Abstract: To address the problems of low efficiency, poor accuracy, and high missed-detection rates in traditional methods for detecting quality defects such as damage and liquid leakage during the production of flavor-capsule slim cigarettes, this paper proposes a defect detection method for flavor capsules in slim cigarettes based on Dual-optical-path symmetric photon imaging. An oblique-angle illumination and dual-optical-path complementary strategy were adopted to design a dual-channel symmetric X-ray detection system, resolving the misjudgment problem caused by capsule overlap in single-optical-path detection. A defect recognition algorithm based on a convolutional neural network was constructed to achieve row-by-row automated detection. Finally, a "detection–conveying–rejection" timing control model was established to ensure precise rejection under high-speed production conditions. The results showed that: 1) The system achieves a recognition accuracy of 99.90% and a rejection accuracy of 100.00%, with a detection speed of 300 packs/min (6,000 cigarettes/min), and can simultaneously identify multiple quality defects including missing capsules, inverted capsules, missing cigarettes, and empty packs. 2) Timeliness verification shows that hand-pinched-damage-type flavor capsule defect samples can be detected immediately, while needle-puncture-damage-type flavor capsule defect samples can be stably detected within 40 minutes. 3) In on-site application testing, continuous detection of 136,200 packs of single-flavor-capsule slim cigarette samples (2.8 mm in diameter) showed no misjudgments, verifying the reliability and stability of the system. This research provides a new technical approach for quality control of flavor-capsule cigarettes.