Reinforcing steel bar detecting and counting method based on deep neural network model

Abstract

The invention discloses a reinforcing steel bar detecting and counting method based on a deep neural network model, and relates to the technical field of milling machine auxiliary tools, an original picture data set of reinforcing steel bars is expanded and enhanced by adopting a sliding window method and a mosaic enhancement method, and the size of an anchor frame is adaptively determined by adopting a clustering algorithm; firstly, a reinforcing steel bar picture is used as input, image features are extracted through a deep neural network, a prediction frame is output through a detection head, a loss function is obtained through calculation according to the prediction frame and a real frame, model parameters are optimized through a back propagation algorithm, and a used model training algorithm is a stochastic gradient descent method based on momentum. And finally, soft non-maximum suppression is performed on the prediction frame to obtain the final number of reinforcing steel bars.

Description

technical field [0001] The invention relates to the field of computer vision target detection, and more specifically relates to a steel bar detection and counting method based on a deep neural network model. Background technique [0002] In the construction industry, steel bars are one of the indispensable building materials. The number of steel bars must be accurately calculated in every link of steel bar production, transportation, and sales. The staff on the construction site need to check and confirm the quantity of purchased steel bars. [0003] The traditional way of counting steel bars is manual counting, which is cumbersome, time-consuming and labor-intensive, and has high labor costs. The inventory method based on traditional digital image processing is susceptible to interference from many factors such as shooting background, shooting angle, and light intensity in complex environments, and is prone to re-inspection and missed inspections. [0004] With the develo...

AI Technical Summary

Problems solved by technology

[0003] The traditional method of steel bar inventory is manually counted, which is cumbersome, time-consuming and labor-intensive, and has high labor costs

The inventory method based on traditional digital image processing is susceptible to interference from many factors such as shooting background, shooting angle, and light intensity in complex environments, and is prone to re-inspection and missed inspections.

Images