In recent years, holography has developed rapidly, and has been applied to many scientific and technical fields. It has become a widely used experimental device and light source in three-dimensional (3D) imaging and display. After years of research and development, digital holography in visible wavelengths has become a crucial observation and detection method in the imaging field. However, the limitations of traditional visible light digital holography are becoming more and more obvious in the industrial area, where specific environments with smoke and flame interference exist. Scholars found that the ability of infrared (IR) light to pass through high-density particle fields is significantly higher than that of visible light, and infrared holography can be imaged through smoke and flames. This opens up new applications for holographic technology. However, in the smoke environment, the infrared digital holographic reconstruction image has serious scattering noise and prominent shadow areas, which can no longer meet the requirements for imaging quality. In this study, we propose an algorithm that applies to the reconstructed image of an infrared hologram in the presence of smoke motion. This algorithm can effectively suppress the scattering noise and achieve the effect of brightness enhancement. Thus, high-quality infrared hologram reconstructed images can be obtained using this algorithm.
The optical path of this study is designed and built based on the Mach-Zehnder interferometric optical path (Fig. 1). The optical path consists of laser, beam splitter, reflector, collimated beam expansion system, smoke generation device, and light sensor. The hologram acquisition time interval is set to keep the experimental setup stable and acquire 15 samples of infrared holograms under a relatively complete smoke change cycle. The algorithm in this paper is divided into four main steps: image pre-processing, segmentation of the target object, segmentation of the shadow region, and brightness enhancement of the shadow region. The specific algorithm flow is shown in Fig. 4. First, the image is pre-processed using a bilateral filter to remove significant scattering noise. Next, the image is binarized, the image edge information is extracted for expansion and filling operations, and the target object is segmented by combining the maximum connected area algorithm. An edge extraction algorithm and expansion and filling operations are then used to segment the shadow region from the target object area. Finally, an optimized sparrow search algorithm is used to enhance the shaded areas, and the brightness-enhanced reconstructed infrared holographic image is output.
In this paper, six infrared holograms with significant continuous smoke changes were selected for reconstruction. The brightness enhancement algorithm of infrared digital holograms through smoke was used to enhance the brightness of the shadow region. Finally, the comparison results were obtained between the reconstructed and enhanced images (Fig. 10). We can judge and analyze the enhancement effect from Fig. 10 that the proposed algorithm eliminates the speckle noise of the image and retains the texture details. The shadow region’s brightness enhancement effect is obvious. Taking three infrared holographic reconstructed images [Figs. 10(a)-10(c)] as examples, four algorithms are compared (Fig. 11). The comparison shows that although the other three algorithms have their advantages, the final image quality cannot meet the experimental requirements. The proposed algorithm gives the best results. Next, the evaluation parameters including peak signal-to-noise ratio (PSNR), mean square error (MSE), and feature similarity index measure (FSIM) are calculated separately, and the results of the four algorithms are compared (Table 1). The PSNR value of the proposed algorithm is better than those of the other three algorithms, the MSE value is the lowest among the four algorithms, and the calculated FSIM value is also the closest to 1. The results show that the proposed algorithm achieves good luminance enhancement in the shadow region. It can achieve the brightness enhancement effect while suppressing the noise and maintaining the original structure of the reconstructed infrared hologram image with smoke motion.
The through-smoke infrared holographic reconstruction image has many problems, such as heavy scattering noise, uneven brightness and lack of object detail information. We propose a new image brightness enhancement algorithm, which can effectively improve the image quality of the through-smoke infrared holographic reconstructed image, with significant effects in brightness enhancement and removal of scattering noise. The experimental results show that, compared with other algorithms, the brightness enhancement algorithm of the through-smoke infrared digital holographic image has a better image enhancement effect, suppressing the scattering noise and showing evident brightness enhancement of the shadow region. In addition, from the analysis of image evaluation indexes, it is seen that the PSNR values of the proposed algorithm are all greater than 21, and they are higher than those of the comparison algorithms, indicating that the proposed algorithm has a better noise suppression effect. The MSE values are all less than 411, the smallest among the comparison algorithms, indicating that the image quality after processing by the proposed algorithm is the best. The FSIM values are all greater than 0.9511, the closest to 1. Compared with other algorithms, the results processed by the proposed algorithm have the highest similarity with the reconstructed images without smoke, exhibiting the complete structural recovery and the minor distortion. In summary, the algorithm proposed in this paper has a noticeable brightness enhancement effect in the shadow region, can intelligently process the acquired holograms, and can effectively improve the image quality of infrared holographic imaging in smoke permeable environment.
