Aiming at the shortcomings of low real-time, low applicability, and low welding precision of automatic welding system, a seam tracking system based on laser vision is designed. Use the laser vision sensor to collect the weld image and transmits it to the industrial control computer for processing. Using a median filter to eliminate noise impacts such as arc and splash. Then, this paper focuses on the combination of an improved image threshold segmentation algorithm is used to solve the optimal threshold to obtain the binary image. And the information of laser stripe and the background are separated, overcomes the problems that the researchers have encountered before, such as the unrecognized global optimal solution, and the inaccuracy of the segmentation caused by system jitter. Finally, combined with the improved upper and lower average method, least square method, and Hough transform, the weld feature points are identified and more ideal real-time weld tracking is realized. The experimental results show that the method can accurately track the weld feature points, and improve the detection speed.

In the field of autonomous driving, cooperative perception through vehicle-to-vehicle communication significantly enhances environmental understanding by leveraging multi-sensor data, including LiDAR, cameras, and radar. However, traditional early or late fusion methods face challenges such as high bandwidth and computational resources, which make it difficult to balance data transmission efficiency with the accuracy of perception of the surrounding environment, especially for the detection of smaller objects such as pedestrians. To address these challenges, this paper proposes a novel cooperative perception framework based on two-stage intermediate-level sensor feature fusion specifically designed for complex traffic scenarios where pedestrians and vehicles coexist. In such scenarios, the model demonstrates superior performance in detecting small objects like pedestrians compared to mainstream perception methods while also improving the cooperative perception accuracy for medium and large objects such as vehicles. Furthermore, to thoroughly validate the reliability of the proposed model, we conducted both qualitative and quantitative experiments on mainstream simulated and real-world datasets. The experimental results demonstrate that our approach outperforms state-of-the-art perception models in terms of mAP, achieving up to a 4.1% improvement in vehicle detection accuracy and a remarkable 29.2% enhancement in pedestrian detection accuracy.

Aiming at the problems of lack of intelligent fire extinguishing means, insufficient timeliness of manual fire extinguishing, and difficult security guarantee in the existing valve hall, an intelligent fire extinguishing robot system for the converter station valve hall was developed this paper. The overall structure, robot body, water supply system, navigation system and software architecture of the system were introduced. Water spray test was conducted, and the robot could switch between columnar spray and fog spray. Finally, the robot has been applied in the valve hall for one year. The results showed that the robot has a high degree of intelligence, and can extinguish the fire automatically, accurately and quickly. The robot is suitable for firefighting of equipment in valve hall of converter station and provides a new technical means and method.

Although various automatic or semi-automatic recognition algorithms have been proposed for tiny part recognition, most of them are limited to expert knowledge base-based target recognition techniques, which have high false detection rates, low recognition accuracy and low efficiency, which largely limit the quality as well as efficiency of tiny part assembly. Therefore, this paper proposes a precision part image preprocessing method based on histogram equalization algorithm and an improved convolutional neural network (i.e. Region Proposal Network(RPN), Visual Geometry Group(VGG)) model for precision recognition of tiny parts. Firstly, the image is restricted to adaptive histogram equalization for the problem of poor contrast between part features and the image background. Second, a custom central loss function is added to the recommended frame extraction RPN network to reduce problems such as excessive intra-class spacing during classification. Finally, the local response normalization function is added after the nonlinear activation function and pooling layer in the VGG network, and the original activation function is replaced by the Relu function to overcome the problems such as high nonlinearity and serious overfitting of the original model. Experiments show that the improved VGG model achieves 95.8% accuracy in precision part recognition and has a faster recognition speed than most existing convolutional networks trained on the same test set.