Dimensional measurement

New in version 1.1

In this application, we use 2 steps:

1. Make calibration
2. Measuring the dimensons of object

During the second step, an image will be passed through this pipeline

[Segmentation] --> [Measurement] --> [Overlay] --> [Saving results]

1. Prepare your dataset

# your dataset structure should be like this
data/
    calib.jpeg
    -test/
        -*.jpg

Notes

The calib.jpeg is the image of object using for calibration (pixel/mm). It's recommended that the calibration object should have the same depth as objects will be used for test.

2. Usages

Calibration

Go to /projects/calculate_solution_opencv

Execute this command:

python calib.py --calib_file_path "<PATH_TO_CALIB_FILE>" --output "<PATH_TO_OUTPUT_DIR>"

Measurement

Go to /projects/calculate_solution_opencv

Execute this command:

python predict.py --img_path "<PATH_TO_IMAGE_FILE>" --options "<PATH_TO_OPT_FILE>" --output "<PATH_TO_OUTPUT_DIR>"

3. Tips for improving performance

1. Use the calibration object with same depth as measuring object (Could be a 3D printed cube with specified width and height)
2. Align the measuring object to be straight with camera/view
3. Align camera position for avoiding distortion