Distance estimation of colored objects in image
Arif Rahman, Nuryono Satya Widodo · AIP conference proceedings · 2016
Detecting objects with a specific color is one of the topics in computer vision. An object captured by the camera will be represented as an image. Closer objects will appear larger in the image and vice versa; however, size changes do not give the actual distance of the object to the camera. This paper proposes a method to estimate the distance between the colored object to the camera by processing the image captured. First, image color filtering based on HSV color model performed and then to detect the colored object in the image, object detection applied to labeling nearest eight connected components with the same value with Moore-neighborhood tracing algorithm. The distance of the object estimates with the single point projection principle of the object height in the image, represented by the bounding rectangle, and the actual object’s height. An experiment conducted using objects with different color and object distance variance in centimeters. Results show that the method can estimate the real distance of the colored object from the camera.