發(fā)布時(shí)間：2018-05-18 18:47

  本文選題：智能交通 + 目標(biāo)檢測(cè)�。� 參考：《華南理工大學(xué)》2014年博士論文

【摘要】：視頻目標(biāo)的檢測(cè)、識(shí)別是目前智能交通和計(jì)算機(jī)視覺(jué)領(lǐng)域中的一個(gè)重要研究方向。但是，由于檢測(cè)和識(shí)別環(huán)境下存在背景復(fù)雜、光照變化、目標(biāo)遮擋等原因，導(dǎo)致該應(yīng)用仍面臨著許多困難，檢測(cè)和識(shí)別的魯棒性及準(zhǔn)確性都有待進(jìn)一步提高。 本論文對(duì)視頻目標(biāo)檢測(cè)和識(shí)別中的幾個(gè)關(guān)鍵問(wèn)題進(jìn)行了研究，主要包括：復(fù)雜場(chǎng)景下目標(biāo)與背景、陰影的準(zhǔn)確分割；對(duì)提取的前景目標(biāo)準(zhǔn)確分類；復(fù)雜背景下的目標(biāo)識(shí)別。針對(duì)這些問(wèn)題，本論文提出了相應(yīng)的解決方法。具體工作如下： 1.提出了一種基于自適應(yīng)模糊估計(jì)的背景建模方法。該方法從函數(shù)估計(jì)的角度對(duì)背景進(jìn)行建模，并采用TSK模糊系統(tǒng)作為估計(jì)函數(shù)。為了訓(xùn)練函數(shù)估計(jì)算子，分別使用粒子群優(yōu)化（PSO）算法和遞歸最小二乘估計(jì)（RLSE）算法來(lái)優(yōu)化模糊系統(tǒng)的前件參數(shù)和后件參數(shù)。為了有效估計(jì)背景，將前景像素看作背景像素的異常樣例，并提出了異常樣例的去除方法，然后用去除后的結(jié)果去訓(xùn)練模糊估計(jì)算子。該方法在動(dòng)態(tài)背景、光照變化、攝像機(jī)振動(dòng)等環(huán)境下都具有較高的運(yùn)行效率和檢測(cè)效果。 2.提出了一種基于模糊積分的運(yùn)動(dòng)陰影檢測(cè)方法。在提取前景區(qū)域的基礎(chǔ)上，選擇顏色和紋理作為陰影檢測(cè)的特征，并分別定義了這兩種特征的相似性和重要性測(cè)度函數(shù)，然后通過(guò)Choquet模糊積分將這兩種特征融合，實(shí)現(xiàn)陰影和前景目標(biāo)的分類，最后通過(guò)后續(xù)處理，找到真正的陰影區(qū)域。 3.提出了一種基于JointBoost I2C距離度量的目標(biāo)分類方法。針對(duì)經(jīng)典I2C距離計(jì)算量大且易受噪聲干擾等不足，首先提出了一種原型特征集的生成方法，該集合中的樣本數(shù)量較少，但更具有代表性，計(jì)算測(cè)試圖像到該原型特征集的距離花費(fèi)較少時(shí)間；然后借助JointBoost算法的思想，聯(lián)合多個(gè)I2C距離度量生成一個(gè)強(qiáng)分類器；最后還提出了一種將空間信息融合到強(qiáng)分類器的方法。實(shí)驗(yàn)證明，該方法在前景目標(biāo)和圖像分類實(shí)驗(yàn)中，具有更高的分類性能。 4.提出了基于特征碼本樹(shù)和能量最小化的目標(biāo)識(shí)別方法。該方法考慮了特征的空間位置信息和特征之間的空間關(guān)系，集成了目標(biāo)檢測(cè)和目標(biāo)識(shí)別。首先從目標(biāo)圖像提取的大量特征中過(guò)濾掉噪聲特征；然后對(duì)單特征和空間上鄰近的串聯(lián)雙特征分別使用層次k均值聚類算法構(gòu)建特征碼本樹(shù)，，利用樹(shù)模型可以實(shí)現(xiàn)特征快速定位和分類；最后建立一個(gè)能量函數(shù)來(lái)融合單、雙特征碼本樹(shù)的類別概率匹配結(jié)果，并通過(guò)在測(cè)試圖像中尋找滑動(dòng)窗口所在區(qū)域的能量最小化來(lái)確定所屬類別目標(biāo)的位置。 5.提出了基于優(yōu)化Hough森林代價(jià)損失的目標(biāo)識(shí)別方法。首先在充分利用訓(xùn)練圖像中對(duì)象位置是已知的基礎(chǔ)上,提出了改進(jìn)的偏移量不確定性度量方法；其次借助Boosting算法的思想，學(xué)習(xí)圖片塊樣本和目標(biāo)對(duì)象樣本的自適應(yīng)權(quán)重分布，并分別優(yōu)化用于構(gòu)造隨機(jī)樹(shù)和Hough森林的代價(jià)損失函數(shù)；最后根據(jù)圖片塊樣本的權(quán)重分布，提出了改進(jìn)的類標(biāo)志不確定性度量方法�；贖ough森林的代價(jià)損失函數(shù)，還提出了隨機(jī)樹(shù)權(quán)重的學(xué)習(xí)方法。
[Abstract]:The detection and recognition of video targets is an important research direction in the field of intelligent traffic and computer vision. However, because of the complicated background, illumination change and target occlusion in the detection and recognition environment, the application still faces many difficulties. The robustness and accuracy of detection and recognition need to be further improved.
In this paper, several key problems in video target detection and recognition are studied, including: the target and background of the complex scene, the accurate segmentation of the shadow, the accurate classification of the foreground object and the target recognition under the complex background.
1. a background modeling method based on adaptive fuzzy estimation is proposed. This method models the background from the angle of function estimation and uses the TSK fuzzy system as the estimation function. In order to train the function estimation operator, the particle swarm optimization (PSO) algorithm and the recursive least double multiplicative estimation (RLSE) algorithm are used to optimize the pre fuzzy system. In order to effectively estimate the background, the foreground pixels are considered as an abnormal example of the background pixels, and the removal method of the anomaly samples is proposed. Then the fuzzy estimation operator is trained by the removal results. The method has high efficiency and detection in the dynamic background, the illumination change, the camera vibration and so on. Effect.
2. a motion shadow detection method based on fuzzy integral is proposed. On the basis of extracting foreground region, color and texture are selected as the feature of shadow detection, and the similarity and importance measure function of the two features are defined respectively. Then the two features are fused by Choquet fuzzy integral to realize the shadow and foreground object. Classification, and finally through the subsequent processing, find the real shadow area.
3. a target classification method based on JointBoost I2C distance measurement is proposed. In view of the shortage of classical I2C distance computation and easy to be disturbed by noise interference, a new method of generating prototype feature sets is proposed. The number of samples in the set is less, but more representative, the distance cost of the test image to the prototype feature set is calculated. Less time; then the idea of JointBoost algorithm is used to combine multiple I2C distance metrics to generate a strong classifier. Finally, a method of fusion of spatial information to a strong classifier is proposed. Experiments show that the method has a higher classification performance in the foreground object and the image classification experiment.
4. a target recognition method based on characteristic codebook tree and energy minimization is proposed. This method takes into account the spatial location information of the feature and the spatial relationship between features, and integrates target detection and target recognition. First, the noise features are filtered out from the large number of features extracted from the target image, and then the single feature and the adjacent space in the space are connected in series. The double feature uses the hierarchical K mean clustering algorithm to construct the characteristic tree tree. The tree model can be used to locate and classify the features quickly. Finally, an energy function is established to fuse the probability matching results of the single, double feature codebook, and to find the energy minimization of the region in which the sliding window is located in the test image. The position of the category target.
5. the target recognition method based on optimized Hough forest cost loss is proposed. Firstly, based on the known location of the object in the training image, an improved measurement method of offset uncertainty is proposed. Secondly, the adaptive weight distribution of the sample of picture block and target object is learned with the help of the thought of the Boosting algorithm. The cost loss functions used to construct random trees and Hough forests are optimized respectively. Finally, based on the weight distribution of the block samples, an improved method for measuring the uncertainty of the class marks is proposed. Based on the cost loss function of the Hough forest, the learning method of the weight of the random tree is also proposed.
【學(xué)位授予單位】：華南理工大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2014
【分類號(hào)】：U495

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 張澤旭,李金宗,李寧寧;基于光流場(chǎng)分割和Canny邊緣提取融合算法的運(yùn)動(dòng)目標(biāo)檢測(cè)[J];電子學(xué)報(bào);2003年09期

2 王志明;張麗;包宏;;基于混合結(jié)構(gòu)神經(jīng)網(wǎng)絡(luò)的自適應(yīng)背景模型[J];電子學(xué)報(bào);2011年05期

3 李玲玲;金泰松;李翠華;;基于局部特征和隱條件隨機(jī)場(chǎng)的場(chǎng)景分類方法[J];北京理工大學(xué)學(xué)報(bào);2012年07期

4 姜柯;李艾華;蘇延召;;基于全局紋理和抽樣推斷的自適應(yīng)陰影檢測(cè)算法[J];光電子.激光;2012年11期

5 張超;吳小培;周建英;戚培慶;王營(yíng)冠;呂釗;;基于改進(jìn)高斯混合建模和短時(shí)穩(wěn)定度的運(yùn)動(dòng)目標(biāo)檢測(cè)算法[J];電子與信息學(xué)報(bào);2012年10期

6 李文輝;倪洪印;;一種改進(jìn)的Adaboost訓(xùn)練算法[J];吉林大學(xué)學(xué)報(bào)(理學(xué)版);2011年03期

7 李闖;丁曉青;吳佑壽;;一種改進(jìn)的AdaBoost算法——AD AdaBoost[J];計(jì)算機(jī)學(xué)報(bào);2007年01期

8 查宇飛;楚瀛;王勛;馬時(shí)平;畢篤彥;;一種基于Boosting判別模型的運(yùn)動(dòng)陰影檢測(cè)方法[J];計(jì)算機(jī)學(xué)報(bào);2007年08期

9 凌志剛;趙春暉;梁彥;潘泉;王燕;;基于視覺(jué)的人行為理解綜述[J];計(jì)算機(jī)應(yīng)用研究;2008年09期

10 戴斌;方宇強(qiáng);孫振平;王亮;;基于光流技術(shù)的運(yùn)動(dòng)目標(biāo)檢測(cè)和跟蹤方法研究[J];科技導(dǎo)報(bào);2009年12期

本文編號(hào)：1906779