A method for oil and gas pipeline marker recognition based on neural network
A neural network, oil and gas pipeline technology, applied in the field of oil and gas pipeline safety, can solve problems such as large interference, and achieve the effects of eliminating interference information, small detection model, and fast detection speed
Active Publication Date: 2022-07-29
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF5 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] In pipeline safety management, it is more hoped that the aerial pictures can be detected in real time to determine whether there is any abnormality; for the aerial pictures, the landmarks are relatively small compared to the whole background, and due to the complex and varied terrain of the scene, the interference is large, and with the aerial photography There are also frequent scale changes in highly variable markers; therefore, the main challenge at present is how to use smaller and faster models to identify multi-scale targets and smaller targets
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0029] The specific embodiments and working principles of the present invention will be further described in detail below with reference to the accompanying drawings.
[0030] like figure 1Shown is the structure model diagram of the target detection network constructed based on the neural network of the present invention, that is, the structure diagram of the Mobilenet-SSD network embedded in the CBAM. The input is a picture with a size of 300×300×3, the first layer is a standard convolutional layer, represented by Conv, and the convolution with stride 2 is performed by 32 3×3×3 convolution kernels, and the output is 150× The feature map of 150×32; the next 13 layers represent 13 depthwise separable convolution modules, represented by Depthwise; the depthwise separable convolution module of the second layer is firstly performed by a 3×3×32 convolution kernel with a stride of The Depthwise convolution of 1 outputs a feature map of 150 × 150 × 32, and then 64 1 × 1 × 32 convolu...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention provides a method for identifying oil and gas pipeline markers based on a neural network. The method includes the steps of: building a model, obtaining a training set, a verification set and a test set, obtaining an optimal model after training, using the test set for testing, and obtaining detection results and calculate the detection accuracy. The invention uses Mobilenet with few parameters as the basic network, and the SSD algorithm that directly returns on the multi-scale feature map as the target detection algorithm, realizes multi-scale target detection, and has a small detection model and a fast detection speed, so that the detection model can be deployed in mobile At the same time, the attention mechanism is embedded behind each layer of the basic network, and the salient features of the markers are highlighted from the global and local scope through the CBAM attention mechanism. It can enhance the feature expression ability layer by layer from the first layer, effectively remove the interference information, and improve the detection accuracy.
Description
technical field [0001] The invention belongs to the field of oil and gas pipeline safety, in particular to a method for identifying oil and gas pipeline markers based on a neural network. Background technique [0002] Oil, natural gas and other resources play an important strategic role, and their transportation has the characteristics of long distance and wide range. Due to the wear, corrosion and other factors of the oil and gas pipeline system itself, leakage occurs from time to time. Oil and gas leakage can easily lead to explosion, fire, and toxic and harmful Therefore, it is of great significance to strengthen the safety management of oil and gas pipelines. [0003] In the field of oil and gas pipeline safety, in addition to time-consuming and labor-intensive manual inspections, there are currently methods for pipeline leak detection based on neural network training pipeline weld defect data or pressure wave signals during pipeline leakage. The characteristics of the ...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06V10/774G06V20/17G06K9/62G06N3/04
CPCG06V20/13G06V2201/09G06V2201/07G06N3/045G06F18/214
Inventor 于永斌唐倩彭辰辉陆瑞军买峰汤亦凡戚敏惠邓权芯毛宇涵
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA