my idea is; after taking the image from the camera, i make edge detection filtering to view the edges then divide it into parts "arrays" and start to compare each part to the arrow,is that right?
note: the environment the camera capture in, may have many shapes and colors but not an arrow.
try starting at both ends of the line and try to find 2 lines going off the tip at roughly 45 degrees. then find the end of each 45 degree line and see if they are about same to each other, unless the arrow is draw then it wont be exatly right. it would be probably simliar to getting the line jsut starting at the end and going at an angle.
i find something called correlation method for image processing and i dont know if its useful to search a known shape in an image and catch the hole shape using it. because its my first time to deal with image processing any notes about correlation method can be useful.
i can use the shape i want and not forced for a special shape so i choose the left and right arrows here.