Ilham Ari Elbaith Zaeni, Siti Sendari, Dyah Lestari, Yogi Dwi Mahandi, Bintang Wahyu Utama
A mobile robot should have the ability to sense the environmental conditions around it using sensors that attached to the robot. A camera can be used in the robot to detect or recognize the object. The camera must be robust to the environmental changes. Multi-object detection is the robot processes to detect and to recognize the multi-object as human perceive. In the object detecting process, the illumination changes were very influential. Therefore, to avoid this effect, we propose the multi-object detection that robust to the illumination changes. The multi-object detection was conducted using hue, saturation, and lightness of the camera image. Contrast Limited Adaptive Histogram Equalization (CLAHE) was applied to the lightness channel to reduce the effect of illumination. The experiment result shows that accuracies to detect object were 90.63%, 93.55%, 64.29%, and 100.00% for the object with red, green, blue, and yellow color, respectively. © 2018 IEEE.
State University of Malang, Department of Electrical Engineering, Malang, Indonesia; Intitut Teknologi Sepuluh Nopember, Department of Electrical Engineering, Surabaya, Indonesia
