Analog Integrated Circuit for Motion Detection against Moving Background Based on the Insect Visual System

A network for displaying the velocity signal of a target edge which has larger velocity than the background was proposed and designed based on an insect visual system. In the insect visual system, the velocity signals generated by the background are inhibited by the signals generated by averaging the velocity signals from each cell and can only display the velocity signals of the target. The proposed network was constructed with simple analog circuits. The measured results of the test chip fabricated with the 1.2$mUm complementary metal-oxide-semiconductor (CMOS) process showed that the edge detection circuit inserted to the first stage of the proposed network can detect the edge position. It was apparent from the measured results of the test circuit constructed with discrete metal-oxide-semiconductor (MOS) transistors that the proposed unit circuit constructing the network can generate the velocity signals by inputting the edge signals. The results with the simulation program with integrated circuit emphasis (SPICE) showed that the proposed network can only display the velocity signals of the target edges but not those of the background. Thus, it was clarified that the various motion directions of the target can be detected independent of the background by using the proposed network.     

Simple Analog Complementary Metal Oxide Semiconductor Circuit for Generating Motion Signal

We proposed in this study a novel analog complementary metal oxide semiconductor (CMOS) circuit for generating a motion signal when an object moves, which is a simple structure. The proposed unit circuit was constructed using a previously proposed edge detection circuit and a novel proposed circuit for generating a motion signal which accepts an edge signal. The part for generating the motion signal was constructed using six metal oxide semiconductor (MOS) transistors and one capacitor. Results obtained by the simulation program with integrated circuit emphasis (SPICE) and the measured results of a test circuit constructed with discrete MOS transistors and the test circuit fabricated with a 1.2 μm CMOS process showed that the proposed unit circuit can output pulsed current (motion signal) when an object moves on the circuit. It was clarified from the SPICE results that the two-dimensional network constructed with proposed unit circuits can output motion signals. The size of the novel unit circuit is expected to be about 110 × 110μm² obtained by the 1.2 μm CMOS process. It is possible to arrange 90 × 90 unit circuits on a chip which has an area of 1 × 1cm². The aperture ratio is expected to be about 21%, which is twice as large as that of the previously proposed circuit. An integrated circuit for image processing in real time can thus be realized by applying the two-dimensional network constructed with the proposed circuits.     

Analog Metal-Oxide-Semiconductor Integrated Circuit Implementation of Approach Detection with Simple-Shape Recognition Based on Visual Systems of Lower Animals

We proposed an analog network for motion detection of an approaching object with simple-shape recognition based on the visual systems of lower animals. Locusts can detect an approaching object by performing a simple process in a descending contralateral motion detector (DCMD) existing in their brain, which senses the increase in size and expansion velocity of the image projected on a retina just before collision. The responses correspond to the approaching velocity and direction. Frogs can recognize a simple shape by performing a simple process in a tectum and thalamus existing in their brain based on retinal information; this is called the Ewert von Seelen model. The proposed network was constructed with simple analog metal-oxide-semiconductor (MOS) circuits. Simulation results with a simulation program with integrated circuit emphasis (SPICE) showed that the network was able to detect the approaching velocity and direction, and also recognized simple shapes such as a circle, square, triangle and rectangle.     

Analog Integrated Circuit for Motion Detection of Approaching Object Based on the Insect Visual System

A network for detection of an approaching object was proposed and fabricated based on the transient response of a descending contralateral movement detector (DCMD) existing in the brain of locusts. The proposed network was constructed with simple analog circuits. The experimental results of a test chip fabricated with a 1.2 $mUm complementary metal-oxide-semiconductor (CMOS) process and the results with a simulation program with integrated circuit emphasis (SPICE) showed that the proposed network is able to detect the approach by generating a peak current just before collision; the peak current allows detection of the approaching velocity and direction without collision. The proposed network could be applied to two-dimensional arrays for three-dimensional motion detection.     

A Network of Analog Metal Oxide Semiconductor Circuits for Motion Detection with Local Adaptation to a Background Image

Inspired by a mechanism of biological vision systems, a model and a network of analog metal oxide semiconductor (MOS) circuits are proposed which display an optical flow with local adaptation to the relative velocity of a background image. A function of displaying an optical flow successfully worked as a result of simulations using the simulation program with integrated circuit emphasis (SPICE). A function of varying an optical flow at a certain instant by local adaptation was also demonstrated. The proposed network is suitable for the realization of a large-scale integrated circuit (LSI), which displays an optical flow with local adaptation to the local velocity of a background.     

Analog vision chip for motion detection of an approaching object against a moving background based on the insect visual system

A network for motion detection of an approaching object without influence of the moving background was proposed based on the insect visual system. The two-dimensional array of the analog complementary metal oxide semiconductor (CMOS) circuits for extracting the edge signals of the approaching object without outputting those of the moving background was inserted at the first stage of the network. At the next stage, analog CMOS circuits for detection of approaching velocity and direction, which accept the extracted edge signals, were inserted based on the locust vision system. A chip of the proposed network was fabricated with the 1.2 mm CMOS process. It was clarified from the measured results and the simulation results with the simulation program with integrated circuit emphasis (SPICE) that the two-dimensional circuits of the first stage can only generate the signals of target edges. The measured results showed that the chip can detect the approaching velocity and direction of the target in the image which contains the moving background.     

An Integrated Circuit for Two-Dimensional Edge-Detection with Local Adaptation Based on Retinal Networks

We designed an integrated circuit for edge detection of a two-dimensional image based on the vertebrate outer retina, which has wide dynamic range in image processing. The unit circuit is simple, and operates as a current-mode analog metal-oxide-semiconductor (MOS) circuit. In order to extract edges from an image composed of bright and dark domains, the circuit realizes a function called local adaptation in which the sensitivity adapts to local brightness of the image. Simulation results, using the simulation program with integrated circuit emphasis (SPICE), of two-dimensional Gaussian-distributed images in which the intensity ranged over four orders of magnitude, showed the local adaptation. As a result, the intensity of output images was in the range of one order of magnitude. Furthermore, as the simulation result of real images, it was shown that edges in the dark domain, which was five times darker than the bright domain, were successfully detected as the bright domain in which input photocurrents ranged over two orders of magnitude.     

A Complementary Metal-Oxide-Semiconductor Vision Chip for Edge and Motion Detection with a Function for Output Offset Cancellation

We have designed and fabricated a complementary metal-oxide-semiconductor (CMOS) vision chip by modeling cells of the human retina as hardware that are involved in edge and motion detection. There are several fluctuation factors which affect the characteristics of metal-oxide-semiconductor field effect transistors (MOSFETs) through the CMOS fabrication process and this effect appears as the output offset of the vision chip, which is composed of pixel arrays and readout circuits. The vision chip which detects edge and motion information from an input image is used for the input stage of other systems. Therefore, the output offset of the vision chip determines the efficiency of the entire system. In order to eliminate the offset at the output stage, we designed a vision chip utilizing the correlated double sampling (CDS) technique. The chip has been fabricated using a 0.6 m standard CMOS process. With reliable output characteristics, this chip can be used at the input stage for various applications. © 2005 The Optical Society of Japan     

Signal Formation of Image-Edge Motion Based on Biological Retinal Networks and Implementation into an Analog Metal-Oxide-Silicon Circuit

A simplified model was proposed for the formation of edge signals and generation of motion signals of a target based on the information processing mechanisms of outer and inner retinas of a vertebrate. Analog metal-oxide-silicon (MOS) integrated circuits were designed based on the model. Simulation program with integrated circuit emphasis (SPICE) simulation results showed the performance of local adaptation over a wide dynamic range in the outer retinal circuit and generation of the velocity signal of a moving edge in the inner retinal circuit. Preliminary experimental results showed local adaptation in a given input range in the outer retinal circuit and the generation of motion pulsed signals in the inner retinal circuit.     

A Bio-Inspired 128 × 128 Complementary Metal–Oxide–Semiconductor Vision Chip for Edge Detection with Signal Processing Circuit Separated from Photo-Sensing Circuit

In this paper, a vision chip for edge detection based on the structure of biological retina is introduced. The key advantage of retinal structure is speed of operation. However, bio-inspired vision chips have suffered from low resolution which is caused by their complex circuit structure. To sufficiently improve the resolution for real application, the circuits for analog processing were separated from circuits for image capturing. In addition, we compensated the low speed problem of operation which is caused by bottleneck of data transfer between photo-sensors and analog processing circuit by adding a reset decoder. The vision chip was fabricated using 0.35 μm 2-poly 4-metal complementary metal---oxide---semiconductor technology. Using the vision chip, we could obtain a contrast-enhanced image without any other cost-increase for fabrication of chips. Then, the edge image was easily achieved by thresholding the previous contrast-enhanced image.     

基于神经网络的零件缺陷机器视觉识别系统

零件缺陷检测是保证零件使用安全的重要手段。传统的零件缺陷检测法需要有操作人员参与其中,易受主观因素影响,检测的效率及精度得不到良好的保证。而采用机器视觉技术的检测法可实现实时在线的自动检测,无需人工参与,这就极大的提高了生产效率。本文以小轴承表面为研究对象,针对微小轴承的表面结构、尺寸、检测精度和缺陷特征,设计了基于BP神经网络的零件缺陷机器视觉在线自动检测系统,其采用机器视觉技术,构建了BP神经网络检测识别模型,采用进行图像特征提取的间接识别方法,对微小轴承缺陷进行实时检测。实验结果证明了人工神经网络模型的检测能力的可靠性。     

基于机器视觉的西林瓶尺寸检测

在西林瓶生产过程中,尺寸是一项重要的产品质量判断标准,与传统的西林瓶尺寸人工检测方法相比,基于机器视觉的自动检测具有巨大优越性。为实现西林瓶尺寸的检测,提出了一种基于机器视觉的西林瓶尺寸检测方案,设计了系统的图像采集和背光源照明方案,通过中值滤波对图像进行去噪,利用对图像像素点的运算算法,对图像的灰度进行了校正变换,增强图像的对比度,采用Canny算子成功提取西林瓶边缘,在HALCON平台下实现了西林瓶尺寸测量。设定系统标定方法并选取15个2mL样品西林瓶进行测试,结果表明,该方法对西林瓶尺寸检测快速准确,边缘量化精度达到了亚像素级别,检测精度为0.02mm,满足西林瓶生产的参数测量精度要求,为工业生产产品尺寸的自动检测提供了一种有效的新途径。     

基于节点导纳矩阵的模拟电路故障诊断

针对模拟电路故障诊断中的容差问题,提出了基于节点导纳矩阵(NAM)的模拟电路故障诊断方法。该方法以NAM为基础,提取被测电路(CUT)的故障特征向量。测试前,用仿真的方法生成被测电路中某一故障对应的故障样本子集,所有类别的故障样本子集构成故障样本集。测试时,测量被测电路的故障特征向量,并根据其与故障样本集中样本的相似性来判断电路发生的故障类型。由于电路的NAM对元件容差不敏感,所以可以很好地克服模拟电路故障诊断中的容差问题。实验结果证明了该方法的有效性。     

基于运动检测的多车辆跟踪方法研究

针对交通监控场景中多目标粘连造成跟踪上的困难和前后两帧车辆关联困难,提出了区域运动相似性分割方法和相似度关联矩阵的解决方案;在运动目标检测过程中, 首先使用背景差分法提取运动区域,经过消除缺口、空洞和分离等处理,在运动区域所在范围内进行块匹配搜索和局部光流计算区域运动矢量,然后使用模糊聚类方法对运动矢量区域融合,完整的分割出粘连运动目标;在目标跟踪部分,目标跟踪建立在目标关联的基础上,提出建立连续两帧目标间距离和局部二元模式相似度关联矩阵的方法进行运动目标标定,从而实现多目标关联;使用公共视频库的图像序列进行测试,所提算法都能实现连续的跟踪和准确的运动目标分割,且处理速度快,表明了算法具有鲁棒性和适用性。     

Fabricating Microbolometer Array on Unplanar Readout Integrated Circuit

In this paper, the integration of an experimental 32 × 32 uncooled IR microbolometer array with an unplanar CMOS Readout Integrated Circuit (ROIC) is presented. A vanadium oxide film fabricated by low temperature reactive ion beam sputtering is utilized as thermal-sensitive material in the bolometric detectors Before the integration, the unplanar ROIC for commercial use is first planarized by bisbenzocyclobutene film, then a electroless nickel-plating on ohmic contact areas is accomplished. Finally the bolometer array is fabricated using a micromachining process, which is completely compatible with CMOS technology. Measurements and calculations for the as-fabricated samples show that the responsivity of 1.4 × 10⁴ V/W and the detectivity of 2.1 × 10⁸cmHz^1/2W^–1 and a thermal response time of 10ms are obtained at a pulse bias of IV.     

基于光谱探测与图像识别的反伪装目标系统研究

为了在野外环境中快速有效地识别敌方伪装的机动目标,设计了基于光谱探测与视频图像目标识别方法联用的目标识别系统。采用视频图像识别技术获取被测区域的二维影像,再通过光谱探测技术识别目标,最终将目标重建在图像相应位置上从而实现目标识别的可视化。理论推导得到了系统可识别目标的函数关系式,根据该函数关系进行了目标识别的量化实验。实验采用汽车模拟被测机动目标,在不同距离上分别以平坦荒地、灌木丛和废弃建筑物为背景,对明显目标、涂覆迷彩色的目标以及遮挡伪装物的目标分别进行光谱探测。实验结果显示,测试背景对光谱探测效果有一定影响,背景的连续性有利于目标识别;伪装方式以伪装物遮挡最难识别,且随着目标与系统的距离增大而信噪比随之降低。综上所述,采用光谱探测技术克服了传统图像目标识别无法识别伪装目标的缺点,可以实现对伪装目标的有效识别。     

基于IEEE1232的模拟电路TPS自动生成系统开发

目前在模拟电路故障诊断及测试过程中存在两个问题：测试信号的连续性及容差特性造成的测试信号数量巨大,故障知识表示复杂,测试程序（Test Program,简称TP）的编写多用基于决策知识的人工生成方法。通过对IEEE1232标准的体系结构和诊断推理机要求的分析,论文对IEEEE1232模型体系进行扩充,提出一种包含特征提取技术和多种AI诊断方法的诊断知识库生成协议,设计并实现了符合1232标准知识库的TPS自动生成测试系统。提高了诊断知识的移植性,实现了TPS的自动生成。仿真结果证明了该方案的可行性     

基于双目视觉的边缘检测算法

提出了一种基于立体视觉的物体边缘检测的方法。先对立体图像对进行基于图割的立体匹配方法求取场景的视差图，然后再用Canny的边缘检测方法对视差图进行边缘检测。立体视觉方法有效解决了单目视觉检测方法中的一些难点，利用了物体在空间的深度信息，对复杂背景下的物体和具有复杂纹理物体的边缘检测有很高的鲁棒性。实验结果表明该边缘检测方法优于传统的单目视觉边缘检测方法。     

面向自动测试的集成电路制造装备测试模型

对于典型的集成电路制造装备,提出了一种面向自动测试系统的多层次测试模型;使用层次化建模方法进行建模,将典型集成电路制造装备的测试模型按层次划分为装备级、子系统级和仪器级;利用XML建立了模块化的测试模型的描述,并使用引用的方式实现测试模型的重用;提出了使用时间信息和GUID确定被引用测试模型可用性的算法;该模型被用于具有工艺腔室的典型集成电路制造装备中,提高了测试模型的重用性,并拓展了对不同层次被测对象的测试能力。     

A Study on Railway Surface Defects Detection Based on Machine Vision

The detection of rail surface defects is an important tool to ensure the safe operation of rail transit. Due to the complex diversity of track surface defect features and the small size of the defect area, it is difficult to obtain satisfying detection results by traditional machine vision methods. The existing deep learning-based methods have the problems of large model sizes, excessive parameters, low accuracy and slow speed. Therefore, this paper proposes a new method based on an improved YOLOv4 (You Only Look Once, YOLO) for railway surface defect detection. In this method, MobileNetv3 is used as the backbone network of YOLOv4 to extract image features, and at the same time, deep separable convolution is applied on the PANet layer in YOLOv4, which realizes the lightweight network and real-time detection of the railway surface. The test results show that, compared with YOLOv4, the study can reduce the amount of the parameters by 78.04%, speed up the detection by 10.36 frames per second and decrease the model volume by 78%. Compared with other methods, the proposed method can achieve a higher detection accuracy, making it suitable for the fast and accurate detection of railway surface defects.