Workpiece recognition and inspection by a model-based scene analysis system

Rummel, P.; Beutel, W.

doi:10.1016/0031-3203(84)90041-4

Cited by 40 publications

(6 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…According to the algorithms used, it can be generally divided into two categories: templatebased matching algorithm and learning-based classifier algorithm. The template-based matching algorithms usually use gray (Rummel and Beutel, 1984), edges, textures, local features (Lowe, 1999) and other features to match the correct class. Template-based algorithms mostly use global search, they are time-consuming (Yan et al, 2009) and are sensitive to shadows and noises (Yang et al, 2012).…”

Section: Fast Workpiece Detection Methodsmentioning

confidence: 99%

A fast workpiece detection method based on multi-feature fused SSD

Shi

Zhang

Zeng

2021

View full text Add to dashboard Cite

Purpose Workpiece sorting is a key link in industrial production lines. The vision-based workpiece sorting system is non-contact and widely applicable. The detection and recognition of workpieces are the key technologies of the workpiece sorting system. To introduce deep learning algorithms into workpiece detection and improve detection accuracy, this paper aims to propose a workpiece detection algorithm based on the single-shot multi-box detector (SSD). Design/methodology/approach Propose a multi-feature fused SSD network for fast workpiece detection. First, the multi-view CAD rendering images of the workpiece are used as deep learning data sets. Second, the visual geometry group network was trained for workpiece recognition to identify the category of the workpiece. Third, this study designs a multi-level feature fusion method to improve the detection accuracy of SSD (especially for small objects); specifically, a feature fusion module is added, which uses “element-wise sum” and “concatenation operation” to combine the information of shallow features and deep features. Findings Experimental results show that the actual workpiece detection accuracy of the method can reach 96% and the speed can reach 41 frames per second. Compared with the original SSD, the method improves the accuracy by 7% and improves the detection performance of small objects. Originality/value This paper innovatively introduces the SSD detection algorithm into workpiece detection in industrial scenarios and improves it. A feature fusion module has been added to combine the information of shallow features and deep features. The multi-feature fused SSD network proves the feasibility and practicality of introducing deep learning algorithms into workpiece sorting.

show abstract

Section: Fast Workpiece Detection Methodsmentioning

confidence: 99%

A fast workpiece detection method based on multi-feature fused SSD

Shi

Zhang

Zeng

2021

View full text Add to dashboard Cite

show abstract

“…This method is based on structural relations that describe an identifiable image. Image (scene) is represented as some structure that consists of graphic primitives [5]. First the contours chooses to find objects on the image.…”

Section: Image Transformationmentioning

confidence: 99%

“…Check for validity of categorized object. The specially trained Hamming neural network is capable to recognize the image that has been compounded of primitives [4], therefore, it will be used for confirmation of classification on grammar [5]. In case of error there is a rollback of operatings up to step Synthesis of interdependent objects.…”

Section: Object Recognitionmentioning

confidence: 99%

The Component Method of Scene Analysis and Object Recognition

Vechur¹,

Chayka²,

Chemodakov³

2014

IJC

View full text Add to dashboard Cite

show abstract

“…In these techniques (Rummel and Beutel 1984, Ayache and Faugeras 1986, Knoll and Jain 1986, Chaudhury et al 1990), some estimated positions and orientations are found from the matched set of features. Then a tree search technique based on heuristic reasoning is employed to ® nd the match for other features.…”

Section: Heuristic Search Techniquesmentioning

confidence: 99%

Connectionist object recognition: Methods and methodologies

Basak¹,

Pal²

1999

International Journal of Systems Science

View full text Add to dashboard Cite

An overview of di erent connectionist approaches for object recognition is presented. These approaches are ® rst of all classi® ed on the basis of functional characteristics and structural characteristics of the networks. At the next level, these are partitioned on the basis of their underlying assumptions and the types of recognition problem that they handle. Their principles and key features are highlighted. The issue of mixed category perception (in the context of multiple object recognition) is addressed. A comparison of the characteristics of various application-speci® c systems is provided in a tabular form. Some of the basic issues to be considered while building a connectionist system for object recognition are ® nally listed.

show abstract

Workpiece recognition and inspection by a model-based scene analysis system

Cited by 40 publications

References 1 publication

A fast workpiece detection method based on multi-feature fused SSD

A fast workpiece detection method based on multi-feature fused SSD

The Component Method of Scene Analysis and Object Recognition

Connectionist object recognition: Methods and methodologies

Contact Info

Product

Resources

About