China
Africa
Explore chapters and articles related to this topic
Machine Vision System Components
Published in Sheila Anand, L. Priya, A Guide for Machine Vision in Quality Control, 2019
It is important to match machine vision cameras with appropriate lenses to obtain good-quality images. We know that machine vision applications require quality images with appropriate image resolution, contrast, and sharpness to enable identification of the required features in the image. The choice of lens has a direct impact on the speed and accuracy of image capture. The main challenge is to be able to obtain an image that has clarity and sharp focus throughout the entire image.
Image Processing for Knowledge Management and Effective Information Extraction for Improved Cervical Cancer Diagnosis
Published in Kavita Taneja, Harmunish Taneja, Kuldeep Kumar, Arvind Selwal, Eng Lieh Ouh, Data Science and Innovations for Intelligent Systems, 2021
Image enhancement is a process of improving contrast adjustment and sharpness in the image. This is the initial step for the entire image processing application. The main working process is adjusting the pixel values until better visualization. The basic image enhancement methods are shown (Figure 5.7).
Linear Filter Applications
Published in David C. Swanson, ®, 2011
While directional derivatives can be very useful in detecting the orientation of image features such as edges, sometimes it is desirable to detect all edges simultaneously. The geometry of the detected edges can then be used to identify important information in the image such as shape, relative size, and orientation. A straightforward edge detection method, known as Sobel edge detection, computes the spatial derivatives in the x- and y-directions, sum their squares, and compute the square-root of the sum as the output of the filter. A less complex operator, known as the Kirsh operator, accomplishes a more economical result without the need for squares and square roots by estimating all eight gradients and taking the maximum absolute value as the edge detection output. The application of Sobel edge detection, with the Kirsh approximation to our test image in Figure 5.9 can be seen in Figure 5.14, where we have inverted the image to see the edges as black in a white field. This line-art appearance to the image is much easier to let the computer calculate than manually drawing it or even using photographic techniques! While edge detection is useful for extracting various features from the image for use in pattern recognition algorithms, it can also be used to enhance the visual quality of the image. The edge detector operator can easily be seen as a type of high-pass filter allowing only abrupt changes in spatial brightness to pass through to the output. If one could amplify the high frequencies in an image, or attenuate the low frequencies, one could increase the sharpness and apparent visual acuity. Typically, sharpness control filtering is done using a rotationally invariant Laplacian operator as follows: () ∇2B=∂2B∂x2+∂2B∂y2
Adaptive rule-based colour component weight assignment strategy for underwater video enhancement
Published in The Imaging Science Journal, 2023
Jitendra P. Sonawane, Mukesh D. Patil, Gajanan K. Birajdar
Prasath et al. gave attention to the attenuation and dispersion caused by underwater image enhancement and proposed colour correction and contrast correction algorithm using local and global contrast correction integrated model with a new optimization model named the Distance Oriented Cuckoo search (DOC) algorithm [16]. Compared to methods such as CLAHS, dynamic stretching, IGLCC, IGLCC-GA, IGLCCPSO, IGLCC-FF, and IGLCC-CS, the form of DOC has accomplished a better result. An approach that addresses the problem of green–blue colour casting by enhancing the gain factor of the red channel and balancing the intensity value for the remaining channels is presented in ref. [17]. To improvise the overall sharpness of the image, the unsharp masking technique was applied, solving the problem of poor contrast.
A Novel Infrared Image Enhancement Based on Correlation Measurement of Visible Image for Urban Traffic Surveillance Systems
Published in Journal of Intelligent Transportation Systems, 2020
Jingyue Chen, Xiaomin Yang, Lu Lu, Qilei Li, Zuoyong Li, Wei Wu
Figure 10 shows five IR images with their registered VIS images, where the VIS images 1, 4 are clearer than others. The VIS image 3, 5 are poor in detail and clarity, and the VIS image 2 is worst. Figures 11–15 shows the experimental results of five IR images and their corresponding detail maps. As can be seen from the result images of HE in the Figures 11–15, the HE can significantly improve the brightness, contrast, and enhance the edge sharpness of IR images. However, the detail in highlighted regions is missing and noise is also enhanced. For instance, the noise of the background shown in Figures 12, 13, and 15 is severe. The detail of highlighted regions in Figures 11 and 14 is absence seriously. The brightness of the result images by SSR, MSR, MSRCR is obviously enhanced, whereas the contrast is poor and the edge information is seriously deficiency. For example, shown in Figures 12 and 15 the edge between the people and background is obscure, especially MSR. In Figure 11, the sharpness and brightness of the IR images via AHPBC is enhanced. However, the detail is damaged, where highlighted areas in the middle of image. In Figures 11 and 13, we can see that there are many dark noises on the building and the sea, and the effect is undesirable. In addition, the overall contrast is not ideal. For the proposed method, the center playground regions in the Figure 11, the background in Figure 12, the sea in Figure 13, the pillar close to the people in Figure 14, the door and window at the middle regions in Figure 15, all surpasses other methods. It can be shown that the proposed method cannot only enhance the brightness and contrast, the image information, sharpness, and visual effect are also ameliorated.
Investigating the contribution of different sizes of pore spaces to the permeability of heterogeneous carbonate rocks using Markov Chain Monte Carlo and lattice-Boltzmann simulation
Published in Geosystem Engineering, 2020
Javad Ghiasi-Freez, Mansour Ziaii, Ali Moradzadeh
The pore differentiation includes several image processing and analyzing steps. First, the light intensity of the images was equalized and the pixel sharpness was increased using low-pass filter and median filter, respectively. Both filters applied one time with a neighborhood of 3 × 3 pixels to remove the possible small noises or incongruity alongside saving the natural representation of images.