三維光聲斷層掃描是一種兼具高對比與成像深度的非侵入式影像技術，特別適用於血管結構的可視化。然而，光聲訊號的寬頻特性受限於超音波換能器的靈敏度以及聲波在組織中的頻率依賴性衰減，使得同時解析大小尺度結構面臨挑戰。本研究提出一套應用於三維 光聲斷層掃描的頻帶分離重建技術，透過球形陣列換能器將訊號分離為高頻與低頻子頻帶，分別進行重建與影像強度歸一化後再合併，以增強多尺度血管結構的呈現能力。為驗證此方法之有效性，本研究設計並製作多種尺寸的墨水混合Agar而成的「Agar tube」，進行多尺度結構模擬驗證，透過對比雜訊比（CNR）量化分析顯示，影像品質在整體上獲得顯著提升，惟增益幅度受原始訊號強度影響而有所不同。進一步採用不同中心頻率的超音波換能器（4 MHz、7 MHz、10 MHz）進行實測，顯示低中心頻率的換能器對大結構具更明顯的增強效果，而高頻換能器能更進一步提升小結構解析力，同時也能強化大結構的表現。此外，臨床手指樣本實測結果亦證實頻帶分離重建可同時提升大小血管的對比度與可視化表現。本研究驗證了頻帶分離重建法於三維光聲斷層掃描在臨床多尺度血管成像中的應用潛力，對血管結構與相關疾病的精確診斷提供更多可能性。

Three-dimensional Optoacoustic tomography (3D-OAT) is a non-invasive imaging modality that offers high optical contrast and substantial imaging depth, making it particularly suitable for vascular visualization. However, the broadband nature of optoacoustic signals is limited by the sensitivity of ultrasound transducers and frequency-dependent acoustic attenuation in tissue, posing challenges for simultaneously resolving both large and small structures. In this study, we propose a frequency band-separation reconstruction method for 3D-OAT. By separating the received signals into low- and high-bands using a spherical array transducer, each band is independently reconstructed and normalized before being combined to enhance multiscale vascular features. Validation using agar tube phantoms with different diameters demonstrated improved image quality, as confirmed by quantitative contrast-to-noise ratio (CNR) analysis, although the degree of enhancement varied with the original signal strength. Further testing with transducers of different center frequencies (4 MHz, 7 MHz, 10 MHz) showed that low-frequency transducers provide more pronounced enhancement for large structures, while high-frequency transducers further improve the resolution of small features. Finally, in vivo imaging of human fingers confirmed the method’s ability to improve contrast and visibility of both large and small vessels. This study demonstrates the clinical potential of band-separated reconstruction for accurate multiscale vascular imaging in 3D-OAT and supports its potential utility in improving the accuracy of vascular analysis and facilitating future diagnostic applications.