China
Africa
Explore chapters and articles related to this topic
Image and Its Properties
Published in Ravishankar Chityala, Sridevi Pudipeddi, Image Processing and Acquisition using Python, 2020
Ravishankar Chityala, Sridevi Pudipeddi
Here is an example code snippet where we read an image and write an image. The imwrite function takes the file name and the ndarray of an image as input. The file format is identified using the file extension in the file name. import cv2 img = cv2.imread('image1.png') # cv2.imwrite will take an ndarray. cv2.write('file_name', img)
Data Models for Storage and Retrieval
Published in Praveen Kumar, Jay Alameda, Peter Bajcsy, Mike Folk, Momcilo Markus, Hydroinformatics: Data Integrative Approaches in Computation, Analysis, and Modeling, 2005
A data model provides a conceptual view of objects, but does not address the need to organize the objects in ways that facilitate the operations we need to perform. The choice of a file format can be of critical importance for storing, accessing, and managing these objects by computers. We examine data formats in detail in Chapter 8.
A time series decomposition approach to detect coal fires in parts of the Gondwana coalfields of India from VIIRS data
Published in Journal of Spatial Science, 2023
Ritesh Mujawdiya, R. S. Chatterjee, Dheeraj Kumar
The VNP21A2 data files are provided in Hierarchical Data Format version 5 (HDF5). A single file contains 11 bands. These bands include emissivity maps in wavelength ranges 8.4–8.7 µm, 10.26–11.26 µm, and 11.54–12.49 µm, daytime and nighttime LST maps, as well as quality control, view angle, and view time bands for LST maps. The required bands were extracted from HDF files and converted into Tag Image File Format (TIFF). The LST maps have sinusoidal projection; therefore, their projection was changed to the WGS84 UTM zone 45 projected coordinate system. The data is available in rescaled digital numbers that must be multiplied by 0.02 to obtain actual LST values. The LAADS web interface provides VIIRS images in tiles. Due to the large size of the study area, a total of three tile locations were covered by the study area. Each tile location had 322 daytime and 322 nighttime LST maps. Therefore, the corresponding tiles were mosaicked to form full images covering the entire study area.
Simplified Prediction Method for Detecting the Emergency Braking Intention Using EEG and a CNN Trained with a 2D Matrices Tensor Arrangement
Published in International Journal of Human–Computer Interaction, 2023
Hermes J. Mora, Esteban J. Pino
The TensorFlow (TF) algorithm is a Python-open source library for numerical calculus making machine learning faster and easier (Python, 2019). The usage of TF to design and train CNNs is reasonably easy due to the strong support in Artificial Intelligence that has a vast number of functions to manage the input data. We can build a data generator with the (.numpy) extension that corresponds to a Python library implemented for working with N-dimensional arrays. With this in mind, the input data can be used directly as a large array configured by matrices. It is no necessary a dataset as RGB or grayscale images. Implementing a tensor (n-dimensional matrix) implies that the network designer reduces the processing time and computer resources when training the network. Based on Table 2, the dataset for each of the six electrode groups is configured directly as a (.numpy) array, Figure 3(b). By contrast, the image groups used to train our CNN through grayscale images and compare the CNN results are converted into the common image file format (.png). There are six electrode groups (4, 8, 13, 18, 33, 59) that result in six different 2 D-tensors. The height of each 2 D matrix varies according to the number of electrodes. The width of the matrix (400 samples) because there is no variation in the length of segments. The number of images in each image-set is significantly reduced given the number of electrodes used in each group.
Colour filter array demosaicking over compression through modified grey wolf optimization technique
Published in The Imaging Science Journal, 2018
M. S. Safna Asiq, W. R. Sam Emmanuel
The proposed demosaicking algorithm combined with optimized compression techniques is compared using the Kodak dataset of Kodak Eastman Company. The dataset consists of 24 true colour images of 768 × 512 dimensions in tagged image file format. The work was implemented in the MATLAB R2015a platform. The performance is measured using the Peak Signal Noise Ratio (PSNR), Structural Similarity Index Measure (SSIM) and Feature Similarity Index Measure (FSIM). The PSNR, SSIM and FSIM values express the quality measure. Tables 1–3 show the performance of the proposed work. The conventional demosaicking algorithms used for analysis includes Bilinear Interpolation (BI), Gradient-Based Threshold-Free (GBTF) Algorithm [31], (Figures 7 and 8).