Design and implementation of video processing controller for pipeline robot based on embedded machine vision

With the rapid development of pipeline transportation, there are more and more hidden safety risks behind the rapid development of pipeline transportation. However, the current pipeline inspection methods are incomplete, and the inspection technology fails to meet the requirements. This has led to pipelines with potential safety hazards being missed during the inspection process. In order to perform effective video detection on the pipeline, this paper is based on the pipeline robot on the embedded platform to complete the control operation of the embedded machine vision system on the video processing in the pipeline. The experiments show that the system has better recognition and positioning effect, less error, and the system occupancy and real-time performance can meet the design requirements. This paper proposes an improved LwIP-based congestion prevention method for video processing in embedded machine vision systems. When the RTT value is higher than the RTO value, it will cause false retransmission of data, which will affect the data transmission efficiency. The experimental results show that the improved congestion control method can more effectively deal with the problem of packet loss, effectively reduce the false retransmission of data, and improve the data throughput. Through comparison, it is found that under the same object change interference, the video processing controller based on embedded machine vision can more effectively suppress the video control error in the pipeline than the traditional video processing controller, and it will not affect the stability of the system.