CN104915918A - High-robustness digital watermarking method for image maps - Google Patents

High-robustness digital watermarking method for image maps Download PDF

Info

Publication number
CN104915918A
CN104915918A CN 201510349570 CN201510349570A CN104915918A CN 104915918 A CN104915918 A CN 104915918A CN 201510349570 CN201510349570 CN 201510349570 CN 201510349570 A CN201510349570 A CN 201510349570A CN 104915918 A CN104915918 A CN 104915918A
Authority
CN
Grant status
Application
Patent type
Prior art keywords
map
watermark
value
random sequence
pj
Prior art date
Application number
CN 201510349570
Other languages
Chinese (zh)
Other versions
CN104915918B (en )
Inventor
孙建国
李佳楠
李博权
Original Assignee
哈尔滨工程大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date

Links

Abstract

The invention specifically relates to a high-robustness digital watermarking method for image maps. The high-robustness digital watermarking method comprises the steps of reading a linked list of nodes according a data structure of an IMG image map, and acquiring a pixel value of each node according to a corresponding field; acquiring watermarking binary codes, and generating a corresponding pseudorandom sequence; setting a value which is less than one pixel unit as a threshold value according to the map scale and the size of the pseudorandom sequence, and translating the map in eight directions; calculating a difference value between each feature point and a pixel point around, reading the pseudorandom sequence, and accumulating a product of a corresponding random sequence value and the pixel value difference to the pixel value of the feature point; and randomly selecting several of the maps in which a watermark is embedded to publish. The method provided by the invention realizes high-robustness watermark embedment and extraction. A watermark embedding carrier is based on the brightness difference of images, the watermark is embedded into a plurality of maps with extremely strong correlation, and the watermark can be restored through solving the correlation and an average pixel value of the maps once the watermark is damaged.

Description

一种面向影像地图的强鲁棒数字水印方法 Strong robust digital watermarking method oriented image map

技术领域 FIELD

[0001] 本发明具体涉及的是一种面向影像地图的强鲁棒数字水印方法。 [0001] The present invention particularly relates to a method for digital watermark robustness image map.

背景技术 Background technique

[0002] 遥感影像(卫星,航空,地面近景)作为对地观测获取地球表面覆盖与结构信息的载体,在地学分析应用领域是不可或缺的信息源。 [0002] The remote sensing images (satellite, aerial, terrestrial close-range) as a surface covering for earth observation acquired configuration information carrier earth, in the field of the analysis applications are indispensable information source. 而如何将地理影像转化为有价值的信息对成功实施GIS和制图工程又是至关重要的。 And how the geographic imagery into valuable information for the successful implementation of GIS and mapping projects is crucial. 目前,在我们周围越来越多的人们能够利用全范围的地理影像产品来提取和使用有价值的信息。 At present, around us more and more people can take advantage of the full range of geographic imaging products to extract and use the valuable information.

[0003] 作为众多数字地图的原始数据来源,影像地图的真实性和安全性受到格外关注, 既要维护地图所有者的合法利益,又要缺乏地图内容不被篡改,为此,影像地图数字水印研宄得到了较为快速的发展。 [0003] as the original source of many digital map data, image maps of authenticity and safety has received special attention, both to safeguard the legitimate interests of the owner of the map, but also the lack of map content is not tampered with, to this end, the digital watermark image maps a Subsidiary obtained a more rapid development.

[0004] 目前,对影像地图的数字水印算法很少,对于图像的水印算法主要包括两类:即空间域和频率域。 [0004] Currently, digital watermarking algorithm for image maps of little watermarking algorithm for image mainly include two categories: spatial domain and frequency domain. 空间域算法主要是通过修改像素信息来嵌入水印,算法实现简单,但鲁棒性较差,对图像的损伤也较大,无法抵抗多种攻击;频率域算法通过离散余弦变换、傅立叶变换或小波变换等数学方法获得频域系数,通过调整频域系数来嵌入水印,该类算法较复杂, 对图像的质量也有一定影响,且无法抵抗多种组合攻击。 Spatial domain algorithm is mainly used to embed a watermark by modifying the pixel information, the algorithm is simple, but less robust, damage to the image is large, can not resist a variety of attacks; in the frequency domain by a discrete cosine transform algorithm, a Fourier transform or wavelet mathematical transform to obtain frequency-domain coefficients, by adjusting the coefficients in frequency domain watermark is embedded, such more complicated algorithm, also have some impact on the quality of the image, and can not resist the attack various combinations.

[0005] 基于空间域算法和频率域算法的不足,一些学者提出了无损影地图水印的算法, 主要要研宄如下:王贤敏等在2004年提出了小波用于基于遥感影像特征的自适应二维盲水印算法;王勋等提出了一种鲁棒的矢量地图数字水印算法;刘九芬等提出了一种抗几何攻击的小波变换域图像水印算法;张军等发明了用于图像认证的基于神经网络的水印技术;沃焱等提出了基于视觉特性的灰度级自适应盲水印算法。 [0005] Based on the lack of space domain and frequency domain algorithm algorithm, some scholars proposed algorithm lossless watermark shadow map, mainly to study based on the following: Wang Xianmin such as an adaptive two-dimensional wavelet based remote sensing image characteristics in 2004 watermarking algorithm; Xun Wang et vector map presents Robust digital watermarking algorithm; Liujiu Fen put forward wavelet transform domain image watermarking algorithm against geometric attacks; Zhang et invented image authentication based on neural network the watermark technology; Yan et fertile gradation proposed adaptive blind watermarking based on visual characteristics. 本算法为基于面向影像地图的强鲁棒水印书的算法,与以上算法相比,让空间域算法和频率域算法更优化。 This algorithm is a strong robust watermarking algorithm for image map book based on comparison with the above algorithm, so that the spatial domain algorithm and the frequency domain algorithm is more optimized.

发明内容 SUMMARY

[0006] 本发明的目的在于提供一种能有效抵抗多种攻击,具有高安全性以及不可见性的面向影像地图的强鲁棒数字水印方法。 [0006] The object of the present invention is to provide a more effective against attack, a strong robust digital watermarking methods for images of high security and map invisibility.

[0007] 本发明包括如下步骤: [0007] The present invention comprises the steps of:

[0008] (1)根据MG影像地图的数据结构,读取节点的链表,并按照对应字段分别获得每个节点的像素值; [0008] (1) The MG map image data structure, reading the list of nodes, and each pixel value of each node in a corresponding field;

[0009] (2)扫描水印位图,获得水印二进制编码,生成相应的伪随机序列L, [0009] (2) scanning a bitmap watermark, the watermark is obtained binary encoding, generates a corresponding pseudo-random sequence L,

Figure CN104915918AD00031

[0011] 其中,La e {+1,-1},L是一个由N个元素组成的伪随机序列,N是水印编码的长度; [0011] where, La e {+ 1, -1}, L is a pseudo-random sequence of N elements, N is the length of the watermark encoder;

[0012] (3)根据地图规模和伪随机序列大小,设定小于1个像素单位的值作为阈值,对地图进行八个方向的平移,包括上T、右上RT、右R、右下RD、下D、左下LD、左L、左上LT ; [0012] (3) the scale of the map and the size of the pseudo-random sequence, set to less than one pixel unit as a threshold value, the map translational eight directions, including T, RT right upper, and right R, bottom right RD, the D, the lower left LD, left L, left upper LT;

[0013] (4)对于位移后的八幅地图,按照顺时针顺序,随机选取其中3幅地图,联同原始地图共同组成待嵌入水印的地图集合{V| Vi|l < i <4}; [0013] (4) for the displacement map by eight, is clockwise, wherein three randomly selected map, the map associated with the original watermark to be embedded together form a map set {V | Vi | l <i <4};

[0014] (5)利用人类视觉系统,检测每一幅地图,据地图集合IVlviIl彡i彡4},按照每个节点仅嵌入1比特水印的编码规则,选取最大规模的特征点集合,同时提取对应节点的像素值; [0014] (5) using the human visual system, each detecting a map, a map data set IVlviIl San San 4} i, according to the coding rules of the watermark embedding is only 1 bit for each node, and selecting the largest set of feature points while extracting corresponding pixel values ​​of the nodes;

[0015] (6)计算每一个特征点与周围像素点的差值,并读取伪随机序列,将对应随机序列值与像素差值的乘积累加到该特征点的像素值; [0015] (6) calculating a difference value for each feature point and pixels around the point, and reads the pseudo-random sequence, the random sequence corresponding accumulated value by the pixel value of pixel difference values ​​is added to the feature point;

[0016] (7)对于嵌入水印后的地图,随机选取其中几幅发布; [0016] (7) The map watermarked, wherein the several randomly selected publisher;

[0017] (8)在水印检测时,将待检测地图与其它嵌入水印但未发布的地图相互融合,求得具有平均像素值的地图; [0017] (8) during watermark detection, and the other map to be detected but watermark embedding merging published map, the map is obtained having an average pixel value;

[0018] (9)根据原始的水印的伪随机序列,以及融合后的水印地图,计算关联系数,如果相关系数的绝对值大于等于所设定的阈值,则表明检测到水印编码,否则未检测到; [0018] (9) The pseudo-random sequence of the original watermark, the watermark and fused map, calculate correlation coefficient, if the absolute value of the correlation coefficient is not less than the set threshold value, it indicates that the watermark is detected, otherwise it is not detected to;

[0019] (10)将编码还原为水印位图,并计算相似度。 [0019] (10) reduced to encoding the watermark bitmap, and calculates the similarity.

[0020] 所述的待嵌入水印的地图集合的生成步骤包括: [0020] The step of generating a map to be watermarked comprising the set of:

[0021] 对于每幅地图Vi,计算对应的水印嵌入强度: [0021] For each of maps Vi, the corresponding calculated watermark embedding strength:

[0022] S (Vi) = {s (Pj) >0 II ^ j ^ N} [0022] S (Vi) = {s (Pj)> 0 II ^ j ^ N}

[0023] 其中,S(Pj)表示地图中特征点Pj在HVS系统内的不可察觉程度,即P j与其他周围其他八个像素节点的亮度差值; [0023] where, S (Pj) denotes the map extent not perceive characteristic points Pj in HVS system, i.e., the luminance difference value of other nodes P j eight pixels surrounding the other;

[0024] 依次读取水印序列,对每个特征点: [0024] sequentially read the watermark sequence, for each feature point:

[0025] s' (Pj) = s (Pj)+s (Pj) Lj, [0025] s' (Pj) = s (Pj) + s (Pj) Lj,

[0026] Y1' =V^s(Pj)Lj, [0026] Y1 '= V ^ s (Pj) Lj,

[0027] 得到{V, |v/ I 1 彡i 彡4}。 [0027] to give {V, | v / I 1 i San San 4}.

[0028] 本发明的有益效果在于 Advantageous Effects [0028] The present invention is

[0029] 本发明完全基于影像地图的色彩特性,通过多图共存的方式,保持地图的基本像素特征,已达到水印横存的目的,具体包括: [0029] The present invention is based solely on the color characteristics of the image map, by the coexistence of multi FIG embodiment, the basic pixel map holding characteristics, has the purpose of cross-stored watermark comprises:

[0030] 实现了强鲁棒性的水印嵌入和提取。 [0030] to achieve a strong robust watermark embedding and extraction. 水印嵌入载体基于影像的亮度差值,且嵌入多幅相关性极强的地图内,水印一旦遭到破坏,可通过求取相关性和地图的平均像素值而得以恢复。 Embedding vector difference value based on the luminance of the image, and map embedded in the plurality of correlation strong watermark once destroyed, it may be restored by obtaining the average pixel value and the correlation maps.

[0031] 极强的鲁棒性。 [0031] strong robustness. 经过测试,水印遭到严重破坏,只需影像地图载体具有可用性,则可通过相关地图进行平均恢复,获得相似度较高的水印信息。 After testing, the watermark was severely damaged, the carrier has just image maps available, can be restored by average map, get high similarity watermark information.

[0032] 较好的隐蔽性。 [0032] better concealment. 在水印嵌入节点的选择上结合了HVS原理,特征点都是一些视觉不易察觉的位置,且通过计算亮度差值,有效控制了由于水印嵌入造成的色彩扰动,由于嵌入的位置极为隐蔽,攻击者很难专门地对水印发起攻击。 In the choice of embedding HVS node combines principles, some visual feature points are difficult to detect a position, and by calculating the luminance difference value, since the effective control of the watermark embedding color disturbance caused due to the embedding location very subtle, the attacker it is difficult to specifically watermark attack.

附图说明 BRIEF DESCRIPTION

[0033] 图1影像地图文件结构; [0033] FIG 1 image map file structure;

[0034] 图2a影像地图原始地图; [0034] FIG. 2a original map image map;

[0035] 图2b影像地右侧平移效果图; [0035] FIG. 2b imagewise FIG right panning effect;

[0036] 图2c影像地右下平移效果图; [0036] FIG. 2c imagewise FIG lower right panning effect;

[0037] 图3实验程序界面; [0037] FIG 3 experimental program interface;

[0038] 图4系统整体结构图; [0038] FIG 4 showing an overall system configuration;

[0039] 图5水印嵌入流程图。 [0039] The flowchart of FIG. 5 watermark embedding.

具体实施方式 detailed description

[0040] 下面结合实验实例对本发明做更详细地描述: [0040] The following examples of the present invention binding experiments described in more detail:

[0041] 本发明提出的是一种面向影像地图的强鲁棒水印方法。 [0041] The proposed invention is a potent method for robust watermarking image map. (1)读取影像地图文件,获得完整的数据结构;(2)按照IMG影像的数据结构,提取地图色彩特征值,对像素值进行编码;(3)结合设定阈值,将原始的地图按照上、下、左、右、左上、左下、右上、右下八个方向在阈值范围内随机平移;(4)从八幅平移得到的地图中随机选取4幅,并和原图共同组成图像序列;(5)制作水印位图并转换为二值序列,对应水印编码产生一个可供嵌入的伪随机序列;(6)对选定的图像序列,结合HVS(人类视觉系统)原理和阈值选取有限个特征点,计算特征点像素与其周围另外八个像素点的差值;(7)将像素差值与伪随机序列的乘积结果对应地叠加到特征点上;(8)递归完成上述操作,然后随机选取已嵌入水印的5幅地图中的任意几幅发布出去。 (1) reads the map image file, a complete data structure; (2) The data structure of image IMG, a color characteristic value extracting map, the pixel values ​​are encoded; (3) in conjunction with setting a threshold value, in accordance with the original map upper, lower, left, right, upper left, lower left, upper right, lower-right eight directions randomly translate within the threshold range; (4) obtained from the map eight randomly selected pan 4, and together constitute the original image sequence and ; (5) and FIG watermarking bit into a binary sequence corresponding to the watermark encoder for generating a pseudo-random sequence embedded; (6) on the selected image sequence, bind the HVS (human visual system) and the principle of thresholding limited feature points, the feature point calculating additional pixel and its surrounding eight pixels of the difference; (7) the product resulting pixel difference pseudorandom sequence corresponding to the feature point superimposed; (8) recursively completion of the operation, and then any randomly selected watermark is embedded in the five map pieces advertised.

[0042] 本发明是关于地图学及地理信息科学的信息处理方法,主要是一种MG格式影像地图的水印方法。 [0042] The present invention relates to an information processing method cartography and geographic information science, is primarily a map of MG format image watermarking method.

[0043] 本发明的目的是这样实现的:包括水印的嵌入和提取,其步骤包括: [0043] The object of the present invention is implemented: a watermark embedding and extraction, comprising the steps of:

[0044] (1)读取IMG影像地图文件,获得所有像素信息的链表。 [0044] (1) to read image IMG map files, get lists of all pixel information.

[0045] (2)扫描水印位图,获得水印二进制编码,并生成伪随机序列。 [0045] (2) scanning a bitmap watermark, the watermark is obtained binary encoding, and generates a pseudo-random sequence.

[0046] (3)设定阈值,并结合阈值,对地图进行八个方向的平移。 [0046] (3) setting a threshold value, and in conjunction with the threshold value, the map translational eight directions.

[0047] (4)得到位移后的八幅地图,随机选取其中三幅,联同原始地图一并作为待嵌入水印的地图。 [0047] (4) obtained after eight shift map, wherein the three randomly selected, together with the original map together with the map as a watermark to be embedded.

[0048] (5)利用HVS(人类视觉系统)系统,检测每幅地图,获取特征点。 [0048] (5) the use of the HVS (human visual system) system that detects each map, obtaining feature points.

[0049] (6)计算每一个特征点与周围像素点的差值,将对应随机序列值与像素差值的乘积累加到该特征点。 [0049] (6) calculating a difference value for each feature point and pixels around the point, and a corresponding random sequence by a value added to the accumulated pixel difference feature point.

[0050] (7)对于嵌入水印后的地图,随机选取其中几幅发布。 [0050] (7) for watermarked map, randomly selected among the several released.

[0051] (8)在水印检测时,将待检测地图与其它嵌入水印但未发布的地图相互融合,求得具有平均像素值的地图。 [0051] (8) during watermark detection, to be detected with other Map but watermark embedding released are blended to obtain a map with average pixel value.

[0052] (9)根据原始水印序列,以及融合后的水印地图,计算关联系数。 [0052] (9) The watermark map the original watermark sequence, and fusion, correlation coefficient is calculated. 检测水印编码。 Detect the watermark encoding.

[0053] (10)将检测到的编码还原为水印位图,并计算与原始水印的相似度。 [0053] (10) reduction of the detected encoded watermark bitmap, and calculates the similarity with the original watermark.

[0054] 本发明还可以包括: [0054] The present invention may further comprise:

[0055] 2、所述的读取MG影像地图文件的步骤中,直接根据MG金字塔型的数据组织结构,自链表头节点开始,读取每个节点的像素值字段。 [0055] 2, the step of reading the document image map MG, the MG directly from the data organization of the pyramid, beginning from the head node list, reads the pixel value field for each node.

[0056] 3、所述的伪随机序列生成的步骤中,具体方法为: [0056] 3, said step of generating a pseudo-random sequence, the specific method:

[0057] 根据定义好的水印位图,将位图转换为二进制的水印编码序列,然后依次扫描每个水印比特倌,#桉照加下挪_构律伪晡机序列L。 [0057] The defined watermark bitmap, the bitmap is converted to binary watermark coding sequence, and then sequentially scanning each watermark bit groom, eucalyptus irradiation and the # _ Norway Asr machine configuration law pseudo sequence L.

Figure CN104915918AD00051

[0059] 其中,La e {+1,-1},则L是一个由N个元素组成的伪随机序列,N是对应二进制水印编码的长度。 [0059] where, La e {+ 1, -1}, then L is a pseudo-random sequence of N elements, N is the length of the corresponding binary watermark code.

[0060] 4、所述的阈值设定和地图平移过程中,为了降低对于水印地图的过强扰动,限定平移单位不能超过1个像素,故阈值小于1,同时,为了起到混淆和扩散的目的,按照逆时针方向规定了八个平移方式,得到不同方向的八幅衍生地图。 [0060] 4, and the threshold is set according to the process of translating the map in order to reduce the map for the watermark is too strong disturbance, defined translation unit can not be more than one pixel, so the threshold is less than 1, while, in order to play the confusion and diffusion of the purpose, in accordance with the provisions of the eight translation counterclockwise manner, a derivative eight different directions map.

[0061] 5、所述的地图选取过程中,是为了平衡安全性和效率二者的关系,不能过多频繁的嵌入水印信息,故从混淆的衍生地图中选取3幅,并结合原始的地图一并作为可嵌入水印的地图对象{V| Vi|l彡i彡4}。 [0061] 5, the process of selecting the map in relation to both the safety and efficiency of balance, not too frequent embedding watermark information, it is selected from 3 confusion derived map, combined with the original map collectively as a map objects embedded watermark {V | Vi | l i San San 4}.

[0062] 6、所述的每幅地图特征点选取的过程中,是按照HVS(人类视觉系统)原理,扫描每一幅地图,找到视觉不敏感区域内的关键节点,最终节点的选取策略是参照每节点仅嵌入1比特水印的规则,选取最多元素集合。 [0062] 6, the process maps each feature point in the selection, in accordance with the HVS (Human Visual System) principle, a map of each scan, to find the key nodes insensitive visual region, is finally selected policy node Referring to embed only 1 bit per node watermark rules, select up elements of the set.

[0063] 7、所述的数字水印嵌入策略中,将水印伪随机序列同亮度差值的乘积叠加到视觉不敏感的特征点上,具体方法是: [0063] 7, the digital watermark embedding strategy, the product of a difference with the luminance watermark is superimposed on the pseudo-random sequence insensitive visual feature points, the specific method is:

[0064] 对于每幅地图Vi,计算对应的水印嵌入强度,即水印容量: [0064] For each of maps Vi, the corresponding calculated watermark embedding intensity, i.e. watermark capacity:

[0065] S (Vi) = {s (Pj) >0 II ^ j ^ N} [0065] S (Vi) = {s (Pj)> 0 II ^ j ^ N}

[0066] 其中,s (Pp表示特征点&在HVS系统内保持视觉不敏感情况下,p #其他周围其他八个像素节点的亮度差值。 [0066] wherein, s (Pp represents the feature point & insensitive visual holding case, the luminance difference value other eight pixels surrounding the other nodes in the p # HVS system.

[0067] 具体地,读取水印的伪随机序列值后,对地图内的每个特征点嵌入水印信息: [0067] Specifically, the value of the pseudo-random sequence to read the watermark, for each feature point watermark information is embedded within the map:

[0068] Sr (Pj) = s (Pj)+s (Pj) Lj [0068] Sr (Pj) = s (Pj) + s (Pj) Lj

[0069] 对每幅地图由此递归计算: [0069] Thus for each map recursive computation:

[0070] Vi' = Vj+s (Pj) Lj [0070] Vi '= Vj + s (Pj) Lj

[0071] 最终得到嵌入数字水印后的地图集合{V' I Vi' I 1彡i彡4}。 Map of the [0071] finally obtained digital watermark embedding set {V 'I Vi' I 1 i San San 4}.

[0072] 8、所述的水印地图发布过程中,即选取地图集合{V' |Vi' 中的部分地图发布。 [0072] 8. The process according to the published map of the watermark, i.e. select the map set {V '| Vi' published in the portion of the map.

[0073] 9、所述的水印信息检测过程中,将待检测的地图V'与其它嵌入水印但未发布的地图V tl相融合,根据阈值T,逆向平移V ^,并对待检测地图V'进行逆向恢复,求得平均像素值的地图样本t [0073] 9, the watermark detection process, the detection of the map to be V 'fused with other published map but watermark embedding V tl, according to the threshold T, the reverse translation of V ^, to be detected and the map V' reverse recovery, determine the average value of the pixel map sample t

[0074] 10、具体的水印提取方式为:根据原始水印Wtl的生成规则,重新生成伪随机序列L,计算L与地图样本P的关联度C : [0074] 10, in particular watermark extraction way: The original watermark Wtl generation rule, the pseudo-random sequence regenerate L, C L and the degree of association calculated map of the sample P:

Figure CN104915918AD00061

[0076] 如果|C|>T,表明地图内嵌入了疑似水印Wt;否则,表明伪随机序列不存在于地图内,即地图内无被检测水印,具体表示为: [0076] If | C |> T, the map shows that the embedded watermark Wt suspected; otherwise, show the pseudo-random sequence is not present in the map, i.e., no watermark is detected within the map, specifically expressed as:

Figure CN104915918AD00062

[0078] 特别地,对于单位水印比较过程中,有如下关系成立: [0078] In particular, the unit watermark for the comparison process, there is the following relationship:

[0080]其中,Wi ewt,Si Ci= c。 [0080] wherein, Wi ewt, Si Ci = c.

Figure CN104915918AD00071

[0081 ] 11、对于求得的水印序列,通过与原始水印的相似度计算,确定水印的真实性和地图的完整性。 [0081] 11, to the watermark sequence obtained by calculating the similarity of the original watermark, the watermark to determine the authenticity and integrity of the map. 对于检测到的水印标识Wt,由于地图受到攻击造成部分信息位丢失,造成^与原水印W内容不一致。 For detected watermark logo Wt, since the map is under attack causing some bit of information is lost, resulting in the original watermark W ^ inconsistent. 对此,需要进行相似度水印检测: In this regard, the need for similarity watermark detection:

Figure CN104915918AD00072

[0083] Wci为原始水印,Wt为检测出的水印。 [0083] Wci original watermark, Wt is detected watermark. N为水印容量。 N watermark capacity. 对于R 有两种情况:1.比特位数值相反;2.体特位数值为空。 For R, there are two cases: the opposite bit value of 1; 2 digits Laid body is empty...

[0084] 实施例1 [0084] Example 1

[0085] -种面向影像地图的强鲁棒数字水印方法,包括水印的嵌入和提取,其步骤包括: [0085] - Robust digital watermarking method of strongly-oriented image map, including embedding and extracting a watermark, comprising the steps of:

[0086] (1)读取MG影像地图文件,获得所有像素信息的链表。 [0086] (1) reads MG image map files, get lists of all pixel information.

[0087] (2)扫描水印位图,获得水印二进制编码,并生成相应的伪随机序列。 [0087] (2) scanning a bitmap watermark, the watermark is obtained binary encoding, and generates a corresponding pseudo-random sequence.

[0088] (3)根据地图规模和伪随机序列大小,自动设定阈值,并结合阈值,对地图进行八个方向的平移。 [0088] (3) the scale of the map and the size of the pseudo-random sequence, automatically set a threshold value, and in conjunction with the threshold value, the map translational eight directions.

[0089] (4)对于位移后的八幅地图,随机选取其中的三幅,联同原始地图一并作为待嵌入水印的地图。 [0089] (4) for the displacement map by eight, wherein the three randomly selected, together with the original map together with the map as a watermark to be embedded.

[0090] (5)利用HVS(人类视觉系统)系统,检测每一幅地图,获取其特征点及其像素值。 [0090] (5) the use of the HVS (human visual system) system, a map of each detection, and feature points acquired pixel values.

[0091] (6)计算每一个特征点与周围像素点的差值,并读取伪随机序列,将对应随机序列值与像素差值的乘积累加到该特征点的像素值。 [0091] (6) calculating a difference value for each feature point and pixels around the point, and reads the pseudo-random sequence, the random sequence of the corresponding pixel value of the pixel value difference is added to accumulated by the feature point.

[0092] (7)对于嵌入水印后的地图,随机选取其中几幅发布。 [0092] (7) for watermarked map, randomly selected among the several released.

[0093] (8)在水印检测时,由于应用过程中的地图会受到不同程度的改变,为此,将待检测地图与其它嵌入水印但未发布的地图相互融合,求得具有平均像素值的地图。 [0093] (8) When the watermark detection, since the map application process will be different degrees of change, for the map to be detected merging with other map but watermark embedding released, is obtained with an average pixel value map.

[0094] (9)根据原始的水印的伪随机序列,以及融合后的水印地图,计算关联系数。 [0094] (9) The watermark map the original pseudo-random sequence of watermark, as well as fusion, correlation coefficient is calculated. 如果相关系数的绝对值大于等于所设定的阈值,则表明检测到水印编码,否则未检测到。 If the absolute value of the correlation coefficient is greater than the threshold value is equal to the set, it indicates that the watermark is detected, otherwise it is not detected.

[0095] (10)通过上述方式,将编码还原为水印位图,并计算相似度。 [0095] (10) the above-described embodiment, the coding is reduced to the watermark bitmap, and calculates the similarity.

[0096] 根据IMG影像地图的数据结构,读取节点的链表,并按照对应字段分别获得每个节点的像素值。 [0096] The map data structure of the image IMG reads the linked list of nodes, and each pixel value of each node in a corresponding field.

[0097] 根据定义好的水印位图,将其表示为二进制的水印编码序列,并按照如下规则构建伪随机序列L。 [0097] The watermark bitmap defined, expressed as a binary watermark coding sequence, pseudo-random sequence and constructed in accordance with the following rules L.

Figure CN104915918AD00073

[0099] 其中,La e {+1,-1},L是一个由N个元素组成的伪随机序列,N是水印编码的长度。 [0099] where, La e {+ 1, -1}, L is a pseudo-random sequence of N elements, N is the length of the watermark encoder.

[0100] 随机选取一个小于1个像素单位的值作为阈值,并按照阈值大小按照顺时针方向分别平移得到八幅地图。 [0100] randomly selected value is less than one pixel unit as a threshold value, and according to a threshold size, respectively, in the clockwise direction to give eight translation map. 方向为:上(T)、右上(RT)、右(R)、右下(RD)、下(D)、左下(LD)、 左(L)、左上(LT)。 Direction: upper (T), upper right (RT), the right (R), lower right (RD), the (D), bottom left (LD), the left (L), upper left (LT).

[0101] 考虑鲁棒性和计算效率,按照顺时针顺序,随机选取其中3幅地图,联同原始地图共同组成待嵌入水印的地图集合IVlviIl彡i彡4}。 [0101] consideration of the robustness and computational efficiency, is clockwise, wherein three randomly selected map, the map associated with the original watermark to be embedded composed atlas set IVlviIl San i San 4}.

[0102] 根据地图集合IVlviIl彡i彡4},按照HVS (人类视觉系统)原理,按照每个节点仅嵌入1比特水印的编码规则,选取最大规模的特征点集合,同时提取对应节点的像素值。 [0102] pixel values ​​IVlviIl San San 4} i, according to the HVS (Human Visual System) principle, each node is only 1-bit encoding rules embedded watermark, selecting the largest set of feature points, and extracts the corresponding node according to the map set .

[0103] 对于每幅地图Vi,计算对应的水印嵌入强度: [0103] For each of maps Vi, the corresponding calculated watermark embedding strength:

[0104] S (Vi) = {s (Pj) >0 IK j < N} (2) [0104] S (Vi) = {s (Pj)> 0 IK j <N} (2)

[0105] 其中,S(Pj)表示地图中特征点Pj在HVS系统内的不可察觉程度,即P j与其他周围其他八个像素节点的亮度差值。 [0105] where, S (Pj) denotes the map extent not perceive characteristic points Pj in HVS system, i.e., the luminance difference value of eight pixels of other nodes P j and the other surrounding. 以此确保水印嵌入对地图不造成过大的扰动。 In order to ensure the embedding of the map does not cause too much disturbance.

[0106] 具体地,依次读取水印序列,对每个特征点: [0106] Specifically, the watermark sequence sequentially reads, for each feature point:

[0107] Sr (Pj) = s (Pj)+s (Pj) Lj (3) [0107] Sr (Pj) = s (Pj) + s (Pj) Lj (3)

[0108] 由此: [0108] Thus:

[0109] Vi' = Vj+s (Pj) Lj (4) [0109] Vi '= Vj + s (Pj) Lj (4)

[0110] 得到嵌入水印后的地图集合{V' Ivi' |1彡i彡4}。 Map of the [0110] obtained watermark embedding set {V 'Ivi' | 1 i San San 4}.

[0111] 本发明为了便利后续的水印提取,可从V'任意选取至多3幅地图发布。 [0111] The present invention, in order to facilitate the subsequent extraction of the watermark can be 'up to three arbitrarily selected from the published map V.

[0112] 本发明根据未发布的水印地图和待检测地图,根据阈值定义,选取部分区域像素值进行对比后,对所有平移地图进行逆向恢复,并最后求得平均像素值的地图样本。 After [0112] According to the present invention is not to be released and the watermark detection map maps, according to a defined threshold, to select a partial region of the pixel values ​​are compared for all reverse recovery pan maps, a map and finally determined average pixel sample values.

[0113] 本发明根据原始水印生成规则,重新生成伪随机序列,计算其与平均像素值的地图样本的关联度,如果随机序列存在于地图内,则关联度的绝对值必大于阈值;关联度过小,则表明伪随机序列不存在于地图内,即地图内无被检测水印。 [0113] The present invention is based on the original watermark generation rule, again generates a pseudo-random sequence, calculate the degree of association with the average pixel value map of the sample, if the random sequence is present in the map, the absolute value of the degree of association will be greater than a threshold; Correlation is too small, it indicates that the pseudo-random sequence is not present in the map, i.e., no watermark is detected inside the map.

[0114] 本发明对于求得的水印序列,通过与原始水印的相似度计算,确定水印的真实性和地图的完整性。 [0114] For the present invention, the watermark sequence obtained by calculating the similarity of the original watermark, the watermark to determine the authenticity and integrity of the map.

[0115] 本发明面向影像地图的强鲁棒数字水印方法,载体为IMG格式的数字影像地图, MG可以直接转换为TIFF、JPG、CNG等多种格式。 [0115] The present invention is a method for watermarking robustness image map, the carrier is a digital image IMG map format, the MG can be directly converted to TIFF, JPG, CNG and other formats. 开发环境为VC. 6. 0,如图4所示,主要验证数字水印的嵌入方法和提取方法。 Development environment for VC. 6. 0,, primary verified digital watermark embedding method and the extraction method shown in FIG. 4.

[0116] (1)读取影像地图文件中节点像素信息的数据结构 [0116] Data structure of the pixel information (1) to read the image file in the node map

[0117] 由于MG文件的节点的组织方式类似于二叉树(如图1所示)。 [0117] Since the nodes MG organization similar binary file (as shown in Figure 1). 读取MG文件的节点像素数据时,采用先顺序遍历,每一个节点都有自己的链表和头文件,头文件的存储结构EhfaEntry格式如下: When the pixel data read node MG file, using the first order traversal, each node has its own list and header files, storage structure EhfaEntry header format is as follows:

[0118] Longnext ;/*下一个节点的位置*/ [0118] Longnext; / * the next node position * /

[0119] Longprey ;/*前一个节点的位置*/ [0119] Longprey; / * previous node position * /

[0120] Longparent ;/* 父节点的位置*/ [0120] Longparent; / * parent node position * /

[0121] Longchild ;/*第一个子节点的位置*/ [0121] Longchild; / * position of the first child node * /

[0122] Longdata ; /*像素数据的存放位詈*/ [0122] Longdata; / * bit pixel data stored curse * /

[0123] Longdatasize :/* 数据大小*/ [0123] Longdatasize: / * Data Size * /

[0124] Char [64] name ;/* 节点的名字*/ [0124] Char [64] name; / * node name * /

[0125] Char [32] type ;/* 节点的存储结构*/ [0125] Char [32] type; / * storage structure node * /

[0126] HMEmodTime ;/*此节点的修改时间*/ [0126] HMEmodTime; / * modified this node * /

[0127] 每个节点的data字段之后都有具体的像素信息。 [0127] After the data field of each node has a specific pixel information.

[0128] (2)由用户扫描水印位图产生水印序列 [0128] (2) generated by a user to scan watermark serial watermark bitmap

[0129] 程序读取水印位图,扫面位图的像素值,得到二值序列。 [0129] the program reads the watermark bitmap, the bitmap scan plane pixel value to obtain a binary sequence. 为了提高安全性,生成对应的伪随机序列。 To improve security, generating a corresponding pseudorandom sequence.

[0130] (3)抽取每幅地图的特征点 [0130] (3) feature points are extracted in each map

[0131] 如图2所示,选择平移后需要嵌入水印的对象,将水印编码长度写入HVS系统,具体地: [0131] 2, the selection object to embed watermark posterior translation, the length of the writing HVS watermark encoder system, in particular:

[0132] a读取影像地图文件,按照伪随机序列大小,确定特征点的数目; [0132] a read image map file, in accordance with the size of the pseudorandom sequence, determining the number of feature points;

[0133] b利用HVS系统,扫描全幅地图,搜索最大数量的不敏感特征点,如果特征点有相邻情况,则比较其亮度值; [0133] b using HVS system, the whole scanning maps, search for the maximum number of feature points is not sensitive, if there is an adjacent feature points, the brightness value of the comparator;

[0134] c若可选数目仍大于水印编码长度,则按照亮度值进行排序,确定优先顺序; [0134] c optional if the number is still greater than the length of the watermark encoding, then sorted according to the luminance value, prioritization;

[0135] d保存特征点集合。 [0135] d save feature point set.

[0136] (4)将伪随机序列依次写入地图 [0136] (4) The pseudo-random sequence are sequentially written map

[0137] 具体步骤如下: [0137] the following steps:

[0138] a读取首个对象节点的像素信息。 [0138] a first pixel information read target node.

[0139] b比较该节点与周围相邻节点的亮度差值,记录该差值。 [0139] b comparing the luminance difference between the node and the neighboring node around, recording the difference.

[0140] C如图5所示流程图,将水印编码序列依次写入每个结点的像素数据内,即将亮度差值(注意可正可负)与伪随机序列对应编码的乘积累加到该节点的像素值。 [0140] the pixel data of the flowchart shown in FIG. 5 C, the watermark coding sequence are sequentially written in each node, i.e. the luminance difference value (note may be positive or negative) corresponding to the pseudo-random sequence encoded by the accumulation added the pixel value of the node.

[0141] d保存地图文件。 [0141] d save the map file.

[0142] 如图3所示,为主程序界面。 [0142] As shown in FIG. 3, the main program interface.

[0143] (5)获得新的地图样本,提取数字水印 [0143] (5) to get a new map sample, extract the digital watermark

[0144] 读取待检测地图文件,根据相同内容的地图文件,重新对地图进行平移恢复,求取检测地图的平均像素值,获得地图样本,并计算地图样本与伪随机序列的相关度,如果相关系数的绝对值大于阈值则相应水印提取成功。 [0144] reads the map file to be detected, according to the map files of the same content re-map recovery pan, obtains the average pixel value detected maps, maps the samples obtained, the sample map and calculates the correlation with the pseudo-random sequence, if the absolute value of the correlation coefficient is greater than a threshold value corresponding watermark extraction success.

[0145] (6)根据读取的二值水印序列和水印位图的大小,生成水印位图。 [0145] (6) The binary watermark sequence and size of the watermark bitmap is read, generating a watermark bitmap.

[0146] 根据提取出的水印位图的字节数,位图字节数=位图宽度*位图高度,以及读取的二值水印序列可以生成原始水印。 [0146] The number of bytes in the extracted watermark bitmap, the bitmap = number of bytes Bitmap Bitmap Width * Height, and reading the binary watermark sequence may be generated the original watermark. 并对原始水印位图和检测到的水印位图进行NC值比较来确定水印的真实性和完整程度。 And the original watermark bitmap and detected watermark bitmaps NC value compared to determine the authenticity and completeness of the watermark.

Claims (2)

  1. 1. 一种面向影像地图的强鲁棒数字水印方法,包括水印的嵌入和提取,其特征在于,包括如下步骤: (1) 根据MG影像地图的数据结构,读取节点的链表,并按照对应字段分别获得每个节点的像素值; (2) 扫描水印位图,获得水印二进制编码,生成相应的伪随机序列L, CLAIMS 1. A method for watermarking robustness of the image map, including embedding and extracting a watermark, characterized by comprising the steps of: (1) The data structure of the image map MG, read the list of nodes, and in a corresponding field pixel value of each node, respectively; and (2) scanning a bitmap watermark, the watermark is obtained binary encoding, generates a corresponding pseudo-random sequence L,
    Figure CN104915918AC00021
    其中,LaG{+1,-1},L是一个由N个元素组成的伪随机序列,N是水印编码的长度; (3) 根据地图规模和伪随机序列大小,设定小于1个像素单位的值作为阈值,对地图进行八个方向的平移,包括上T、右上RT、右R、右下RD、下D、左下LD、左L、左上LT; (4) 对于位移后的八幅地图,按照顺时针顺序,随机选取其中3幅地图,联同原始地图共同组成待嵌入水印的地图集合{V|Vi|l彡i彡4}; (5) 利用人类视觉系统,检测每一幅地图,据地图集合{V|Vi 11 <i< 4},按照每个节点仅嵌入1比特水印的编码规则,选取最大规模的特征点集合,同时提取对应节点的像素值; (6) 计算每一个特征点与周围像素点的差值,并读取伪随机序列,将对应随机序列值与像素差值的乘积累加到该特征点的像素值; (7) 对于嵌入水印后的地图,随机选取其中几幅发布; (8) 在水印检测时,将待检测地图与 Wherein, LaG {+ 1, -1}, L is a pseudo-random sequence of N elements, N is the length of the watermark encoder; (3) the scale of the map and the size of the pseudo-random sequence, is set smaller than a pixel unit as the threshold value, the map translational eight directions, including T, RT right upper, and right R, bottom right RD, at D, the LD bottom left, left L, left upper LT; (. 4) for the displacement of the eight map , is clockwise, wherein three randomly selected map, the map associated with the original watermark to be embedded together form a map set {V | Vi | l i San San 4}; (5) using the human visual system, each detecting a map , according to the map set {V | Vi 11 <i <4}, each node only embedded watermark bit coding rule 1, selecting the largest set of feature points, and extracts the corresponding pixel values ​​of the nodes; (6) is calculated for each the difference between the feature point and pixels around the point, and reads the pseudo-random sequence, the corresponding pixel value by accumulating the difference between the pixel value of the random sequence is added to the feature point; (7) to the map watermarked, selected randomly pieces of release; (8) in the watermark detection, will be detected and map 它嵌入水印但未发布的地图相互融合,求得具有平均像素值的地图; (9) 根据原始的水印的伪随机序列,以及融合后的水印地图,计算关联系数,如果相关系数的绝对值大于等于所设定的阈值,则表明检测到水印编码,否则未检测到; (10) 将编码还原为水印位图,并计算相似度。 But watermark embedding it merging published map, the map is obtained having an average pixel value; (9) The pseudo-random sequence of the original watermark, the watermark and fused map, calculate correlation coefficient, if the absolute value of the correlation coefficient is greater than is equal to the set threshold value, it indicates that the watermark is detected, otherwise it is not detected; (10) is reduced to the encoded watermark bitmap, and calculates the similarity.
  2. 2. 根据权利要求1所述的一种面向影像地图的强鲁棒数字水印方法,其特征在于,所述的待嵌入水印的地图集合的生成步骤包括: 对于每幅地图Vi,计算对应的水印嵌入强度: S(v^ ={s(pj) >0 11 ^j^N} 其中,s(Pj)表示地图中特征点口」在HVS系统内的不可察觉程度,即p」与其他周围其他八个像素节点的亮度差值; 依次读取水印序列,对每个特征点: s'Pj=s(pj)+s(Pj)Lj, v/ =vi+s(pJ)LJ, 得到IV,|Vi' |1 彡i彡4}。 According to one of the claim 1 for the robustness of the watermarking method of map image, wherein said step of generating a map set to be watermarked comprising: Vi for each map, calculates the corresponding watermarking embedding strength: S (v ^ = {s (pj)> 0 11 ^ j ^ N} where, s (Pj) denotes the map is not the degree of perceived feature points port "in the HVS system, i.e., p" and the other around the other the luminance difference value of pixels eight nodes; sequentially read the watermark sequence, for each feature point: s'Pj = s (pj) + s (Pj) Lj, v / = vi + s (pJ) LJ, to give IV, | Vi '| 1 i San San 4}.
CN 201510349570 2015-06-23 2015-06-23 Strong robust digital watermarking method oriented image map CN104915918B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201510349570 CN104915918B (en) 2015-06-23 2015-06-23 Strong robust digital watermarking method oriented image map

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201510349570 CN104915918B (en) 2015-06-23 2015-06-23 Strong robust digital watermarking method oriented image map

Publications (2)

Publication Number Publication Date
CN104915918A true true CN104915918A (en) 2015-09-16
CN104915918B CN104915918B (en) 2017-12-19

Family

ID=54084960

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201510349570 CN104915918B (en) 2015-06-23 2015-06-23 Strong robust digital watermarking method oriented image map

Country Status (1)

Country Link
CN (1) CN104915918B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20020090110A1 (en) * 1996-10-28 2002-07-11 Braudaway Gordon Wesley Protecting images with an image watermark
US20080022114A1 (en) * 1996-07-02 2008-01-24 Wistaria Trading, Inc. Optimization methods for the insertion, protection, and detection of digital watermarks in digital data
CN101149835A (en) * 2007-10-29 2008-03-26 中国人民解放军信息工程大学 Map data rasterizing based robust blind water mark embedding and extraction method
CN101458810A (en) * 2008-12-08 2009-06-17 哈尔滨工程大学 Vector map watermark method based on object property characteristic

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080022114A1 (en) * 1996-07-02 2008-01-24 Wistaria Trading, Inc. Optimization methods for the insertion, protection, and detection of digital watermarks in digital data
US20020090110A1 (en) * 1996-10-28 2002-07-11 Braudaway Gordon Wesley Protecting images with an image watermark
CN101149835A (en) * 2007-10-29 2008-03-26 中国人民解放军信息工程大学 Map data rasterizing based robust blind water mark embedding and extraction method
CN101458810A (en) * 2008-12-08 2009-06-17 哈尔滨工程大学 Vector map watermark method based on object property characteristic

Also Published As

Publication number Publication date Type
CN104915918B (en) 2017-12-19 grant

Similar Documents

Publication Publication Date Title
Li et al. A steganographic method based upon JPEG and particle swarm optimization algorithm
Fridrich et al. Steganalysis based on JPEG compatibility
Kutter et al. Watermark copy attack
US20030188166A1 (en) Data processing apparatus and method
Fridrich Visual hash for oblivious watermarking
Alattar Reversible watermark using difference expansion of quads
Lin et al. Semifragile watermarking for authenticating JPEG visual content
Khan et al. A recent survey of reversible watermarking techniques
Caldelli et al. Reversible watermarking techniques: An overview and a classification
Chang et al. Four-scanning attack on hierarchical digital watermarking method for image tamper detection and recovery
Fridrich et al. Robust digital watermarking based on key-dependent basis functions
Niu et al. A survey of digital vector map watermarking
Farzam et al. A robust multimedia watermarking technique using Zernike transform
Li et al. Geometrically invariant image watermarking using Polar Harmonic Transforms
US20020191812A1 (en) Object edge watermarking
Phadikar et al. Region based QIM digital watermarking scheme for image database in DCT domain
Bravo-Solorio et al. Passive forensic method for detecting duplicated regions affected by reflection, rotation and scaling
EP0962886A2 (en) Method and system for watermarking in the frequency domain
CN102156952A (en) Copyright-protection-service-oriented remote sensing image digital watermarking method
Song et al. Recent advances and classification of watermarking techniques in digital images
Ni et al. Secure semi-blind watermarking based on iteration mapping and image features
Guo et al. Lossless watermarking for verifying the integrity of medical images with tamper localization
Ramkumar et al. Image watermarks and counterfeit attacks: Some problems and solutions
US20030172275A1 (en) Real-time blind watermarking method using quantization
Kuo et al. A reversible data hiding scheme based on block division

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
GR01