On-line detection device based on multi-modal registration

Abstract

The invention discloses an online detection device based on multi-mode registration, which includes a mode registration unit and an intelligent lighting assembly installed above the mechanical bearing platform from top to bottom; the manual intervention unit is installed on the right side of the mode registration unit. On the other hand, the pattern registration unit is composed of a plurality of pattern registration subunits, and each pattern registration subunit and manual intervention unit are respectively connected to the central processing unit through an interactive data communication bus. The present invention performs registration and analysis on multiple feature information of the detection target at the same time, realizes the classification, collection and synthesis of multiple feature information, and realizes efficient and accurate online detection and recognition; integrates the registration results of multiple feature information, and effectively realizes The compression of multi-feature information reduces redundant feature information, greatly reduces the pressure on the hardware system to process image information, improves the detection accuracy of the system, and provides effective support for the analysis and decision-making of automatic control of the production line by integrating the results of multi-mode registration information.

Description

technical field [0001] The invention belongs to an online detection device, in particular to an online detection device based on multi-mode registration. Background technique [0002] In the process of modern industrial and agricultural production, online detection technology has been widely used in the fields of product production, quality inspection and quality control, and has gradually become one of the basic technologies of modern industry. In these applications, through registration and identification of a certain feature of various product components, the detection of product defects, scratches, stains, shapes, and installation positions can be realized, thereby eliminating or reducing defective products. Relatively speaking, this is a traditional image matching method, which requires strict processing of the image grayscale and other information of the detection target, and a series of optimized processing of image algorithms in order to achieve satisfactory recognit...

AI Technical Summary

Problems solved by technology

[0003] However, it is unavoidable that most of the production lines have various tension fluctuations, noise interference and other phenomena, and the actually obtained detection target images more or less have problems such as stretching deformation and complex background in different situations.

With the complexity of online detection applications and the continuous improvement of detection requirements, it is difficult to describe all the information in the detection scene by relying on a single feature information, while the traditional online detection system uses a single image feature pattern for detection and recognition. Under conditions such as light sources, only one characteristic information of the detection target can be obtained at a time, and the characteristic information used for detection, identification and matching after final processing often does not meet the expected requirements.

[0004] At the same time, the uncertainty of the detection site has brought many difficulties to the operating efficiency, hardware cost, and algorithm complexity of the online detection system. Since the detection target position, light source and other conditions are changing, in many cases, a single feature To identify the detection target, it is difficult to achieve the desired effect

[0005] In many sites where online testing is required, it is often necessary to detect multiple features of the product, such as the surface condition, internal structure damage, installation position, etc. of mechanical products on the production line. At this time, a single pattern registration method is used. Efficiency and accuracy are obviously difficult to meet the requirements. In order to quickly and accurately realize the online detection of the detection target, it is necessary to design a simple, reliable and low-cost technical solution

Images