Objective To develop a novel recognition algorithm that can assist physicians in locating pulmonary nodules. MethodsSixteen patients with pulmonary nodules who underwent thoracoscopic surgery at the Department of Thoracic Surgery, the Affiliated Drum Tower Hospital of Nanjing University Medical School were enrolled, including 9 males and 7 females with an average age of (54.9±14.9) years. Chest surface exploration data of 60 frames per second and1 920×1 080 resolution were collected from patients, and frame images were saved at regular intervals for block processing. An algorithm database for lung nodule recognition using the above data was built. ResultsIn the optimized multi topology convolutional network model, the test results showed an accurate recognition rate of 94.39%. Furthermore, by integrating micro-variation amplification technology into the convolutional network model, the accuracy of identifying lung nodules was improved to 96.90%. By comprehensively evaluating the performance of these two models, the overall recognition accuracy reached 95.59%. Based on this, we infered that the proposed network model was suitable for the recognition task of lung nodules, and the convolutional network incorporating micro-variation amplification technology performs better in accuracy. Conclusion Compared with traditional methods, our proposed technique can significantly improve the accuracy of lung nodule identification and localization, and help surgeons locate lung nodules during thoracoscopic surgery.