Image matching method for monitoring interior of pipeline based on nine-grid principle

Abstract

The invention provides an image matching method for monitoring interior of a pipeline based on a nine-grid principle. The image matching method comprises the steps that: firstly, a system performs machine learning on a picture set of various conditions in the pipeline, wherein each human face image is sequentially subjected to gray scale conversion, normalization, image grid division, calculationof pixel feature values of each grid and embedment of feature values of each grid into one-dimensional nine-grid vector, and lightweight storage in a robot single-chip microcomputer is realized; and after a robot carries a camera and enters the pipeline, frames of collected videos are taken in real time, the frames of pictures are rapidly converted into one-dimensional nine-grid vectors, the similarity is calculated with the nine-grid vector set already stored in the single-chip microcomputer, and when the similarity is larger than a preset matching degree, the matching succeeds. According tothe image matching method, the process of machine learning is simple, complex mathematical operation is not needed, lightweight algorithm and data set storage can be achieved, the equipment cost is reduced, meanwhile, the monitoring efficiency inside the pipeline is improved, the engineering progress is accelerated, and a powerful guarantee is provided for engineering safety.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a video image matching method for internal monitoring of pipelines based on the Jiugong principle. Background technique [0002] Modern industry and life are inseparable from various pipelines, such as oil pipelines, natural gas pipelines, power transmission pipelines and water pipelines. Today's actual demand for pipeline monitoring automation creates conditions for the development and application of pipeline monitoring technology. However, at present, many related technologies are still immature, the operation is dangerous, and the cost of equipment is high. For example, on-line corrosion monitoring, monitoring based on radioactive sources, and monitoring based on magnetic flaw detection, although the monitoring effect can meet the needs, but the equipment and technology are complicated, often monopolized by large companies, and the supporting equipment is expensive and the cos...

AI Technical Summary

Problems solved by technology

The former mainly uses one-dimensional or two-dimensional sliding templates in space for image matching, and the difference between different algorithms is mainly in the selection of templates and related criteria. This type of method generally has a high matching rate, but has a large amount of calculation and a slow speed; the latter By extracting salient features such as points, lines, and regions in the original image as matching primitives, and then using them for feature matching, the matching speed is generally faster, but the matching accuracy is not necessarily high

Images