Quality and Structure Detection Algorithm of Chocolate with X-ray Computed Tomography

Yoga Arif Firmansyah, Bayu Nugraha, S.T.P., M.Sc., Ph.D.; Dr. Arifin Dwi Saputro, S.T.P., M.Sc.

2025 | Tesis | S2 Mekanisasi/Teknik Pertanian

Cokelat praline adalah salah satu produk konfeksioneri yang sangat populer, ditunjukkan dengan lapisan cangkang cokelat yang diisi dengan berbagai isian. Namun, kombinasi isian tersebut berkontribusi besar terhadap stabilitas struktur internal yang memengaruhi tekstur dan kualitas produk. Penelitian ini bertujuan untuk mengembangkan metode berbasis X-ray Computed Tomography yang dapat memvisualisasikan dan menganalisis struktur mikro internal cokelat praline. Ketiga cokelat berbeda (dark-, milk-, dan white chocolate) yang mengandung dua jenis isian: berbasis air (selai nanas) dan berbasis lemak (selai kacang) dievaluasi dengan membandingkan tiga metode yang berbeda (Otsu segmentation-based, histogram-based, dan adaptive segmentation-based) untuk mengidentifikasi struktur cokelat praline. Metode berbasis histogram (30–255) terbukti paling efektif untuk segmentasi secara keseluruhan struktur cokelat praline, sementara metode Otsu optimal untuk deteksi persebaran gelembung udara. Berbagai pendekatan diterapkan untuk mengkarakterisasi isian, seperti entropy-based segmentation untuk selai nanas, dan adaptive histogram equalization untuk selai kacang dengan mempertimbangkan karakteristiknya. Hasil penelitian menunjukkan bahwa karakteristik material cokelat secara umum menurun dari dark chocolate diikuti oleh milk chocolate dan white chocolate. Namun, milk chocolate menawarkan volume gelembung udara terendah (0,18±0,05%) daripada dark chocolate (0,30±0,01%) dan white chocolate (0,63±0,04%) karena profil viskositas yang lebih baik. Selai nanas (kadar air tinggi) menyebabkan perubahan struktur yang lebih signifikan daripada selai kacang. Korelasi densitas R² antara hasil digital dan eksperimen sangat kuat (0,9255) tetapi bervariasi pada isian (selai nanas: 0,8242; selai kacang: 0,2269) yang dipicu oleh karakteristik isian. Kadar air dan warna permukaan meningkat seiring penyimpanan berkaitan dengan volume gelembung udara awal yang lebih rendah, memicu efek bloom yang dapat mengurangi kualitas.

Praline chocolate is a widely popular confectionery characterized by a chocolate shell filled with various fillings. However, the combination of different filling types leads to internal structural stability, affecting texture and product quality. This study aimed to develop a method based on X-ray computed tomography to visualize and analyze the internal microstructure of praline chocolates. These three different chocolates (dark, milk, and white), which contained two types of fillings: water-based (pineapple jam) and fat-based (peanut butter) were evaluated by comparing three different methods (Otsu segmentation-based, histogram-based, and adaptive segmentation-based) to identify the structure of praline chocolate. The histogram-based method (30–255) proved most effective for general segmentation of praline chocolate structure, while the Otsu method was optimal for air bubble detection. The different approaches were implemented to characterize the filling, such as entropy-based segmentation for pineapple jam, and adaptive histogram equalization for peanut butter by considering the characteristic. The results showed that material characteristic of chocolate generally decreased from dark followed by milk and white chocolate. However, milk chocolate offers the lowest air bubble volume (0.18±0.05%) than dark (0.30±0.01%) and white (0.63±0.04%) due to the good viscosity profile. Pineapple jam (high moisture) caused more significant structural changes than peanut butter. Density correlation of R² between digital and experimental results was strong (0.9255) but varied in correlation of filling segmentation (pineapple jam: 0.8242; peanut butter: 0.2269) corresponding to filling characteristics. Moisture content and surface colour increased during storage along with lower air bubble volume, triggering bloom effect that reduce the quality.

Kata Kunci : chocolate, praline, internal structure, x-ray computed tomography, image processing