CAJ | 학술논문

对数字热成像的噪点信息矢量量化是实现图像去噪，提高图像识别能力的关键技术。传统方法中对数字热成像中噪点信息矢量量化算法采用LBG算法，当图像出现灰色关联度噪点时，量化算法的抗噪性能不好。提出一种基于强边缘保留和误差补偿编码的数字热成像中噪点信息矢量量化算法。用量化误差补偿法对量化的热成像进行补偿，消除当前图像块的瑕疵特征点，提高误差补偿编码处理精度。仿真表明，本算法对数字热成像噪点信息的矢量量化效果较好，去噪性能优越，进行矢量量化处理后，实现噪点去除，PSNR值较传统方法都有明显提高,提高对图像目标的识别性能，较传统方法优越。
대수자열성상적조점신식시량양화시실현도상거조，제고도상식별능력적관건기술。전통방법중대수자열성상중조점신식시량양화산법채용LBG산법，당도상출현회색관련도조점시，양화산법적항조성능불호。제출일충기우강변연보류화오차보상편마적수자열성상중조점신식시량양화산법。용양화오차보상법대양화적열성상진행보상，소제당전도상괴적하자특정점，제고오차보상편마처리정도。방진표명，본산법대수자열성상조점신식적시량양화효과교호，거조성능우월，진행시량양화처리후，실현조점거제，PSNR치교전통방법도유명현제고,제고대도상목표적식별성능，교전통방법우월。
The noise information of vector quantization for digital thermal imaging is the realization of image denoising, the key technology to improve image recognition ability. In the traditional method of noise information vector quantization algo?rithm for digital thermal imaging using LBG algorithm, when the image appears gray correlation noise, anti noise perfor?mance quantization algorithm is not good. A noise information vector quantization algorithm for digital thermal imaging strong edge preserving and error compensation coding based on. To retain the compression processing digital thermal imag?ing strong edges, using strong edge preserving compression, the elimination of the current image block defect feature point, improve the processing accuracy error compensation code. Simulation results show that the better effect of the vector quanti?zation algorithm for digital thermal imaging noise information is obtained, the denoising performance is superior, vector quantization is performed after treatment, achieves the noisy filter, improves the recognition performance of the image, it compared with the traditional method is superior.