Railway fastener defect form rapid detection vehicle

Abstract

The invention discloses a railway fastener defect form rapid detection vehicle which is divided into two parts of hardware and a built-in model algorithm, the hardware comprises a vehicle body, a solar cell panel installed at the top of the vehicle body and a computer transmission control platform inside the vehicle body, and two fixing rods are arranged on the left side and the right side of thefront portion of the vehicle body respectively; a light supplement device, a three-axis stabilizer and a camera are arranged on the rod, and a storage battery pack is arranged behind the vehicle body;the built-in model algorithm is adopted to train a mature DenseNet201 deep migration learning model on a self-made railway fastener defect form special data set; the camera is used for shooting fasteners along a railway, when collected fastener pictures are recognized as defective fasteners through the built-in algorithm, an alarm is triggered, geographic position information of the fasteners isautomatically recorded and uploaded, detection is completed, and the railway fastener defect form rapid detection vehicle is reasonable in design, convenient to use, practical and capable of greatly improving detection efficiency.

Description

technical field [0001] The invention relates to the field of defect shape detection of railway fasteners, in particular to a rapid inspection vehicle for defect shape of railway fasteners combined with modern technologies such as computer vision and deep learning. Background technique [0002] Railway fasteners are extremely important connecting parts in the railway system, which provide a huge guarantee for the safe running of trains. Therefore, the detection of railway fasteners is of great significance. [0003] At present, most of the existing detection methods use manual inspection and detection vehicles. Manual detection has the disadvantages of strong subjectivity, time-consuming and labor-intensive, and cannot guarantee the safety of the detection workers. However, most of the existing detection methods through detection vehicles are It is a detection method based on traditional computer vision. The entire detection process has to go through steps such as data collec...

AI Technical Summary

Problems solved by technology

[0003] At present, most of the existing detection methods use manual inspection and detection vehicles. Manual detection has the disadvantages of strong subjectivity, time-consuming and labor-intensive, and cannot guarantee the safety of the detection workers. However, most of the existing detection methods through detection vehicles are It is a detection method based on traditional computer vision. The entire detection process has to go through steps such as data collection, data image preprocessing, fastener positioning, artificially designed features, feature extraction, and classifier classification. This method mainly has high detection costs and high detection process. Complicated and poor generalization performance

Images