Assessment and evaluation are the essential processes of industrially manufactured products for the determination of the quality and quantity of products. They give justifications in a practical way about whether the machine is perfect or imperfect, which can lead to a better or poorer production. In this study, the authors propose an algorithm that uses morphological geodesic active contour and image processing techniques to perform segmentation and assess the performance of a robot used to manufacture welding beads. The algorithm has four parameters which are pre-processed images, balloon force, smoothing parameter, and number of iterations. To pre-process the images, the algorithm uses an inverse Gaussian gradient operator for edge detection and applies the histogram equalization method to level the distribution. To detect the external contour of the bead, the level set is initialized as the region of interest whereby a balloon force can inflate or deflate towards the edges. To smoothen the contour, a smoothing parameter is applied to convert the jagged lines into a curve over a reasonable number of iterations. Based on the experimental results, the authors' algorithm used a fixed balloon force of −2, a smoothing parameter value of 4, and 40 iterations to segment images obtained from three different environments. The computation time for the segmentation and evaluation of one image was 0.70, 0.61, and 0.67 s for datasets with high brightness, low brightness, and normal brightness, respectively. Additionally, the authors' proposed algorithm achieved an outstanding performance of 0.9954, 0.9843, 0.9892, and 0.9435 in terms of recall, precision, F-measure, and IOU, respectively. To justify the performance of the authors' proposed algorithm, the authors compared it with the existing algorithms and found that it worked better than all the others for segmentation, although it lagged behind the entropy-based algorithm in terms of speed.
Researches on pattern recognition have been tremendously performed in various fields because of its wide use in both machines and human beings. Previously, traditional methods used to study pattern recognition problems were not strong enough to recognize patterns accurately as compared to optimization algorithms. In this study, we employ both traditional based methods to detect the edges of each pattern in an image and apply convolutional neural networks to classify the right and wrong pattern of the cropped part of an image from the raw image. The results indicate that edge detection methods were not able to detect clearly the patterns due to low quality of the raw image while CNN was able to classify the patterns at an accuracy of 84% within 1.5 s for 10 epochs.
The procedures of white points detection and localization are practically complex on noisy images. In this paper, we propose an algorithm that detects and localizes white points on 3D film images. The proposed algorithm uses the fast Fourier transform to convert the binarized image into real and imaginary parts to obtain the number of white points along the horizontal and vertical. We determine the sorted coordinates of the white points by adding a brute-force solution to the coordinates obtained from the real part of the image. These sorted coordinates are obtained by subtracting the error between the Euclidean distances of the normalized coordinates along the vertical and horizontal direction. The proposed algorithm with and without brute-force achieved an average detection ratio of 0.98 and 0.88 respectively, while the others underperformed. We perform various experiments using the existing algorithms such as template matching, thresholding, and an iterative method to validate the performance of our algorithm. We also compare the rule-based algorithms that detect and localize objects in noisy images with the proposed one to determine the reliability of our algorithm. The experimental results indicate that the proposed algorithm performs better than the template matching, thresholding, and iterative algorithm.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.