Skip to content

Options for the use of computer vision in the framework of the Cup of the Governor of the region in computer vision

License

Notifications You must be signed in to change notification settings

D1ffic00lt/computer-vision-cup

Repository files navigation

Computer Vision

Options for the use of computer vision in the framework of the Cup of the Governor of the region in computer vision

Tasks

Nut defect detection (product quality control)

The video shows the movement of nuts along the conveyor. The nuts move from the lower border of the frame to the upper one. Among the nuts there are defective ones, they are deformed and have distorted contours. For each nut captured on video, it is necessary to determine compliance with quality standards.

image

The 97% accuracy solution is presented in folder folder .

Logo detection

The image shows the logo. Your task is to write a function that determines in which part of the image the logo is located, and what it is called. Each image shows one of the five logos.

Your function should return the name of the logo and the coordinates of the border bounding the logo. Correctly detected and recognized are those logos for which the name is correctly indicated and the IoU is greater than 0.5.

image

The 98% accuracy solution is presented in folder.

Classification of road signs

The images show traffic signs. There is exactly one sign on each image. Your task is to write a function that determines which of the eight characters in the image.

image

The 100% accuracy solution is presented in folder folder.

Traffic light detection

The images show traffic lights. Your task is to write a function that determines in which part of the image the traffic light is located. There is only one traffic light in each image. The function should return the coordinates of the frame bounding the traffic light. Traffic lights for which IoU is greater than 0.5 are considered to be correctly detected.

image

The 100% accuracy solution is presented in folder folder.

Hand detection

Your task is to write a program for gesture recognition and integrate it into aikar. Possible gestures: the thumb up indicates forward movement, the thumb down indicates backward movement, the index finger to the left and to the right indicate turns to the left and to the right, respectively, and the clenched fist indicates a stop.

The solution is presented in folder folder.

Eyes detection

Your task is to write a program for blinking recognition and integrate it into aikar. A single blink indicates a turn of the wheels by 70, 90 and 110 degrees, a double blink indicates a stop and start of movement.

The solution is presented in folder folder.

Aikar on the road (the task of the cup final)

You need to "teach" akar to drive along the highway in a figure of eight, drive through intersections, stop at traffic lights, turn left and let pedestrians pass.

The solution that took the first place is presented in the folder.