Unlock instant, AI-driven research and patent intelligence for your innovation.

Target detection method suitable for small-size parts and stacked parts and application thereof

A target detection and small-volume technology, applied in the field of visual inspection, can solve the problems of high requirements for equipment hardware and data sets, low classification accuracy, slow recognition speed, etc., achieve good recognition effect, strengthen expression ability, and solve the imbalance of quantity Effect

Pending Publication Date: 2021-11-16
SHANGHAI NORMAL UNIVERSITY
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] The purpose of the present invention is to overcome the existing defects of low target detection and classification accuracy for small-volume parts and stacked parts, high requirements for equipment hardware and data sets, and slow recognition speed, and to provide a high-precision classification system for equipment. A target detection method suitable for small-volume parts and stacked parts with low requirements for hardware and data sets and fast recognition speed

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Target detection method suitable for small-size parts and stacked parts and application thereof
  • Target detection method suitable for small-size parts and stacked parts and application thereof
  • Target detection method suitable for small-size parts and stacked parts and application thereof

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0060] A target detection method suitable for small-volume parts and stacked parts, applied to electronic equipment, the steps are as follows figure 1 As shown, specifically:

[0061] (1) Build a known part data set:

[0062] (1.1) Collect images of various parts from different angles and in different industrial production environments, which is the initial known part data set;

[0063] Aiming at environmental factors such as viewing angle changes, illumination changes, and background interference in actual scenes in the industrial manufacturing field, collect enough images of parts in the above situations. Due to differences in actual application scenarios, the lighting system needs to select the appropriate light source type and lighting method according to the actual situation. Because the surface of the collected target parts is smooth, in the case of direct lighting and coaxial lighting, a lot of reflections will be generated, which will weaken the surface features of t...

Embodiment 2

[0093] An electronic device, including one or more processors, one or more memories, one or more programs, and an image acquisition device for acquiring target pictures;

[0094] One or more programs are stored in the memory, and when the one or more programs are executed by the processor, the electronic device executes the object detection method applicable to small-volume parts and stacked parts as described in Embodiment 1.

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

No PUM Login to View More

Abstract

The invention discloses a target detection method suitable for small-size parts and stacked parts and application thereof, and the method comprises the steps: inputting a target picture into a trained YOLOv3 improved model, and outputting a target category through the trained YOLOv3 improved model; compared with a YOLOv3 model, the improvement of the YOLOv3 improved model lies in that an improved Focal Loss formula is adopted to replace a negative sample confidence part in a loss function of the YOLOv3 model, and an SPP module is added between the last feature scale of a YOLOv3 network structure and a detection layer. According to the target detection method, target detection is carried out through the YOLOv3 improved model, the YOLOv3 improved model can be suitable for detection of small targets and stacked targets by combining local and global multi-scale features, meanwhile, the method has the advantages of being high in classification precision, low in requirements for equipment hardware and the number of data sets and high in recognition speed, and the method has great application prospects.

Description

technical field [0001] The invention belongs to the technical field of visual detection, and relates to a target detection method suitable for small-volume parts and stacked parts and its application. Background technique [0002] With the rapid development of industrialization, the field of intelligent manufacturing has become a new focus in the process of industrial production automation of a large amount of repetitive labor. Parts recognition is an important basis for industrial automation and intelligence. It has a wide range of application scenarios in the field of smart manufacturing, such as object detection systems for smart warehousing, parts recognition and classification on conveyor belts, and parts detection for smart assembly workshops. Part identification and detection is the primary problem facing the development of intelligent manufacturing, and solving this problem is the first requirement to realize automated production. [0003] Due to the progress of soc...

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
no application Login to View More
IPC IPC(8): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/047G06N3/045G06F18/23213G06F18/2415G06F18/253G06F18/214
Inventor 安康张文浩朱凯上官倩芡方厚招
Owner SHANGHAI NORMAL UNIVERSITY