Klasifikasi Kualitas Sinyal Radio FM Berbasis Pengolahan Citra Spektrum pada Berbagai Titik Observasi

Abstract

The quality of FM radio broadcast reception is strongly affected by radio wave propagation characteristics, such as distance, geographical contour, building density, and vegetation, while conventional evaluation methods using a spectrum analyzer are relatively expensive and inflexible. This study aims to develop a simple and economical method for evaluating FM radio signal quality based on spectrum image processing captured through GQRX software and an RTL-SDR Software Defined Radio module. Data were collected at 20 observation points spread across four cardinal directions (North, East, South, West) with five distance variations (2, 4, 6, 8, and 10 km) from the central reference point at the Great Mosque of Palembang, producing 1,424 spectrum images. Each image was processed through spectrum panel detection, power-axis scale reading using Optical Character Recognition (OCR), signal curve extraction, Gaussian filter smoothing, and signal peak detection to calculate the Signal-to-Noise Ratio (SNR), which was then classified according to the TIPHON standard. The results show that the average SNR ranged from 27.06–36.90 dB, with a fluctuating and non-linear distribution pattern with respect to distance. Of the 20 observation points, 18 points obtained a dominant Outstanding category and 2 points obtained an Excellent category, with none falling into the Good, Fair, or Bad categories. These results demonstrate that the RTL-SDR and GQRX-based spectrum image processing approach is effective, economical, and replicable as an FM radio signal quality evaluation method.