Removing temperature drift for bee colony weight measurements based on linear regression model and Kalman filter

In precision beekeeping, bee colony weight is an important indicator to monitor the behaviours such as foraging and swarming. However, ambient temperature variations can greatly affect the measured values. In this paper, a combined method with linear regression model and Kalman filter is proposed to reduce the influence of ambient temperature. Monitoring data is collected to validate the effectiveness of the proposed method. Moreover, methods that solely rely on the statistic model or Kalman filter are investigated as a comparison with the proposed method. The compensation results indicate that the proposed method outperforms the two others, and is feasible and effective in temperature drift removal. The mean absolute errors can be decreased over 40% from that before removal during periods of no honeycombs, the coefficient of variations can also be reduced by over 40% and 5% respectively during the periods of no bees and bees in the beehives. As the proposed method can improve the reading accuracy of bee colony weight, it has potential to benefit the precision beekeeping and basic research on bee activities.& COPY; 2023 IAgrE. Published by Elsevier Ltd. All rights reserved.

Automated summary

In this paper, a combined method of linear regression model and Kalman filter is proposed to reduce the influence of ambient temperature on bee colony weight monitoring. The effectiveness of the proposed method is validated through monitoring data and compared with methods solely relying on statistic model or Kalman filter. The results show that the proposed method outperforms the other two methods in temperature drift removal, with over 40% reduction in mean absolute errors and variations during different periods.