MSANet: efficient detection of tire defects in radiographic images

Abstract

Visual inspection has been widely studied and applied in industrial fields. Previous work has investigated the use of established traditional machine learning and deep learning methods to perform automated defect detection for tires. However, intelligent tire defect online detection is still a challenging task due to the complex anisotropic texture background of tire radiographic image. In this paper, we propose an efficient tire defect online detection method named MSANet based on an improved lightweight Yolov4 tiny network. A novel multi-scale self-attention feature enhancement module (MSAM) is designed to extract feature map with rich multi-scale context information. An improved feature pyramid model, named MC-FPN is proposed which utilizes MSAM and convolutional block attention module (CBAM) to enhance the information representation of the feature pyramid. Ablation experiments are conducted to verify the effectiveness of the proposed modules. Comparative experimental results with the state-of-the-art methods validate the effectiveness and efficiency of the proposed method which can achieve a mAP of 96.96% and an average detection time of 30.81ms per image. The proposed method can meet the requirements of industrial online detection by virtue of its lower computational costs, and has good generalization ability in other visual inspection tasks.

Weight-guided feature fusion and non-local balance model for aluminum surface defect detection

“…A weighted bi-directional FPN is proposed in EfficientDet [17] to perform feature fusion. Zhao et al [5] proposed a MSAM-CBAM feature pyramid network, which improves YOLOv4-tiny by multi-scale feature fusion, thus alleviating the small object detection problem. Irrelevant features were weakened by the embedded CBAM.…”

“…However, the conventional inspection methods cannot meet the detection requirements when aluminum surface defects appear with drastic scale changes and similar features of defect areas and background texture features [3,4]. With the development of artificial intelligence technology, more computer vision methods are applied to industrial defect detection [5]. Defect detection based on computer vision and deep learning can solve the problems existing in traditional defect detection.…”

Explainable attention-based fused convolutional neural network (XAFCNN) for tire defect detection: an industrial case study

“…The complex composite structures of tires pose a challenge in recognizing defects, as the presence of varied textures within and between tires' layers, as well as the low contrast of certain faults like bubbles in the tread and foreign objects in the bead, can complicate the process [4]. In order to comprehend the intricate aspects of tires, specialized imaging devices known as x-ray cameras are employed to capture comprehensive 360-degree radiographic images of the tire structure.…”

“…The tire is equipped with steel wires that are uniformly distributed over its left, center, and right sections. A careful examination of these elongated x-ray images can reveal discrepancies that may be challenging to identify by alternative means [4].…”

Automated defect recognition (ADR) for monitoring industrial components using neural networks with phased array ultrasonic images

“…The traditional method of characterizing TFM images by NDE experts is time-consuming and prone to human error as the data becomes more complex. Therefore, there is a growing interest in developing ADR systems to reduce human intervention and improve inspection speed [20,21].…”