Aiming at solving the issues that fire detection is prone to be affected by environmental factors, and the accuracy of flame and smoke detection remains relatively low at the incipient stage of fire, a fire detection algorithm based on GCM-YOLO is put forward. Firstly, GhostNet is introduced to optimize the backbone network, enabling the model to be lightweight without sacrificing model accuracy. Secondly, the upsampling module is reorganized with content-aware features to enhance the detail capture and information fusion effect of the model. Finally, by incorporating the mixed local channel attention mechanism in the neck, the model can enhance the processing capability of complex scenes. The experimental results reveal that, compared with the baseline model YOLOv8n, the GCM-YOLO model in fire detection increases the mAP@0.5 by 1.2%, and the number of parameters and model size decrease by 38.3% and 34.9%, respectively. The GCM-YOLO model can raise the accuracy of fire detection while reducing the computational burden and is suitable for deployment in practical application scenarios such as mobile terminals.

Visual sensor-based infusion monitoring methods have the advantage of being easily integrated into existing medical facilities, but suffer from the problem that monitoring accuracy decreases when the distance between dropper and camera changes. Therefore, this paper proposed a deep learning-based infusion monitoring method that uses improved YOLOv8n network to detect the dropper, droplet, and liquid level, obtains the relative droplet area and relative liquid level height from the ratio of geometric parameters of the detection anchor box, and calculates the infusion rate by the periodic change of the relative droplet area. The experimental results show that the mAP@0.5:0.95 of the improved network reached 91.739%, using relative droplet area to calculate the infusion rate is more accurate than previous method using droplet area, and the error did not exceed 1 drop per minute when the infusion rate is less than 120 drops per minute. It can also accurately measure the infusion remainder in the dropper when the infusion is about to end, and remind the medical staff to change the bottle or remove the needle at the right time.

Fault feature extraction of analog circuits is the main factor affecting their fault diagnosis accuracy. Therefore, in order to increase the accuracy of analog circuit fault diagnosis, this paper proposes a fault diagnosis method for analog circuit based on VMD-CFOA-ELM and DGCCA(Deep Generalized Canonical Correlation Analysis,DGCCA). The method firstly performs VMD(Variational Mode Decomposition,VMD) on the fault time-domain response signal to obtain all the IMF (Intrinsic Mode Function,IMF) components and calculates the CMSE(Composite Multi-Scale Entropy,CMSE) of each IMF component, treating this as a separate subset of fault features. Next, it extracts the time-frequency domain characteristic features of the fault response signal and uses the DGCCA algorithm for supervised feature fusion extraction, constructing a shared representation feature matrix G, which is spliced with the subset of fault features as the input features of the classifier. Finally, the CFOA(Catch Fish Optimization Algorithm, CFOA)algorithm is relied upon to optimize the ELM (Extreme Learning Machine,ELM) classifier to have good classification accuracy to diagnose the type of faults in the test circuits. Fault simulation diagnostics on a Sallen-Key bandpass filter circuit show that, compared to methods such as KPCA-ELM, KSLPP-ELM, EEMD multi-scale entropy-LSSVM, and WP-ICA energy spectrum-SVM, the proposed method achieves the highest accuracy, reaching 99.20%. Additionally, this method is portable and applicable to fault diagnosis of other analog circuits.

In response to the issues of low precision, a large number of parameters and high model complexity in steel surface defect detection, a lightweight algorithm using improved YOLOv8 is proposed. Firstly, GhostNet is utilized as the backbone network in order to reduce the number of model parameters and computational complexity. Secondly, the MPCA((MultiPath Coordinate Attention, MPCA) attention mechanism is integrated to enhance feature extraction capabilities. Finally, the SIoU (Simplified IoU, SIoU) is used to replace the traditional CIoU loss function, which can make the anchor frame more fast and accurate in the regression process, to improve the stability and the robustness of detection. The experimental results indicate that these enhancements have led to a reduction of 37% in calculation amount for the improved YOLOv8n algorithm, a decrease of 32% in parameter count, and an increase in average detection accuracy (mAP) by 1.2%. This model achieves a balance between lightweighting and detection accuracy while providing a viable solution for deployment in computationally resource-constrained edge computing environments such as embedded systems and mobile devices.