本文選題：仿生視覺 + 多光譜��； 參考：《紅外與激光工程》2017年09期

【摘要】：針對傳統(tǒng)仿生視覺系統(tǒng)中目標圖像獲取單一性問題,提出一種仿人眼自適應(yīng)調(diào)節(jié)的多光譜視覺成像技術(shù)。首先,通過改進的自動調(diào)焦算法使成像系統(tǒng)同時采集可見光高分辨率圖像及近紅外低分辨率圖像。然后,對于多光譜成像系統(tǒng)中由于分光棱鏡折射率不同導致的在固定焦距下,可見光和近紅外圖像清晰度有所不同的問題,采用改進的二代小波變換進行近紅外圖像增強,提高圖像對比度,改善視覺效果。最后,搭建基于液體變焦透鏡的多光譜實驗系統(tǒng)驗證自動調(diào)焦算法及圖像增強算法的實際性能。實驗結(jié)果表明:系統(tǒng)完成有效自動調(diào)焦的平均用時為756 ms。同時,近紅外圖像增強后其灰度方差函數(shù)值提高了79.4%,解決了對比度低和細節(jié)模糊的問題,最終實現(xiàn)自適應(yīng)調(diào)節(jié)。
[Abstract]:Aiming at the singularity of target image acquisition in traditional bionic vision system, a multispectral visual imaging technique based on human eye adaptive adjustment is proposed. Firstly, the improved auto-focusing algorithm enables the imaging system to simultaneously collect high resolution image of visible light and low resolution image of near infrared. Then, for the problem that the clarity of the visible and near infrared images in the fixed focal length is different due to the different refractive index of the splitter prism in the multispectral imaging system, the improved second generation wavelet transform is used to enhance the near infrared image. Improve image contrast and visual effect. Finally, a multi-spectral experimental system based on liquid zoom lens is built to verify the actual performance of the auto-focusing algorithm and image enhancement algorithm. The experimental results show that the average time for the system to complete effective auto-focusing is 756 Ms. At the same time, the gray level variance function of NIR image is increased by 79.4 after enhancement, which solves the problems of low contrast and fuzzy details, and finally realizes adaptive adjustment.
【作者單位】： 北京信息科技大學光電信息與儀器北京市工程研究中心;北京信息科技大學光電測試技術(shù)北京市重點實驗室;
【基金】：教育部“創(chuàng)新團隊發(fā)展計劃”(IRT_16R07)
【分類號】：TP391.41

【相似文獻】

相關(guān)期刊論文 前2條

1 姜濤;肖迎元;袁曉潔;;RFID讀寫器功率的自適應(yīng)調(diào)節(jié)策略[J];計算機工程;2010年20期

2 周毅,朱虹,辛威,馬天德;一種擬人視覺自適應(yīng)調(diào)節(jié)原理的數(shù)字補光方法[J];西安理工大學學報;2001年02期

，

本文編號：2054180