drawing API summary (opencv_draw_line.py / opencv_draw_circle.py / opencv_draw_rectangle.py / opencv_draw_text.py)

The following shows how each drawing function is called; all of them modify the original image ndarray in place:

xxxxxxxxxx
# Source file: ~/microros_ws/src/microros_v2_opencv_base_pkg/scripts/opencv_image_processing_and_drawing/
# See the comments on each function for the specific source file
​
# Line: start (50,20) → end (20,100), color (255,0,255)=magenta, line width 10
cv2.line(img, (50, 20), (20, 100), (255, 0, 255), 10)
​
# Circle: center (80,80), radius 50, color magenta, line width 10 (negative value fills it solid)
cv2.circle(img, (80, 80), 50, (255, 0, 255), 10)
​
# Rectangle: top-left (50,20) → bottom-right (100,100), color magenta, line width 10
cv2.rectangle(img, (50, 20), (100, 100), (255, 0, 255), 10)
​
# Ellipse: center (80,80), axes (20,50), angle 0, start angle 0°, end angle 360°, color magenta, line width 5
cv2.ellipse(img, (80, 80), (20, 50), 0, 0, 360, (255, 0, 255), 5)
​
# Polygon: vertex array (5 points), isClosed=True, color magenta, line width 5
points = np.array([[120, 50], [40, 140], [120, 70], [110, 110], [50, 50]], np.int32)
cv2.polylines(img, [points], True, (255, 0, 255), 5)
​
# Text: content 'This is Yahboom!', bottom-left coordinate (50,50), font, scale 1.0, color (0,200,0), line width 2
cv2.putText(img, 'This is Yahboom!', (50, 50), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 200, 0), 2)

Common rules for shape drawing:

• The color parameter is a BGR triplet; (255, 0, 255) is magenta
• The thickness parameter: a positive value is the line width (in pixels); a negative value (such as -1) fills the shape solid
• The coordinate origin is the top-left corner of the image, with the x axis pointing right and the y axis pointing down
• All drawing functions return the drawn ndarray (i.e., the same object modified in place)

5.2 Key Parameters

5.3 Node Explanation

This experiment is a pure Python script and does not use any ROS2 nodes.

6. Common Issues