Abstract: In order to promote the mechanized harvesting of bulk tea and improve the harvesting efficiency and quality of fresh leaves of bulk tea, a fusion 2D-LiDAR and Attitude and heading reference system（AHRS）was proposed in view of the fact that the sensing sensor of the current profiling tea picker is easily affected by contact force, natural light or the gap between the leaves of the tea canopy. Based on the estimation method of profiling distance of cutting knife of tea picker, an accuracy verification test bench and an automatic profiling tea picker were designed and developed, indoor and field experiments were carried out respectively. The tea picker used 2D-LiDAR to measure the profiling distance at first. In order to improve the ranging accuracy and real-time performance, combined with the acceleration sensed by AHRS, a fusion of 2D-LiDAR ranging and acceleration (FLRA) was proposed. The algorithm accuracy verification platform and method were developed to verify the effectiveness of the algorithm. The indoor test results show that the mean value of the ranging error of the profiling distance before the algorithm processing was 36.53 mm, and the standard deviation was 23.21 mm. After the algorithm processing, the mean value of the profiling distance estimation error was 8.56 mm, and the standard deviation was 6.31 mm, which improved the accuracy and real-time performance of profiling distance ranging. Field tests show that the harvesting efficiency reached 180-210 kg·h^-1. The average picking rate of young shoots on the canopy covered by cutter was 92.38%. The integrity rate of bud and leaf was 85.34% and the impurity rate was 4.93%. The young shoots better than one bud and three leaves accounted for 90.72%, which meets the technical standards of bulk tea machine picking and the requirements of subsequent processing technology. Compared with the traditional ultrasonic sensing automatic profiling tea picker, the harvesting effect of bulk tea fresh leaves was improved.

Key words: tea picker, profiling distance, 2D-LiDAR, AHRS, fusion algorithm, accuracy verification