| At present,China has become one of the countries with serious waste problems.Compared with the widely used landfill sites,the incineration technology adopted in China can effectively reduce the volume of waste by more than 90%.Incineration treatment has many advantages,such as high reduction rate,no need for further decomposition.In addition,the residues produced by incineration can be used as building materials.During the incineration process,pathogens will be completely eliminated,and harmful gases produced by perishable organic matter will be oxidized.Under ideal conditions,incineration operation is reliable and clean,and the fully enclosed production mode is generally not affected by natural conditions.Compared with landfill sites,incineration plants have a smaller footprint and may be easier to locate.During the operation of the waste incineration plant,in order to realize the automatic garbage detection and grasping,and solve the problem of whether the garbage is burned enough,improve the veracity of the garbage station operators and reduce the labor demand,firstly,the single-stage target detection algorithm and two-stage target detection algorithm are introduced.At the same time,the image segmentation algorithms represented by full convolution network(FCN)and U-Net are introduced.Next,this research adopted the real-time garbage detection algorithm based on YOLOV7 algorithm,and used the cameras in the garbage pool area to take real-time photos to detect the garbage in each partition,then the garbage is grabbed by crane gripper in real time and transferred to the area to be incinerated,so as to improve the production efficiency.In order to improve the performance of the algorithm,a new algorithm named SMSTHN-YOLO is proposed in this paper.The algorithm uses Swin-Transformer V2 as the backbone network,embeds the Hor Net-based recursive gated convolution,and integrates the attention mechanism of Sim AM.Compared with the original YOLOV7 algorithm,the improvement is 10.1%.The improved algorithm can be applied to the real-time detection of Garbage and crane gripper in real scene,and meet the industrial needs of garbage disposal.At the same time,the U-Net algorithm is used in this study.The algorithm can be applied to the real-time detection of waste incineration in real scene,and meet the industrial requirements of waste disposal and waste incineration.This study provides an effective real-time detection and capture method for waste disposal industry.These methods help to increase the production efficiency of the waste disposal industry,reduce labor costs,and reduce the burden of operating workers.The application of real-time detection and capture technology can also help to solve the problem of adequacy of waste incineration and ensure that the process of waste incineration is more efficient and environmentally friendly.To sum up,this study explored convolutional neural network object detection algorithms and their applications in the waste processing industry,providing strong support for automated waste detection and retrieval at waste stations,it provides theoretical basis and practical reference for the sustainable development of our country’s waste treatment industry.In this study,we explore the application of convolutional neural network object detection algorithm in the waste treatment industry,and provide useful enlightenment for the automation and intelligentization of the waste treatment industry.In the future research,we should continue to explore the application of other advanced technologies in waste treatment,to provide more technical support for the sustainable development and upgrading of waste treatment industry.With the continuous innovation and development of technology,it is believed that the waste treatment industry in the future will make greater breakthroughs in the protection of the environment and improve the utilization of resources.Figure [28] Table [2] Reference [66]... |
