College of Computer Science and Technology, Xi'an University of Science and Technology, Xi'an 710054, China.
Comput Intell Neurosci. 2021 Nov 12;2021:2436486. doi: 10.1155/2021/2436486. eCollection 2021.
The lighting facilities are affected due to conditions of coal mine in high dust pollution, which bring problems of dim, shadow, or reflection to coal and gangue images, and make it difficult to identify coal and gangue from background. To solve these problems, a preprocessing model for low-quality images of coal and gangue is proposed based on a joint enhancement algorithm in this paper. Firstly, the characteristics of coal and gangue images are analyzed in detail, and the improvement ways are put forward. Secondly, the image preprocessing flow of coal and gangue is established based on local features. Finally, a joint image enhancement algorithm is proposed based on bilateral filtering. In experimental, -means clustering segmentation is used to compare the segmentation results of different preprocessing methods with information entropy and structural similarity. Through the simulation experiments for six scenes, the results show that the proposed preprocessing model can effectively reduce noise, improve overall brightness and contrast, and enhance image details. At the same time, it has a better segmentation effect. All of these can provide a better basis for target recognition.
煤矿中的高粉尘污染会影响照明设施，导致煤和矸石图像出现昏暗、阴影或反射等问题，使得从背景中识别煤和矸石变得困难。为了解决这些问题，本文提出了一种基于联合增强算法的煤和矸石低质量图像预处理模型。首先，详细分析了煤和矸石图像的特点，并提出了改进方法。其次，基于局部特征建立了煤和矸石的图像预处理流程。最后，提出了一种基于双边滤波的联合图像增强算法。在实验中，使用 -means 聚类分割来比较不同预处理方法的分割结果与信息熵和结构相似度。通过对六个场景的模拟实验，结果表明，所提出的预处理模型可以有效地降低噪声，提高整体亮度和对比度，增强图像细节。同时，它具有更好的分割效果。所有这些都可以为目标识别提供更好的基础。
