Object identifying apparatus

An object identifying apparatus comprising:

• an Electronic Control Unit (ECU) programmed to:
• determine, as a first region in an image, a region in the image having an object that is an identification target;
• determine, as a border of the object that is the identification target, pixels at each of which a gradient of image characteristics amount is equal to or larger than a predetermined amount;
• determine a main border portion that is at least one portion of the border, the determining of the main border portion including:
• (i) determining which angle range among a plurality of angle ranges the direction of the gradient of the image characteristics amount at each pixel belongs to, (ii) calculating a total number of the pixels belonging to each angle range, (iii) determining the angle range in which the total number of the belonging pixels is at a maximum, in order to determine the angle range in which a distribution range of the belonging pixels is at the maximum, and (iv) determining, as the main border portion, the region at which the pixels belonging to the determined angle range exist;
• determine a first parameter representing a length along a side of the first region and a second parameter representing a sum of lengths of the side of the first region onto which the main border portion is projected on when the main border portion is projected onto the side of the first region, wherein the first region is a polygonal shape;
• determine an identification parameter, which is a third parameter, based on comparing the first parameter and the second parameter;
• and
• identify a type of object of the object that is the identification target based on the determined identification parameter.

An object identifying apparatus comprising:

• an Electronic Control Unit (ECU) programmed to:
• determine, as a target region in an image, a region in the image having an object that is an identification target, wherein the target region is a polygonal shape;
• classify a plurality of pixels included in a border of the object in the target region based on a direction or direction ranges of image characteristic gradients at each pixel wherein the direction ranges correspond to a plurality of directions;
• calculate a total number of the plurality of pixels included in the border for the direction or the direction ranges of the image characteristic gradients;
• determine a main border portion corresponding to a region at which the plurality of pixels belonging to the direction or the direction ranges of which the total number of the plurality of pixels is at a maximum exists;
• identify a first parameter representing a length along a side of the target region;
• calculate a second parameter representing a sum of lengths of the side of the target region onto which the main border portion is projected on when the main border portion is projected onto the side of the target region;
• and
• based on the first parameter and the second parameter, identify a type of object of the object that is the identification target.