CN105654496B - The bionical adaptive fuzzy edge detection method of view-based access control model characteristic - Google Patents
The bionical adaptive fuzzy edge detection method of view-based access control model characteristic Download PDFInfo
- Publication number
- CN105654496B CN105654496B CN201610010368.5A CN201610010368A CN105654496B CN 105654496 B CN105654496 B CN 105654496B CN 201610010368 A CN201610010368 A CN 201610010368A CN 105654496 B CN105654496 B CN 105654496B
- Authority
- CN
- China
- Prior art keywords
- image
- fuzzy
- brightness
- edge detection
- adaptive
- Prior art date
- Legal status (The legal status 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 status listed.)
- Expired - Fee Related
Links
- 238000003708 edge detection Methods 0.000 title claims abstract description 35
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000003044 adaptive effect Effects 0.000 title claims abstract description 30
- 241000282414 Homo sapiens Species 0.000 claims abstract description 42
- 230000006870 function Effects 0.000 claims abstract description 26
- 230000000007 visual effect Effects 0.000 claims abstract description 21
- 230000009466 transformation Effects 0.000 claims abstract description 14
- 230000016776 visual perception Effects 0.000 claims abstract description 6
- 238000000605 extraction Methods 0.000 claims description 29
- 230000000694 effects Effects 0.000 claims description 21
- 238000004364 calculation method Methods 0.000 claims description 16
- 230000004438 eyesight Effects 0.000 claims description 7
- 210000001116 retinal neuron Anatomy 0.000 claims description 7
- 230000002964 excitative effect Effects 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 claims description 5
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 230000005764 inhibitory process Effects 0.000 claims description 4
- 230000023886 lateral inhibition Effects 0.000 claims description 3
- 230000002146 bilateral effect Effects 0.000 claims description 2
- 238000013507 mapping Methods 0.000 claims description 2
- 230000001629 suppression Effects 0.000 claims description 2
- 230000002708 enhancing effect Effects 0.000 claims 7
- 230000000747 cardiac effect Effects 0.000 claims 2
- 230000035807 sensation Effects 0.000 claims 2
- 238000012935 Averaging Methods 0.000 claims 1
- 230000001419 dependent effect Effects 0.000 claims 1
- 238000012545 processing Methods 0.000 abstract description 22
- 239000011664 nicotinic acid Substances 0.000 abstract description 5
- 239000000284 extract Substances 0.000 abstract description 4
- 241000209094 Oryza Species 0.000 description 6
- 235000007164 Oryza sativa Nutrition 0.000 description 6
- 238000010586 diagram Methods 0.000 description 6
- 230000002093 peripheral effect Effects 0.000 description 6
- 235000009566 rice Nutrition 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 230000004044 response Effects 0.000 description 4
- 230000001953 sensory effect Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 206010013142 Disinhibition Diseases 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000002401 inhibitory effect Effects 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000002547 anomalous effect Effects 0.000 description 1
- 230000008827 biological function Effects 0.000 description 1
- 238000005282 brightening Methods 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 235000001968 nicotinic acid Nutrition 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000000638 stimulation Effects 0.000 description 1
- 230000007704 transition Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20004—Adaptive image processing
- G06T2207/20008—Globally adaptive
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20004—Adaptive image processing
- G06T2207/20012—Locally adaptive
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
本发明涉及了一种基于人类视觉特性的图像仿生自适应模糊边缘检测方法，属于数字图像处理技术领域。本发明采用如下步骤实现：对操作图像进行全局亮度自适应增强，将原图像空间域转化为模糊域，对操作图像模糊域进行局部亮度自适应增强，对原图像模糊域进行逆变换，最后采用“Nakagowa”算子对处理后的图像边缘提取。本发明依据人眼视觉感知系统的全局和局部自适应调节特性对传统Pal算法模糊边缘检测进行优化保留了图像的低灰度值边界信息，简化了隶属度函数，具有更强的抗噪能力，能够有效地把边缘从复杂的背景中提取出来，针对不同种类的图像能够自适应地进行边缘检测。
The invention relates to an image bionic adaptive fuzzy edge detection method based on human visual characteristics, and belongs to the technical field of digital image processing. The present invention adopts the following steps to realize: carry out global brightness adaptive enhancement on the operation image, convert the original image space domain into a fuzzy domain, perform local brightness adaptive enhancement on the operation image fuzzy domain, perform inverse transformation on the original image fuzzy domain, and finally adopt The "Nakagowa" operator extracts the edge of the processed image. According to the global and local adaptive adjustment characteristics of the human visual perception system, the present invention optimizes the fuzzy edge detection of the traditional Pal algorithm, retains the low gray value boundary information of the image, simplifies the membership function, and has stronger anti-noise ability. It can effectively extract edges from complex backgrounds, and can adaptively perform edge detection for different types of images.
Description
技术领域：Technical field:
本发明涉及了一种基于人类视觉特性的图像仿生自适应模糊边缘检测方法，属于数字图像处理技术领域。The invention relates to an image bionic adaptive fuzzy edge detection method based on human visual characteristics, and belongs to the technical field of digital image processing.
技术背景：technical background:
仿生学是人类模仿生物功能来发明创造的科学，是利用生物的结构和功能原理来研制机械或各种新技术的一门学科。人类视觉系统具有亮度自适应特性和视网膜神经元感受野非经典侧抑制特性，越来越多的研究人员将仿人类视觉系统思想运用到图像处理、系统辨识和机器视觉方面。Bionics is a science in which human beings imitate biological functions to invent and create. It is a discipline that utilizes biological structure and functional principles to develop machinery or various new technologies. The human visual system has the characteristics of brightness adaptation and non-classical lateral inhibition of the receptive field of retinal neurons. More and more researchers have applied the idea of imitating the human visual system to image processing, system identification and machine vision.
边缘检测是图像处理中最重要的研究内容之一，是进行图像描述、分析和理解的关键技术。1959年Julez首次提出边缘检测概念，之后国内外大量专家学者对边缘检测这一课题进行了广泛深入的研究并提出了很多方法，常见的边缘检测方法有差分法、Robert、Sobel、Prewitt、Log、Canny等方法，但这些方法主要增强了图像的边缘等高频信息，有效提取了图像边缘，但对噪声较为敏感，在处理实际图像中效果并不理想。Pal和King把模糊的概念引入到图像处理的领域中，提出基于模糊集合理论的图像分割方法，该算法能够起到较好的抑躁效果，能够取得较好的图像分割效果，但是经过理论和实验分析：Pal和King算法前期没有进行去噪处理，对于含噪图像效果并不理想，容易造成误判；幂级数形式的隶属度函数仅仅达到了图像映射到模糊集的作用，几乎没有一点模糊增强的效果；图像经过非线性变换后图像中低灰度值会变成0，造成边缘信息的丢失，影响最终效果；迭代次数、分割阀值的选择对图像的模糊增强处理有很大影响，对模糊域图像逆变换存在影响。Edge detection is one of the most important research contents in image processing, and it is a key technology for image description, analysis and understanding. In 1959, Julez first proposed the concept of edge detection. After that, a large number of experts and scholars at home and abroad conducted extensive and in-depth research on the subject of edge detection and proposed many methods. The common edge detection methods include difference method, Robert, Sobel, Prewitt, Log, Canny and other methods, but these methods mainly enhance the high-frequency information such as the edge of the image, and effectively extract the edge of the image, but are sensitive to noise, and the effect is not ideal in processing actual images. Pal and King introduced the concept of fuzziness into the field of image processing, and proposed an image segmentation method based on fuzzy set theory. Experimental analysis: The Pal and King algorithms did not perform denoising processing in the early stage, and the effect on noisy images is not ideal, which may easily cause misjudgment; the membership function in the form of power series only achieves the function of image mapping to fuzzy sets, almost no The effect of blur enhancement; the low gray value in the image will become 0 after the image undergoes nonlinear transformation, resulting in the loss of edge information and affecting the final effect; the selection of the number of iterations and segmentation threshold has a great influence on the blur enhancement processing of the image , which has an impact on the inverse transformation of the fuzzy domain image.
基于以上背景，本发明以传统Pal和King模糊边缘检测算法为框架，以人类视觉系统的全局增强特性和局部自适应调节特性为理论依据，提出了一种基于视觉特性的仿生自适应模糊边缘检测方法，并将该方法应用在实际图像边缘提取中，对图像进行亮度全局自适应和局部增强处理，既能很好地增强图像的边缘对比度，又能提高被传统感受野所滤除的区域亮度对比和亮度梯度信息，保留了低灰度边缘信息，增强后的图像质量符合人眼的主观视觉效果，相关的专利如申请公布号CN 103310461 A的发明专利公开了一种基于块Kalman滤波的图像边缘提取方法，分别采用分层处理和插值运算有效提高了滤波效果和操作对象的信噪比，提高了图像边缘检测的质量。申请公布号CN 104809733 A的发明专利公开了一种古建墙壁受污题记文字图像边缘提取方法，采用gabor滤波器对空间平均，识别和消除污染造成的伪边缘和不连续边缘。申请公布号CN 101286233 A的发明专利公开了一种基于对象云的模糊边缘检测方法，利用最大熵原理自适应进行边缘过渡区处理，弥补基于模糊集理论算法的缺陷。申请公布号CN 104268872 A的发明专利公开了一种基于一致性的边缘检测方法，引入灰度梯度方向，区分真实边缘、噪声引起的灰度梯度变化。但是，以上专利并没有涉及基于人类视觉特性的图像边缘提取。Based on the above background, the present invention takes the traditional Pal and King fuzzy edge detection algorithm as the framework, and takes the global enhancement characteristics and local adaptive adjustment characteristics of the human visual system as the theoretical basis, and proposes a bionic adaptive fuzzy edge detection based on visual characteristics method, and apply this method in the actual image edge extraction, and perform brightness global adaptation and local enhancement processing on the image, which can not only enhance the edge contrast of the image well, but also improve the brightness of the area filtered out by the traditional receptive field Contrast and brightness gradient information, low gray edge information is retained, and the enhanced image quality conforms to the subjective visual effect of the human eye. Related patents such as the invention patent application publication number CN 103310461 A disclose an image based on block Kalman filtering The edge extraction method adopts layered processing and interpolation operation to effectively improve the filtering effect and the signal-to-noise ratio of the operating object, and improves the quality of image edge detection. The invention patent of the application publication number CN 104809733 A discloses a method for extracting the edge of the image of the inscription on the ancient building wall, which uses the gabor filter to average the space to identify and eliminate the false edge and the discontinuous edge caused by the pollution. The invention patent of application publication number CN 101286233 A discloses a fuzzy edge detection method based on object cloud, which uses the maximum entropy principle to adaptively process the edge transition area, and makes up for the defects of the algorithm based on fuzzy set theory. The invention patent of application publication number CN 104268872 A discloses a consistency-based edge detection method, which introduces the gray gradient direction to distinguish real edges and gray gradient changes caused by noise. However, the above patents do not relate to image edge extraction based on human visual characteristics.
发明内容：Invention content:
针对图像对比度较小，边缘检测效果不佳问题，本发明以传统Pal和King模糊边缘检测算法为框架，以人类视觉系统的全局增强特性和局部自适应调节特性为理论依据，提出了一种基于视觉特性的仿生自适应模糊边缘检测方法，依据人眼视觉感知系统的全局和局部自适应调节特性对传统Pal和King算法模糊边缘检测进行优化。Aiming at the problem that the image contrast is small and the edge detection effect is not good, the present invention takes the traditional Pal and King fuzzy edge detection algorithm as the framework, and takes the global enhancement characteristics and local self-adaptive adjustment characteristics of the human visual system as the theoretical basis, and proposes a method based on The bionic adaptive fuzzy edge detection method of visual characteristics optimizes the traditional Pal and King algorithm fuzzy edge detection according to the global and local adaptive adjustment characteristics of the human visual perception system.
本优化算法在图像边缘检测时保留了重要的边缘信息，检测出与人的主观视觉更为一致的图像边缘，提高了模糊边缘检测的自适应性和实用性。This optimization algorithm retains important edge information during image edge detection, detects image edges that are more consistent with human subjective vision, and improves the adaptability and practicability of fuzzy edge detection.
本发明采用如下技术方案：The present invention adopts following technical scheme:
一种基于视觉特性的仿生自适应模糊边缘检测方法，其步骤是：A bionic adaptive fuzzy edge detection method based on visual characteristics, the steps are:
步骤1，对操作图像进行全局亮度自适应增强，采用基于人类视觉系统特性的全局自适应性对数模型，对原始图像的明暗程度整体亮度进行非线性调整，对图像过暗或过亮部分进行调节，使图像的明暗区域对比度增强；Step 1: Carry out adaptive global brightness enhancement on the operating image, and use a global adaptive logarithmic model based on the characteristics of the human visual system to non-linearly adjust the overall brightness of the original image's lightness and darkness, and perform a non-linear adjustment on the dark or bright parts of the image. Adjust to enhance the contrast of the bright and dark areas of the image;
步骤2，将原图像空间域转化为模糊域，定义一个简单有效的隶属度函数取代原有的隶属度函数，提高模糊边缘提取算法的实时性；利用正弦函数的性质和定义，有效的实现模糊域的转化，避免了图像中大部分的低灰度值被硬性设置为0，保存了图像的低灰度边缘信息；Step 2, transform the original image space domain into a fuzzy domain, define a simple and effective membership function to replace the original membership function, and improve the real-time performance of the fuzzy edge extraction algorithm; use the properties and definitions of the sine function to effectively realize the fuzzy Domain conversion avoids most of the low grayscale values in the image being rigidly set to 0, and preserves the low grayscale edge information of the image;
步骤3，对操作图像模糊域进行局部亮度自适应增强，采用视网膜神经元感受野非经典侧抑制性的三高斯模型和高斯滤波相结合的双边滤波计算领域主观感觉亮度，依据当前点亮度和主观感觉亮度间的线性关系增强图像局部细节信息；Step 3: Carry out local brightness adaptive enhancement on the fuzzy domain of the operation image, and use the triple-Gaussian model of the retinal neuron receptive field non-classical lateral inhibition and the bilateral filter combined with Gaussian filtering to calculate the subjective perception brightness of the field, based on the current point brightness and subjective The linear relationship between sensory brightness enhances the local detail information of the image;
步骤4，对原图像模糊域进行逆变换，逆变函数将模糊隶属度矩阵转化为空间域图像；Step 4, perform inverse transformation on the fuzzy domain of the original image, and the inverse function converts the fuzzy membership degree matrix into a spatial domain image;
步骤5，图像边缘提取，采用“Nakagowa”算子对处理后的图像边缘提取。Step 5, image edge extraction, using the "Nakagowa" operator to extract the edge of the processed image.
与现图像边缘检测技术相比，本发明的优点在于：依据人眼视觉感知系统的全局和局部自适应调节特性对传统Pal算法模糊边缘检测进行优化保留了图像的低灰度值边界信息，简化了隶属度函数，具有更强的抗噪能力，能够有效地把边缘从复杂的背景中提取出来，针对不同种类的图像能够自适应地进行边缘检测。Compared with the existing image edge detection technology, the present invention has the advantages of: according to the global and local adaptive adjustment characteristics of the human visual perception system, the traditional Pal algorithm fuzzy edge detection is optimized, and the low gray value boundary information of the image is preserved, simplifying It has a stronger anti-noise ability, can effectively extract edges from complex backgrounds, and can adaptively perform edge detection for different types of images.
本发明的优选方案是：Preferred version of the present invention is:
依据人眼视觉感知系统的全局增强特性，所述步骤1图像全局亮度增强计算公式为：According to the global enhancement characteristics of the human visual perception system, the formula for calculating the global brightness enhancement of the image in step 1 is:
式中：为原始图像位置处的像素值，是经过全局亮度增强后的归一化亮度，、、是根据图像自身的亮度确认全局对数调整的程度；人眼视觉系统根据目标的整体亮度情况，初期自适应地全局增强图像的亮度，通过参数化对数模型自适应地全局增强图像亮度，该非线性调节有效地压缩了图像的动态范围，使图像的暗区域变亮。In the formula: for the original image the pixel value at the location, is the normalized brightness after global brightness enhancement, , , It is to confirm the degree of global logarithmic adjustment according to the brightness of the image itself; the human visual system initially adaptively enhances the brightness of the image globally according to the overall brightness of the target, and adaptively enhances the brightness of the image globally through a parameterized logarithmic model. Non-linear scaling effectively compresses the dynamic range of the image, brightening dark areas of the image.
步骤2中，确定图像隶属度函数：定义一个简单有效的隶属度函数，Pal和King算法在得到图像的模糊集合中元素表示像素的隶属度；计算公式为：In step 2, determine the membership function of the image: define a simple and effective membership function, and the Pal and King algorithm obtains elements in the fuzzy set of the image represent pixels The degree of membership; the calculation formula is:
式中，表示图像像素的灰度值，、分别表示图像中最大、最小灰度级；由正弦函数的性质及定义，知其隶属值域线性度好于Pal和King算法的隶属度函数，避免了图像中的低灰度值被硬性设置为0，保存了图像的低灰度边缘信息。In the formula, Represents image pixels the gray value of , Respectively represent the maximum and minimum gray levels in the image; from the nature and definition of the sine function, it is known that the linearity of its membership range is better than the membership function of the Pal and King algorithms, which avoids the low gray value in the image being rigidly set to 0, the low-gray edge information of the image is saved.
根据人类视觉局部自适应调节特性，所述步骤3图像局部亮度增强计算公式为：According to the local adaptive adjustment characteristics of human vision, the calculation formula for image local brightness enhancement in step 3 is:
式中：为一正值常数，为局部线性关系的比例系数；是模糊变换后的图像值，是当前点处的邻域平均亮度，它反映当前点所在位置人眼感受到的亮度情况；为权重系数取决于定义域核和值域核的乘积，输出像素的值依赖于领域像素值的加权组合；由人类视觉系统特性得知人眼对于局部对比度更为敏感，视觉系统在对信号进行最终处理时，具有侧抑制效应的效果，会使人眼对图像的边缘有增强的感觉。In the formula: is a positive constant, and is the proportional coefficient of the local linear relationship; is the image value after fuzzy transformation, is the current point The average brightness of the neighborhood at , which reflects the brightness perceived by the human eye at the current point; is the weight coefficient depends on the domain kernel and range kernel The value of the output pixel depends on the weighted combination of domain pixel values; from the characteristics of the human visual system, it is known that the human eye is more sensitive to local contrast. The human eye has an enhanced perception of the edges of an image.
步骤4中，对模糊集进行逆变换：对局部增强图像进行逆变换，将模糊隶属度矩阵转化为空间域图像；计算公式为：In step 4, perform inverse transformation on the fuzzy set: perform inverse transformation on the locally enhanced image, and transform the fuzzy membership degree matrix into a spatial domain image; the calculation formula is:
依据人类视觉局部自适应调节特性，所述步骤3视网膜神经元感受野的三高斯模型计算公式是：According to the local adaptive adjustment characteristics of human vision, the calculation formula of the three-Gaussian model of the retinal neuron receptive field in step 3 is:
式中，表示感受野内任意一点的兴奋反应大小，分别表示中心兴奋区、四周抑制区、外周大范围去抑制区的敏感度峰值，分别表示中心兴奋区、抑制区、外周大范围去抑制区的面积系数。In the formula, Indicates the magnitude of the excitement response at any point in the sensory field, Respectively represent the sensitivity peaks of the central excitatory area, the peripheral inhibition area, and the peripheral large-scale disinhibition area, Represent the area coefficients of the central excitatory zone, inhibitory zone, and peripheral large-scale disinhibited zone, respectively.
步骤5，图像边缘提取：利用Nakagowa 提出的最小算子，定义图像的边缘，计算公式为：Step 5, image edge extraction: use the minimum operator proposed by Nakagowa to define the edge of the image, the calculation formula is:
式中：表示处理后的图像，，表示3×3的窗口。In the formula: represents the processed image, , Represents a 3×3 window.
附图说明Description of drawings
图1为本发明所涉及的方法流程图。Fig. 1 is a flow chart of the method involved in the present invention.
图2为人眼主观亮度与光强度的对数关系。Figure 2 shows the logarithmic relationship between the subjective brightness of the human eye and the light intensity.
图3为感受野的三高斯模型。Figure 3 is a three-Gaussian model of the receptive field.
图4为“rice”图像。Figure 4 is the "rice" image.
图5为利用Canny算子对“rice”图像边缘提取效果图。Figure 5 is the effect diagram of edge extraction of "rice" image using Canny operator.
图6为利用Pal和King算法（n=2）对“rice”图像边缘提取效果图。Figure 6 is the effect diagram of edge extraction of "rice" image using Pal and King algorithm (n=2).
图7为利用本发明算法对“rice”图像边缘提取效果图。Fig. 7 is an effect diagram of edge extraction of a "rice" image using the algorithm of the present invention.
图8为“cameraman”图像。Figure 8 is the "cameraman" image.
图9为利用Canny算子对“cameraman”图像边缘提取效果图。Figure 9 is an effect diagram of edge extraction of the "cameraman" image using the Canny operator.
图10 为利用Pal和King算法（n=2）对“cameraman”图像边缘提取效果图。Figure 10 is the effect diagram of edge extraction of "cameraman" image using Pal and King algorithm (n=2).
图11为利用本发明算法对“cameraman”图像边缘提取效果图。Fig. 11 is an effect diagram of edge extraction of "cameraman" image by using the algorithm of the present invention.
具体实施方式：Detailed ways:
下面结合附图和具体实施方式对本发明作进一步说明。The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.
本发明按照图1的步骤流程来进行：The present invention carries out according to the step process of Fig. 1:
步骤1：step 1:
对操作图像进行全局亮度自适应增强，采用基于人类视觉系统特性的全局自适应性对数模型，对原始图像的明暗程度整体亮度进行非线性调整，对图像过暗或过亮部分进行调节，使图像的明暗区域对比度增强，以便于人眼观察。The global brightness adaptive enhancement is performed on the operating image, and the global adaptive logarithmic model based on the characteristics of the human visual system is used to non-linearly adjust the overall brightness of the original image, and to adjust the dark or bright parts of the image, so that The contrast between light and dark areas of the image is enhanced for easier viewing by the human eye.
依据人眼视觉感知系统的全局增强特性，图像全局亮度增强计算公式为：According to the global enhancement characteristics of the human visual perception system, the calculation formula for image global brightness enhancement is:
式中：为原始图像位置处的像素值，是经过全局亮度增强后的归一化亮度，、、是根据图像自身的亮度确认全局对数调整的程度；人眼视觉系统根据目标的整体亮度情况，初期自适应地全局增强图像的亮度，以避免图像增强过亮而后在后续步骤中失去有用的亮度信息，通过参数化对数模型自适应地全局增强图像亮度，则对数函数的特性可知，该非线性调节有效地压缩了图像的动态范围，使图像的暗区域变亮。In the formula: for the original image the pixel value at the location, is the normalized brightness after global brightness enhancement, , , It is to confirm the degree of global logarithmic adjustment according to the brightness of the image itself; the human visual system initially adaptively enhances the brightness of the image globally according to the overall brightness of the target, so as to avoid the image from being too bright and losing useful brightness in subsequent steps Information, through the parameterized logarithmic model to adaptively enhance the brightness of the image globally, the characteristics of the logarithmic function can be seen that the nonlinear adjustment effectively compresses the dynamic range of the image and brightens the dark area of the image.
步骤2：Step 2:
将原图像空间域转化为模糊域，定义一个简单有效的隶属度函数取代原有的隶属度函数，提高模糊边缘提取算法的实时性；利用正弦函数的性质和定义，有效的实现模糊域的转化，避免了图像中大部分的低灰度值被硬性设置为0，保存了图像的低灰度边缘信息。Transform the original image space domain into fuzzy domain, define a simple and effective membership function to replace the original membership function, improve the real-time performance of fuzzy edge extraction algorithm; use the properties and definitions of sine function to effectively realize the conversion of fuzzy domain , which avoids most of the low grayscale values in the image being rigidly set to 0, and preserves the low grayscale edge information of the image.
确定图像隶属度函数：定义一个简单有效的隶属度函数，Pal和King算法在得到图像的模糊集合中元素表示像素的隶属度；计算公式为：Determine the image membership function: define a simple and effective membership function, Pal and King algorithm in the fuzzy set of the obtained image elements represent pixels The degree of membership; the calculation formula is:
式中，表示图像像素的灰度值，、分别表示图像中最大、最小灰度级；由正弦函数的性质及定义，知其隶属值域线性度好于Pal和King算法的隶属度函数，避免了图像中的低灰度值被硬性设置为0，保存了图像的低灰度边缘信息。In the formula, Represents image pixels the gray value of , Respectively represent the maximum and minimum gray levels in the image; from the nature and definition of the sine function, it is known that the linearity of its membership range is better than the membership function of the Pal and King algorithms, which avoids the low gray value in the image being rigidly set to 0, the low-gray edge information of the image is saved.
步骤3：Step 3:
对操作图像模糊域进行局部亮度自适应增强，采用视网膜神经元感受野非经典侧抑制性的三高斯模型和高斯滤波相结合的双边滤波计算领域主观感觉亮度，依据当前点亮度和主观感觉亮度间的线性关系增强图像局部细节信息。Local brightness adaptive enhancement is performed on the fuzzy domain of the operation image, and the subjective sensory brightness in the field is calculated by using the triple-Gaussian model of the retinal neuron receptive field non-classical side-inhibitory and the Gaussian filter. The linear relationship enhances the local detail information of the image.
根据人类视觉局部自适应调节特性，图像局部亮度增强计算公式为：According to the local adaptive adjustment characteristics of human vision, the calculation formula of image local brightness enhancement is:
式中：为一正值常数，为局部线性关系的比例系数；是模糊变换后的图像值，是当前点处的邻域平均亮度，它反映当前点所在位置人眼感受到的亮度情况；为权重系数取决于定义域核和值域核的乘积，输出像素的值依赖于领域像素值的加权组合；由人类视觉系统特性得知人眼对于局部对比度更为敏感，视觉系统在对信号进行最终处理时，类似于进行一种通过权重进行求和的运算过程，与信号处理中的带通滤波处理比较相似，具有侧抑制效应的效果，会使人眼对图像的边缘有增强的感觉。In the formula: is a positive constant, and is the proportional coefficient of the local linear relationship; is the image value after fuzzy transformation, is the current point The average brightness of the neighborhood at , which reflects the brightness perceived by the human eye at the current point; is the weight coefficient depends on the domain kernel and range kernel The value of the output pixel depends on the weighted combination of domain pixel values; from the characteristics of the human visual system, it is known that the human eye is more sensitive to local contrast. The calculation process of the sum is similar to the band-pass filter processing in signal processing, and has the effect of side suppression effect, which will make the human eye have an enhanced feeling for the edge of the image.
视网膜神经元感受野的三高斯模型计算公式是：The calculation formula of the three-Gaussian model of the receptive field of retinal neurons is:
医学研究表明，在传统的细胞感受野同心圆之处还存在着一个大范围的区域，对该区域刺激会对感受野中心的响应起到调制作用。式中，表示感受野内任意一点的兴奋反应大小，分别表示中心兴奋区、四周抑制区、外周大范围去抑制区的敏感度峰值，分别表示中心兴奋区、抑制区、外周大范围去抑制区的面积系数。Medical research has shown that there is a large-scale area at the concentric circle of the traditional cell receptive field, and stimulation of this area will modulate the response of the receptive field center. In the formula, Indicates the magnitude of the excitement response at any point in the sensory field, Respectively represent the sensitivity peaks of the central excitatory area, the peripheral inhibition area, and the peripheral large-scale disinhibition area, Represent the area coefficients of the central excitatory zone, inhibitory zone, and peripheral large-scale disinhibited zone, respectively.
步骤4：Step 4:
对原图像模糊域进行逆变换，逆变函数将模糊隶属度矩阵转化为空间域图像；对模糊集进行逆变换：对局部增强图像进行逆变换，将模糊隶属度矩阵转化为空间域图像；计算公式为：Perform inverse transformation on the fuzzy domain of the original image, and the inversion function converts the fuzzy membership matrix into a spatial domain image; perform inverse transformation on the fuzzy set: perform inverse transformation on the locally enhanced image, and convert the fuzzy membership matrix into a spatial domain image; calculate The formula is:
步骤5：Step 5:
图像边缘提取，采用“Nakagowa”算子对处理后的图像边缘提取。定义图像的边缘，计算公式为：Image edge extraction, using the "Nakagowa" operator to extract the edge of the processed image. Define the edge of the image, the calculation formula is:
式中：表示处理后的图像，，表示3×3的窗口。In the formula: represents the processed image, , Represents a 3×3 window.
表1是本发明设计方法与Pal和King方法边缘提取处理时间对比：Table 1 is the design method of the present invention and Pal and King method edge extraction processing time comparison:
表1 模糊算法边缘提取处理时间Table 1 Processing time of fuzzy algorithm edge extraction
下面给出应用本发明进行图像边缘提取的实例。图像边缘检测的基本要求是：正确检测出边缘、准确定位出边缘、边缘连续且单边响应。但是这些要求并没有准确权威的评价方法，现在最为广泛的评价方法就是主观判断，但评价结果易受评价者的经验、图像类型的影响，不能有力的说明边缘检测效果，本发明采用无需参考基准边缘图像评价方法进行评价。An example of applying the present invention to image edge extraction is given below. The basic requirements of image edge detection are: correctly detect the edge, accurately locate the edge, continuous edge and unilateral response. However, there is no accurate and authoritative evaluation method for these requirements. The most widely used evaluation method is subjective judgment, but the evaluation result is easily affected by the evaluator's experience and image type, and cannot effectively explain the edge detection effect. The edge image evaluation method is evaluated.
实施例1：Example 1:
本发明算法与经典Canny算子、Pal和King算法对“rice”边缘提取效果对比实验，Canny算子是是目前主要的边缘检测方法之一，利用一阶偏导有限差分计算梯度幅值和方向，具备非常好的检测效果，在实际工程和现实应用非常广泛。选定Canny算子与模糊提取算法做对比实验，图4选定了第一个原始图，处理结果指标数值分别为0.834、0.756、0.813。实验效果和主观效果说明，模糊边缘检测算法不仅适用于低对比度图像，同时也适用于一般的图像。但是从图6中明显看到虽然边缘有效提取出来，但存在很多异常亮点，经分析为传统Pal和King算法模糊过增强导致。The comparison experiment of the algorithm of the present invention and the classic Canny operator, Pal and King algorithm on the edge extraction effect of "rice", the Canny operator is one of the main edge detection methods at present, and the gradient amplitude and direction are calculated by using the first-order partial derivative finite difference , has a very good detection effect, and is widely used in practical engineering and reality. Select the Canny operator and the fuzzy extraction algorithm for comparative experiments. Figure 4 selects the first original image, and the processing result index The values are 0.834, 0.756, and 0.813, respectively. The experimental results and subjective results show that the fuzzy edge detection algorithm is not only suitable for low-contrast images, but also for general images. However, it is obvious from Figure 6 that although the edges are effectively extracted, there are many abnormal bright spots, which are caused by the over-enhancement of the traditional Pal and King algorithms after analysis.
实施例2：Example 2:
本发明算法与经典Canny算子、Pal和King算法对“cameraman”边缘提取效果对比实验，实验效果看出，Canny对于对比度较小的图像边缘提取效果差，分割出很多冗余信息，处理效果并不理想。而传统Pal和King算法有效的滤除了一些噪声和伪边缘，但抑制噪声能力较低，又同时存在模糊过增强现象，分析得到因空间域向模糊域转化过程参数选择不当导致，处理结果指标数值分别为0.587、0.675、0.731。本发明实验效果可以看出较传统Pal和King算法，有效抑制了高频噪声点，有效的调整了全局亮度，增强后的图像质量符合人眼的主观视觉效果，同时避免了图像过增强现象，保留了图像原有细节，而且处理过程无需人工调整参数，有效地提高了算法的自适应能力和实用性。The algorithm of the present invention is compared with the classic Canny operator, Pal and King algorithm on the edge extraction effect of "cameraman". The experimental results show that Canny has a poor edge extraction effect on images with less contrast, and a lot of redundant information is segmented, and the processing effect is not good. not ideal. However, the traditional Pal and King algorithms can effectively filter out some noise and false edges, but the ability to suppress noise is low, and at the same time, there is a phenomenon of fuzzy over-enhancement. The analysis shows that the processing result index The values are 0.587, 0.675, and 0.731, respectively. The experimental effect of the present invention can be seen that compared with the traditional Pal and King algorithms, the high-frequency noise points are effectively suppressed, the global brightness is effectively adjusted, the enhanced image quality conforms to the subjective visual effect of the human eye, and the image over-enhancement phenomenon is avoided at the same time. The original details of the image are preserved, and the processing process does not need to manually adjust the parameters, which effectively improves the adaptive ability and practicability of the algorithm.
实施例3：Example 3:
本发明设计方法与Pal和King算法边缘提取处理时间对比，实施例1和实施例2采用的Pal和King算法选择的迭代次数为2，若选取迭代次数为1时，边缘提取图像出现大量冗余信息和较多异常亮点，故没有列出实验效果。The design method of the present invention compares with Pal and King algorithm edge extraction processing time, the number of iterations selected by the Pal and King algorithm adopted in embodiment 1 and embodiment 2 is 2, if the number of iterations is 1, a large amount of redundancy appears in the edge extraction image Information and many anomalous highlights, so the experimental results are not listed.
实施例3通过采用本发明算法和传统模糊算法（本发明的迭代分别为1，传统模糊算法为2时）计算图像边缘提取时间，验证本发明算法的实时性。从表1可以看出两幅图像应用本发明边缘检测处理时间分别是0.677s和0.826s，较传统Pal和King算法分别减少了46%和43%的处理时间，有效较少了模糊边缘提取算法的处理时间，提高了算法的实时性。Embodiment 3 uses the algorithm of the present invention and the traditional fuzzy algorithm (the iterations of the present invention are 1, and the traditional fuzzy algorithm is 2) to calculate the image edge extraction time to verify the real-time performance of the algorithm of the present invention. As can be seen from Table 1, the edge detection processing time of the two images using the present invention is respectively 0.677s and 0.826s, which reduces the processing time by 46% and 43% respectively compared with the traditional Pal and King algorithms, and effectively reduces the fuzzy edge extraction algorithm. The processing time improves the real-time performance of the algorithm.
Claims (6)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610010368.5A CN105654496B (en) | 2016-01-08 | 2016-01-08 | The bionical adaptive fuzzy edge detection method of view-based access control model characteristic |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610010368.5A CN105654496B (en) | 2016-01-08 | 2016-01-08 | The bionical adaptive fuzzy edge detection method of view-based access control model characteristic |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105654496A CN105654496A (en) | 2016-06-08 |
| CN105654496B true CN105654496B (en) | 2018-06-29 |
Family
ID=56490600
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610010368.5A Expired - Fee Related CN105654496B (en) | 2016-01-08 | 2016-01-08 | The bionical adaptive fuzzy edge detection method of view-based access control model characteristic |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105654496B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106530245A (en) * | 2016-10-27 | 2017-03-22 | 长安大学 | Method for enhancing frequency band of infrared image of unmanned aerial vehicle-mounted tunnel fire hazard |
| CN108053415B (en) * | 2017-12-14 | 2020-05-22 | 广西科技大学 | Bionic contour detection method based on improved non-classical receptive field |
| CN107977945A (en) * | 2017-12-18 | 2018-05-01 | 深圳先进技术研究院 | A kind of image enchancing method, system and electronic equipment |
| CN108416788B (en) * | 2018-03-29 | 2020-06-05 | 河南科技大学 | Edge detection method based on receptive field and light and shade adaptability thereof |
| CN108830806B (en) * | 2018-05-29 | 2020-12-18 | 河南科技大学 | Sensitivity of a receptive field model and a dynamic control method for model parameters |
| CN108932702B (en) * | 2018-06-13 | 2020-10-09 | 北京微播视界科技有限公司 | Image processing method, image processing device, electronic equipment and computer readable storage medium |
| CN112396618B (en) * | 2020-11-12 | 2022-02-18 | 北京科技大学 | Grain boundary extraction and grain size measurement method based on image processing |
| CN113128511B (en) * | 2021-03-31 | 2023-07-25 | 武汉钢铁有限公司 | Coke tissue identification method and device |
| CN114861534B (en) * | 2022-04-24 | 2024-09-24 | 昆明理工大学 | A method, system, device and medium for processing visual cognitive response data |
| CN116153267B (en) * | 2022-12-30 | 2024-05-31 | 嘉兴美茵智能电子有限公司 | Backlight control method for multi-contact liquid crystal display |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1489115A (en) * | 2003-08-14 | 2004-04-14 | 中国人民解放军第一军医大学 | Image Optimal Segmentation Method Based on Generalized Fuzzy Random Field |
| CN102700569A (en) * | 2012-06-01 | 2012-10-03 | 安徽理工大学 | Mining electric locomotive passerby monitoring method based on image processing and alarm system |
| CN103871029A (en) * | 2014-01-28 | 2014-06-18 | 西安科技大学 | Image enhancement and partition method |
| CN105069807A (en) * | 2015-08-28 | 2015-11-18 | 西安工程大学 | Punched workpiece defect detection method based on image processing |
-
2016
- 2016-01-08 CN CN201610010368.5A patent/CN105654496B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1489115A (en) * | 2003-08-14 | 2004-04-14 | 中国人民解放军第一军医大学 | Image Optimal Segmentation Method Based on Generalized Fuzzy Random Field |
| CN102700569A (en) * | 2012-06-01 | 2012-10-03 | 安徽理工大学 | Mining electric locomotive passerby monitoring method based on image processing and alarm system |
| CN103871029A (en) * | 2014-01-28 | 2014-06-18 | 西安科技大学 | Image enhancement and partition method |
| CN105069807A (en) * | 2015-08-28 | 2015-11-18 | 西安工程大学 | Punched workpiece defect detection method based on image processing |
Non-Patent Citations (2)
| Title |
|---|
| 一种改进的Pla和King模糊边缘检测算法;杨勇 等;《仪器仪表学报》;20080930;第29卷(第9期);第1918-1922页 * |
| 图象模糊边缘检测的改进算法;周德龙 等;《中国图象图形学报》;20010430;第6卷(第4期);第353-358页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105654496A (en) | 2016-06-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105654496B (en) | The bionical adaptive fuzzy edge detection method of view-based access control model characteristic | |
| CN107527332B (en) | Low-illumination image color retention enhancement method based on improved Retinex | |
| Wang et al. | Adaptive image enhancement method for correcting low-illumination images | |
| Wang et al. | Biologically inspired image enhancement based on Retinex | |
| Ma et al. | Multi-scale retinex with color restoration image enhancement based on Gaussian filtering and guided filtering | |
| Park et al. | A contrast enhancement method using dynamic range separate histogram equalization | |
| CN105654436B (en) | A kind of backlight image enhancing denoising method based on prospect background separation | |
| CN104574293B (en) | Multiple dimensioned Retinex image sharpenings algorithm based on bounded computing | |
| CN106683080B (en) | A retinal fundus image preprocessing method | |
| Jeon et al. | Low-light image enhancement using gamma correction prior in mixed color spaces | |
| CN114331873A (en) | Non-uniform illumination color image correction method based on region division | |
| CN101853490A (en) | A Bionic Image Restoration Method Based on Human Visual Characteristics | |
| CN110675351B (en) | Marine image processing method based on global brightness adaptive equalization | |
| Chen et al. | Blood vessel enhancement via multi-dictionary and sparse coding: Application to retinal vessel enhancing | |
| Mustafa et al. | Image enhancement technique on contrast variation: a comprehensive review | |
| CN103489168A (en) | Enhancing method and system for infrared image being converted to pseudo color image in self-adaptive mode | |
| Wu et al. | Non-local spatial redundancy reduction for bottom-up saliency estimation | |
| CN104616259B (en) | A kind of adaptive non-local mean image de-noising method of noise intensity | |
| CN103700077B (en) | A kind of method for adaptive image enhancement based on human-eye visual characteristic | |
| CN119831890A (en) | Denoising enhancement processing method and system for sensor image data | |
| Khunteta et al. | Fuzzy rule-based image exposure level estimation and adaptive gamma correction for contrast enhancement in dark images | |
| Zhang et al. | A novel tone mapping method for high dynamic range image by incorporating edge-preserving filter into method based on retinex | |
| Pearl Mary et al. | Unified adaptive framework for contrast enhancement of blood vessels | |
| CN109410160A (en) | The infrared polarization image interfusion method driven based on multiple features and feature difference | |
| CN104915934A (en) | Grayscale image enhancement method based on retina mechanism |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180629 Termination date: 20200108 |