The mainstream tracking-by-detection paradigm for multi-object tracking generally conducts detection first, followed by Re-IDentification (Re-ID) and motion estimation. The associations between the predicted boxes and existing tracks are then performed via visual and motion association. However, challenges such as irregular motion patterns, similar appearances, and frequent occlusions often arise, making object tracking a nontrivial task. In this paper, we propose a multi-object tracker based on Spatio-TemporAl Topological (STAT) constraints to address the above issues. More specifically, we design the Feature Adaptive Association Module (FAAM) to establish the association between motion and appearance regionally, completing a complementary combination of appearance and motion features. Among these, the Appearance Feature Update Module (AFUM) is proposed to manage the appearance updates of tracked objects by imposing constraints based on the spatial locations and the degree of object occlusion, while temporal consistency is adopted to smooth the appearance states of tracks to mitigate the accumulation of appearance noise. Moreover, the Robust Motion Tracking Module (RMTM) is established to reduce the impact of irregular motions and certain unreliable detection results. The proposed module includes a higher weighted momentum term to accommodate the excessive motion amplitude and considers low-confidence boxes accompanied by the stage-wise association strategy for high-confidence boxes. Extensive experiments on DanceTrack and benchmark MOT datasets verify the effectiveness of our STAT tracker, especially the state-of-the-art results on DanceTrack, which is characterized by irregular motion and indistinguishable appearance attributes.
In intensive aquaculture systems, accurate water temperature prediction is crucial for aquaculture efficiency. Traditional prediction models often have limitations in dealing with strongly coupled, nonlinear, and time-varying water temperature data. A novel hybrid model for temperature prediction is proposed to improve prediction generalization ability and robustness. The model integrates advanced data processing and prediction techniques. Firstly, the VMD method is utilized to achieve effective data decomposition and noise reduction. Secondly, the CNN algorithm is applied to achieve feature extraction of the data. Finally, the bi-directional LSTM and self-concerned combination are used to obtain the final prediction results. The experimental results show that the MAE, RMSE, MSE, MAPE, and R 2 of the VMD-CNN-BILSTM-SA combination prediction model proposed in this paper are 0.016, 0.143, 0.020, 0.035, and 0.978, respectively. Compared with other deep learning models, the BiLSTM model presented in this paper improves the R 2 by 13.2% compared with LSTM and 13.7% over the GRU model. This study can be applied in fishery farming, which can reduce the risk of farming and promote the modernization of fishery.
This paper presents a novel hybrid capacitor-clamp cascade 13-level inverter. The inverter is based on the series connection between two inverter modules with different dc bus voltage. It is the fact of a hybrid control scheme that higher voltage devices and faster devices operate in synergism. Subharmonic PWM method is employed in lower voltage module. In the case of three carrier dispositions, mathematic models of the modulation process are constructed. The rigorous analytic solution is derived, which facilitates the evaluation of spectra accurately. Spectral characteristics of output waveforms including all operation conditions were studied. The proposed inverter is also verified by simulation to confirm the feasibility of the proposed approach.