alexhex7 / object-position-label

object_position_label

• Based on Blog python-opencv-鼠标事件-画框圈定目标.

Statement

• Crop the object in images by hand.
• Record the position(top left point and bottom right point) of this object.

Details

Input an image into the function get_position(image, title), you can get a window of this image.

• For the first time of clicking the left button of your mouse, you can pin the top left point of the box.
• For the second time of clicking the left button, you can pin the bottom right point of the box.
• For the third time of clicking the left button (which means make sure to crop), this function will return thest two points' location (x1, y1), (x2, y2).
• Before make sure to crop the box, you can click right button of your mouse to cancle the pins of the box.

And what the demo does is that

• invoke the function get_position(image, title) to get two points' location.
• Crop the box, save the box to dst_dir .
• Record these two points' location to record_file.
• Move the input image to done_dir.

Environment

• python3.5.2
• opencv3.2.0