Infrared image detail enhancement based on the gradient field specification

Abstract

Human vision is sensitive to the changes of local image details, which are actually image gradients. To enhance faint infrared image details, this article proposes a gradient field specification algorithm. First we define the image gradient field and gradient histogram. Then, by analyzing the characteristics of the gradient histogram, we construct a Gaussian function to obtain the gradient histogram specification and therefore obtain the transform gradient field. In addition, subhistogram equalization is proposed based on the histogram equalization to improve the contrast of infrared images. The experimental results show that the algorithm can effectively improve image contrast and enhance weak infrared image details and edges. As a result, it can give qualified image information for different applications of an infrared image. In addition, it can also be applied to enhance other types of images such as visible, medical, and lunar surface.

An Enhancement and Detection Method for a Glue Dispensing Image Based on the CycleGAN Model

“…As seen in (7), the original CycleGAN model uses the MEA loss function for the cyclic consistent loss. Although the MEA [28] loss function is more robust to outliers, the gradient of the MEA loss function is always the same when the neural network parameters are updated, resulting in a large gradient even for a small loss value.…”

“…Linear enhancement can only be enhanced within a certain gray range of the image, which has great limitations [5], [6]. Nonlinear enhancement, such as Gaussian function transformation, does not significantly enhance the edge information of the image [7]. In the frequency domain, the enhancement algorithm based on directional filter banks has problems, such as reduced image sharpness and it is missing partial features after enhancement [8].…”

Gradient entropy metric and p-Laplace diffusion constraint-based algorithm for noisy multispectral image fusion

“…A common method for this is using the least squares [27][28][29]. The visualization can be mathematically expressed as:…”

An Adaptive Detail Equalization for Infrared Image Enhancement Based on Multi-Scale Convolution

“…The gradient domain mainly reflects the details information of the image, and it can display the image gradient value in the gray level as possible. Zhao et al [23] introduced the gradient domain into the histogram specification and reconstructed the gradient field of the target image by means of weighted fusion of bimodal Gaussian function and original image. In the wavelet domain, the edge and detail features of the image correspond to the high frequency coefficients, and the background and contour information of the image correspond to the low frequency coefficients [21], [24], [32].…”