Penelitian ini bertujuan untuk mengimplementasikan metode background subtraction pada sistem analisis hasil gerakan release latihan atlet lempar lembing. Implementasi metode background subtraction pada sistem ini berbasis pengolahan citra digital. Data yang diproses adalah video atlet yang berawal dari ketika lari dijalur membawa lembing sampai selesai pelepasan lembing dan akhirnya lembing tertancap. Pengambilan data dilakukan dengan kamera statis yang posisinya telah ditentukan. Jumlah sampel sebanyak 30 video atlet lempar lembing yang berformat .avi. Proses ini diawali dengan pengolahan citra yaitu standarisasi dimensi, background subtraction, min filter, pemilihan foreground, thresholding, dilasi dan erosi. Hasil pengolahan citra digunakan untuk melakukan deteksi objek yaitu berupa atlet, yang kemudian dijadikan masukan dalam menghasilkan keluaran sistem. Pada pengujian sistem didapatkan bahwa sistem telah berhasil mendeteksi atlet pada semua data sampel. Dari pengujian deteksi atlet dihasilkan data tinggi gerakan atlet ketika beberapa saat setelah release  dan sudut lembing ketika release. Hasil dari sistem disimpan kedalam database yang berisi beberapa tabel, yaitu tanggal analisis, jenis kelamin, nama atlet, jauh lemparan, tinggi gerakan atlet ketika release dan sudut pelepasan lembing.

Kata Kunci : lempar lembing, jauh lemparan lembing, sudut pelepasan, tinggi atlet, delphi7, background subtraction.

This study aims to implement the background subtraction method in the analysis system resulting from release movement of javelin throwing athletes. Implementation of background subtraction method on this system based on digital image processing. The data processed is the video of athletes from when running along with the javelin until the javelin is released and finally the javel is stuck. Data retrieval is done by static cameras whose positions have been determined. The number of samples is 30 videos of javelin athletes training in the .avi format. This process begins with image processing, namely standardization of dimensions, background subtraction, min filter, foreground selection, thresholding, dilation and erosion. The results of image processing are used to detect objects in the form of athletes, which are then used as input in generating system output. In testing the system it was found that the system had successfully detected athletes in all sample data. From the detection test the athlete produced athlete's movement height when shortly after released and javelin angle when release. The results of the system are stored in the database contains several table, that is analysis time, athlete's name, gender, throw distance, athlete's movement height when released and javelin release angle.

Keywords : javelin throw, far javelin throw, release angle, athlete's height, delphi 7, background subtraction.