1. RoboCup 2016 - KSL Design and implementation of vision and image processing core Academic Supervisor: Dr. Kolberg Eli Mentors: Dr. Abramov Benjamin & Mr. Amsalem Rafi Hen Shoob Assaf Rabinowitz

2. Vision Team Goals  The Vision module is responsible for image processing. The main goal is to detect meaningful objects - ball, goal and white lines. The implementation uses some functions from the OpenCV image processing library [1].  In the 1 st semester, we were responsible of Goal Detection.  Goal Detection is mandatory input for the other cores of the robot, such as the brain and the localization.  The localization output of where the robot is located in the field, and the brain decision of the kick direction, are based on if and where the goal is located in relation to the robot. [1] G. Bradski and A. Kaehler, Learning OpenCV: Computer Vision with the OpenCV Library, O’Reilly Media, October, 2008.

3. Calibration Tool  First, we were needed to be able to distinguish green (field) and white (goal posts) objects from other objects in a given image. We need to configure the HSV ranges of white and green [2]. In order to do so we developed a calibration tool.  The calibration tool adjusts our image-processing to the current environment colors. The robot defines the green and the white color spectrum.  The user clicks on the green\white pixels in the original image and the tool saves the configuration for each color. [2] R. Gonzalez and R. Woods, Digital Image Processing, Third Edition, Pearson Education, 2008.

4. Goal Detection Algorithm  Input: Raw image from the robot’s camera.  Output: Data structure that contains information about the goal, such as is the goal detected (full\partial detection), the center of the goal, etc.

5. Goal Detection Algorithm – cont.  The algorithm is based on finding objects in the input image, which are suspected as the goal’s posts, and selecting the most relevant ones.  Object is suspected as a post if it satisfies the following terms:  White: First, we use a simple threshold function on the HSV transform of the given image. We get a B&W image, in which only white objects are white.  Vertical: We perform a vertical erosion algorithm on the given image to remove any horizontal white objects from the image. Only vertical white objects are left.  Rectangle-shaped: We surround all these objects with minimum area rectangles. We check the ratio between the output rectangle and the white-object area, and we eliminate any rectangle that does not satisfy the threshold ratio.  Straight-angled: From the robot’s eyes, the posts are orthogonal to the field’s plane. We check that the rectangles angle is close to zero.  Inter-edge ratio: We eliminate any rectangle that does not characterized by long vertical edge and a short horizontal edge.  Goal Detection Video Goal Detection Video

6. Targets  Integrate our code into the system and maintenance.  Develop a robust algorithm for calculating the distance to all detected objects such as the ball, the goal and the white lines on the field.  Develop a goalie system, including brain (i.e. FSM), ball movement algorithm, etc.  Create an interface between the vision core and the localization core.