[Teknologi robot yang berkembang pesat sekarang ini menyebabkan
naiknya kebutuhan akan teknik pengolahan citra dan teknik memprediksi gerak
objek. Salah satu cara yang paling umum dalam mendeteksi suatu benda adalah
dengan menggunakan filter warna dan sensor untuk mendeteksi jarak. Namun,
kelemahan metode tersebut terletak pada rentannya filter warna terhadap
gangguan luar, serta lambatnya sensor dalam mengukur jarak. Dalam penelitian
ini, digunakan sebuah Kalman Filter yang berfungsi untuk mengurangi efek
gangguan dalam filter warna, serta membuat program tetap dapat memprediksi
lokasi bola ketika bola hilang dari pandangan. Sebuah algoritma yang dapat
memprediksi jarak dan koordinat bola tanpa menggunakan sensor juga
dikembangkan demi mempercepat waktu proses. Setelah lokasi bola diketahui,
sebuah algoritma untuk memprediksi koordinat jatuh bola juga dikembangkan.
Koordinat jatuh ini nantinya dapat dimasukkan ke sistem penggerak robot yang
mampu mengarahkan robot ke lokasi jatuhnya benda.;Image processing and trajectory prediction play an important role in today?s
robotic technology and its applications are limitless. The most common method of
detecting an object is through the use of colour filter and a sensor to measure its
distance. However, most of the time, colour filter is very vulnerable to noise and
the robot would consume a huge chunk of processing time using distance sensor.
In this paper, a Kalman Filter is developed base on a constant velocity model.
This Kalman Filter will provide a better estimator on the object?s position and
estimate the position of the object when there are occlusions. In addition to the
Kalman Filter, a novel sensorless approach to measure an object?s distance is also
developed in order to improve processing time. After the coordinates of the object
are obtained, an algorithm that could predict its landing coordinates is also
developed by using MATLAB. This landing coordinates could then be fed into a
motor system in a robot which will then move towards the predicted coordinates., Image processing and trajectory prediction play an important role in today’s
robotic technology and its applications are limitless. The most common method of
detecting an object is through the use of colour filter and a sensor to measure its
distance. However, most of the time, colour filter is very vulnerable to noise and
the robot would consume a huge chunk of processing time using distance sensor.
In this paper, a Kalman Filter is developed base on a constant velocity model.
This Kalman Filter will provide a better estimator on the object’s position and
estimate the position of the object when there are occlusions. In addition to the
Kalman Filter, a novel sensorless approach to measure an object’s distance is also
developed in order to improve processing time. After the coordinates of the object
are obtained, an algorithm that could predict its landing coordinates is also
developed by using MATLAB. This landing coordinates could then be fed into a
motor system in a robot which will then move towards the predicted coordinates.]
S62275-Hoo Kevin Wijaya.pdf :: Unduh