Abstract: A highly reliable defect automatic classifying method and apparatus capable of flexibly coping with a request for classifying a defect given by each user without having to collect lots of teach data items. A classifying class arrangement is defined by a user by combining classes supplied by the system itself or classes defined by the user and, further, a priori knowledge on the defect class is given by the user as a restriction so as to carry out restricted learning.

Abstract: A method of processing a document comprises the steps of (a) scanning a document to capture image data associated with the document, (b) processing at least a portion of the captured image data to calculate a first count of pixels associated with a corner of the document, (c) processing at least a portion of the captured image data to calculate a second count of pixels associated with the corner of the document, and (d) determining if the corner of the document is missing or has a fold based upon the first count of pixels associated with the corner of the document and the second count of pixels associated with the corner of the document. The usability of the captured image data associated with the document is determined based upon the amount of information found missing from the image of the document due to, for examples, a torn or folded corner. If the torn or folded corner is determined to be large enough, then the corresponding image of the document is determined to be unusable.

Abstract: The present invention relates to a system and method for the imaging of surfaces of box like objects conveyed along a path. The objects are conveyed past at least one camera which is positioned at an angle to the object's surface. The at least one camera comprises a sensor array and zoom objective, the latter controlled by a control mechanism. The present system further comprises a control unit and at least one sensor. The sensor detects the location and velocity of an object. The data is conveyed to the control unit which further communicates with the control mechanism. The control unit causes the control mechanism to adjust the zoom on the objected so as to stay in constant focus with a passing surface of an object passing through the field of view of the at least one camera. After an object has passed through the field of view, the object is made to return to a home position.

Abstract: A spout control system controls and aims a spout and a spout cap of a crop harvesting vehicle with respect to a separate crop hauling vehicle moving with the harvesting vehicle. The control system includes a video camera which is mounted on the cap and which views a field of view which includes a portion of the hauling vehicle. An image signal generated by the camera is received by an image processing unit. The image processing unit processes a digitized form of the image signal and automatically generates spout and cap control signals as a function thereof. Actuators automatically aim the spout and the cap in response to the control signal.

Abstract: The read head (1) for a postal sorting installation serves to pick up a digital image of each postal object using a high-resolution camera (3) providing a multiple level gray scale. The read head also has a low-resolution color camera (4), the two cameras (3, 4) being placed in such a manner as to acquire simultaneously two superposable digital images of each postal object. The superposable high-resolution gray scale image and low-resolution color image are intended to be combined in order to provide a high-resolution color image.

Abstract: A lumber grading system for providing a lumber cutting solution for a piece of lumber. Video acquisition circuits acquire complementary images of the sides of the piece of lumber taken by two cameras. A computer unit performs an image processing of the images and generates the lumber cutting solution for the piece of lumber based on preset grade determination parameters. The image processing involves a longitudinal division of the images of the piece of lumber into a number of consecutive zones based on predetermined lumber cutting positions, an attribution of respective grades to the zones based on an image analysis of the zones as a function of the preset grade determination parameters, and a solution search providing the lumber cutting solution as a function of the grades attributed to the zones. A communication circuit transmits the lumber cutting solution.

Abstract: At the separation of different minerals from the ores by means of flotation, it belongs to known art to monitor a larger part of the uppermost froth layer in the flotation cell by means of a video camera whose signals are analyzed in a digital computer, in order to detect the structure and color of the froth. The new method and apparatus direct and limit the monitoring and analysis to the surface of the floated material within a fixed, narrow strip which is parallel to the overflow edge of the cell and within which the material surface passing it is homogeneous in the stationary state. The said strip thus presents a renewing sample of the surface at the location in question, which sample represents the stationary state in the average sense. The strip is subjected to homogeneous illumination, which may meet the strip at different angles depending on the primary object, i.e. whether the color or structure is being determined.

Abstract: The assessment of the effects of yarn faults is carried out by simulating the fabric image. In a first step, the yarn is examined by a measuring member for parameters associated with the volume and/or the surface. In a second step, these parameters are converted into grey values or color values, and these values are assigned to image spots. Finally, the image spots are reproduced on a video display unit and/or a printer. An image is generated thereby, representing a simulation of a woven or knitted fabric produced from the examined yarn.

Abstract: A lumber grading system for providing a lumber cutting solution for a piece of lumber. Video acquisition circuits acquire complementary images of the sides of the piece of lumber taken by two cameras. A computer unit performs an image processing of the images and generates the lumber cutting solution for the piece of lumber based on preset grade determination parameters. The image processing involves a longitudinal division of the images of the piece of lumber into a number of consecutive zones based on predetermined lumber cutting positions, an attribution of respective grades to the zones based on an image analysis of the zones as a function of the preset grade determination parameters, and a solution search providing the lumber cutting solution as a function of the grades attributed to the zones. A communication circuit transmits the lumber cutting solution.

Abstract: The present invention aims at providing a part mounting device for efficiently performing mounting operations by eliminating an amount of time required for moving and stoppage intended for recognizing an image of a part. In the part mounting device according to the present invention, an image recognizing device included in a working tower together with a mounting head comprises an optical fiber, a beam splitter arranged at the top of the optical fiber, and an image recognizing camera arranged at the top of the beam splitter. The image recognizing device further comprises an illumination light source at the side of the beam splitter. A nozzle shaft of the mounting head is arranged by being loosely inserted into a lower portion of the optical fiber. An irradiation light from the illumination light source irradiates a part which is picked up by the picking-up nozzle attached to the tip of the nozzle shaft via the beam splitter and the optical fiber.

Abstract: A video timing generator provides control and timing signals for a variety of different camera types by employing a circular buffer scanned rapidly before the occurrence of each potential control and timing signal. The circular buffer holds operation codes for generating the control and timing signal linked to target event signal times. An operation code having a target event signal time matching with the next event time is decoded and queued to be executed at that event time.

Abstract: The system comprises a high resolution camera (1) mounted above a conveyor plane (4) along which objects are displaced, thereby enabling the top face of each object to be observed. The camera (1) is fitted with an objective lens system (5) having a motor-driven focusing system that operates in association with a pickup (8) that produces a signal (S) representative of the vertical profile of the top face of each object relative to the conveyor plane. The profile signal serves to produce positioning references (C) for the focusing mechanism in order to take account of variations in the height of the top face of each object while that object is moving beneath the objective lens system of the camera. Such an image acquisition system is employed to read addresses automatically from postal packets for the purpose of performing automatic sorting.

Abstract: Disclosed is an appearance inspecting device for a solid formulation, including a supply section for continuously supplying solid formulations such as capsules; an inspecting drum for holding on its outer peripheral portion the solid formulations supplied from the supply section, carrying the solid formulations in such a manner as to allow the solid formulations to be intermittently revolved, and allowing the solid formulations to be rotated on their axes at a specific carrying position; an image pickup device for photographing surfaces of the solid formulations being thus rotated on their axes; a quality determining section for processing images of the solid formulations thus picked up from the image pickup device and determining whether the solid formulations are nondefective or defective; and a classifying section for classifying the solid formulations into defective and nondefective solid formulations on the result determined by the quality determining section and recovering the defective and nondefective

Abstract: A plurality of separate conveyings over each of which a single line camera is mounted and connected to a sensor processor, which itself is connected to a fast data network with a plurality of processing terminals. Because of the high resolution (up to 7000 picture elements per line) and a line scanning frequency of the order of up to 6000 hertz very large amounts of data are received, so in the preferred embodiment a high capacity data network is used, for example a 100 Mbit ethernet with a switching hub.

Abstract: Apparatus and method are described for measuring and controlling the crimp characteristics of a moving crimped tow. A light source illuminates a section of the moving crimped tow and at least one camera acquires a video image of the tow. The acquired image is digitized and a processor decomposes the interlaced image into two non-interlaced field images. Crimp characteristics are derived based on the decomposed images.

Abstract: A lumber video grading apparatus includes a vision system, processors, a grading station having visual displays, a storage and retrieval device, an identification device, and a timing system. The vision system may include cameras mounted proximate to a lumber transfer for optically acquiring real time images of exposed cut surfaces of lumber pieces conveyed on the lumber transfer. The processors process the real time images into individually identifiable retrievably storable images and also into visually displayable real time images. A selectable lumber grade is correlated at a grading station to selectable pieces of lumber on the lumber transfer. The displayable real time images or later retrieved individually identifiable retrievably storable images are displayable on a visual display at the grading station. The individually identifiable retrievably storable images are retrievably storable on, and selectively retrievably from, the storage and retrieval device.

Abstract: An optoelectronic system for reading reliefs, such as a mould number, on a transparent container. The optoelectronic system associates a light source and a reception system, the light source illuminating the entire periphery of part of the body of the container and the reception system being capable of reproducing a complete image of the periphery of that part of the body of the container.

Abstract: An automated system and associated method of counting pillings in textile fabrics. Images of consecutive sections of a fabric sample are captured by a CCD camera. A system processor then processes the images to enhance image quality. Pillings on the processed, captured fabric sample images are counted, and the resulting data is fuzzified to determine the membership of the data in one or more of a plurality of fabric classes. The present invention provides an objective rating system and method with repeatable accuracy for fabric samples that eliminates the inherent subjectivity associated with conventional manual visual fabric inspection methods.

Abstract: An image recording system is described for automatic address location on large letters and parcels, comprising a high resolution grey scale camera and a low resolution color camera. Both cameras are directed at one illumination slit past which the object to be processed is fed. The color camera is operated in a special mode which allows it to manage with only a single illumination slit, where adequate light intensity is provided for both cameras. A method for automatic address recognition on large letters and parcels is also provided, together with a method of segmenting color images in the recognition of address blocks.

Abstract: A method and apparatus for inspecting and sorting agricultural products which includes launching each product individually from an input transport unit with substantially no change in speed or direction across a gap to an output transport unit. Fixed line scan camera and photocell units are mounted at spaced intervals around the gap to simultaneously optically line scan in the same plane the surfaces of the product. A plurality of simultaneous line scans are sequentially taken at intervals timed to the speed of the input transport unit as the product travels across the gap, and data obtained from the sequential line scans is used to determine the size, shape, weight, color and other characteristics of the product. The product is graded based upon this characteristics and is sorted based upon its grade by a mechanical sorting unit.

Abstract: A pair of cameras are mounted on a base member for monitoring an electronic equipment on a printed circuit board. A pair of first mirrors are provided for reflecting light from the electronic equipment in a first reflection direction which corresponds to an X-axis on the electronic equipment, and a pair of second mirrors are mounted for reflecting the light reflected on the first mirrors in a second reflection direction which is different from the first reflection direction and corresponds to a Y-axis. The base member is moved in two directions corresponding to the X- and Y-axis.

Abstract: The Invention includes a method and apparatus which analyzes an image of an object to detect and identify defects in the object. The present invention utilizes a scanning technique which converts a 2-D image of the object into a 1-D image, a transformation technique which extracts relevant features from the image, and a fuzzy inferencing technique which utilizes the features generated by the transformation technique to detect and identify defects. The present invention preferably includes an off-line learning process which selects the optimum transform coefficients for a given set of defects and stores the corresponding features in a rulebase. Preferably, the wavelet transform is used as the transformation technique to provide an analysis of the image which is localized in the frequency and time domains. The present invention may also include an on-line learning process when the present invention is incorporated into a manufacturing process for real time inspection of the object being manufactured.

Abstract: In a method and apparatus for generating colorimetric data useful in the automatic colorimetric sorting of products, such as fruits or vegetables, each product is illuminated by means of a beam producing a succession of lines of light, the energy reflected by the product in preselected wavelengths is reconstituted for each point on each line of light, the light intensity of each point is measured, the measured values are converted so as to form a series of numerical data corresponding, for each wavelength, to the light intensity curves for each line of light, and the series of numerical data are processed, by computing, in accordance with programmed criteria based on a comparison of the values of the homologous points of said series, so as to generate usable colorimetric data.

Abstract: Grain defect scanning takes into account a broad set of data representing both wood grain structure and wood grain image to provide a multi-dimensional scan vector for an inspection point with wide variation therein relative to defect types. A library of similarly structured multi-dimensional training set vectors developed during a preliminary training session with known defect types is referenced by multivariate pattern recognition analysis to classify a collection of scan vectors associated with an article under inspection. By statistically matching scan vectors with training set vectors under pattern recognition analysis, physical locations on a wood article are identified according to known defect types.

Abstract: An object is sorted into different classes by producing a video image of the object. The signal level discontinuities in the image are evaluated by first deriving a video signal gradient of each pixel in the image. Then the number of pixels in three different signal gradient ranges are counted to produce count parameters which combined indicate the areas of differing contrast in the object. The count parameters are utilized in a pair of discriminate functions which define the two classes and solution of the discriminate functions are used to determine the probability that the object fits within one class. In response to the probability the object is directed along either a first path associated with one class or along a second path associated with the other class.

Abstract: A sorting system (10) propels a stream of randomly arranged PET and PVC articles (12, 14) through an inspection zone (20) including a first light polarizer/analyzer combination (24, 26), an article-detecting gap (G), and a second light polarizer/analyzer combination (28, 30). The first and second polarizer/analyzer combinations are oriented to extinguish normally incident light in the absence of articles in the inspection zone and are offset 45 degrees relative to each other such that at least one polarizer/analyzer combination detects a principal axis of birefringence of PET articles. The gap is employed to detect the presence of an article in the inspection zone. A video camera (22) includes first, second, and third CCD arrays (58, 60, 62) positioned to receive respective light rays (48, 64, 50) from the first light polarizer/analyzer combination, the gap, and the second light polarizer/analyzer combination and to generate first, second, and third video signals representative of the light each receives.

Abstract: The invention is a process for inspecting chips and/or fragments of metal or metal alloy to eliminate from them inclusions of a more X-ray absorbent material than the metal or alloy, wherein a field that delimits a portion of these chips and/or fragments is X-rayed. The process produces an X-ray image which is then converted into an electronic image. This image is analyzed in order to detect the inclusion(s) having features. The features include a background correction of the degree of illumination of each pixel in the field in the absence of chips and/or fragments, a field is covered with the portion of chips and/or fragments, and the electronic image is corrected by subtracting the background correction from the degree of illumination of each of its pixels, and the portion of chips and/or fragments is rejected if this corrected image contains at least one relative pixel corresponding to a chosen condition.

Abstract: A method and apparatus used with a printed circuit board (PCB) defect detection system generates a reference database image of the PCB having tolerances for each individual PCB feature, The invention is characterized by a tolerance database of at least three states in which each color is weighted and adjacent pixels are grouped into arrays or "bins". An error signal is generated when the sum of the pixel weights in a bin exceeds a preselected threshold.

Abstract: A continuous-motion flying-spot telecine has twin objective lenses (13,14) which provide beams to scan two adjacent film frames simultaneously. Two sets of light detectors (16,17) are provided, one for each frame being scanned. The detector outputs are applied to a multiple frame store (23; FIG. 2). The scans can take place at half the normal speed, while a conventional rate signal is read out of the store. Problems caused at high frequencies as a result of afterglow on the CRT (10) can thereby be substantially reduced.

Abstract: An integrated bulk food sorting and analysis apparatus comprises a wide-belt product conveyor which receives and conveys a laterally-distributed stream of bulk food articles past a product diverter. An upstream camera is positioned to view the stream of food articles upstream of the product diverter. Automated sorting logic is responsive to the upstream camera to individually determine optical characteristics of each food article. The product diverter is responsive to the automated sorting logic to divert a portion of the food articles from the stream depending upon their individual optical characteristics. The apparatus also includes a downstream camera positioned to view the stream of food articles downstream of the product diverter. A data processor is responsive to both the upstream camera and the downstream camera to periodically examine samples of food articles and to calculate upstream and downstream quality statistics regarding the stream of food articles.

Abstract: The invention provides apparatus and methods for concurrently acquiring video data from multiple video data sources. The invention comprises an asynchronous interface that collects pixel data from multiple video data sources. A packet manager collects the pixel data and generates a packet data signal that represents the pixel data collected from the multiple video data sources. The packet manager transmits the packet data signal to a pixel buffer memory that includes an address generator that separates out the pixel data associated with each of the video data sources and that stores the pixel data in a video memory in a manner suitable for video display.

Abstract: A method is provided for detecting the color of objects such as bottles to be recycled so as to enable the bottles to be separated by color. The method which is particularly concerned with determining the color of an object (bottle) with a label attached thereto, begins with the storing of a reference image. Next, a determination is made of whether a bottle is present or absent. If a bottle is present, the image of the bottle is then separated into an intensity image representation of the bottle image, a saturation image representation of the bottle image, and a hue image representation of the bottle image. Then, the reference image is subtracted from the intensity image so that a resultant image is produced with the background contribution removed. Next, a logical "AND" operation is performed on the resultant image and the hue image to create a further image which has no background components. A histogram is then performed on the further image to determine the hue with the maximum number of occurrences.

Abstract: An apparatus and a method sort items, such as pickles, having defective shapes, such as holes. A video camera scans moving items and acquires lines of image data including item-colored data and background-colored data. The lines of image data are sent to an image processor and compiled into a 64 line "frame" of image data that are further processed to identify the coordinates of individual items and to compute the size of each item. Because holes in an item are the same color as the background color, the invention provides means of differentiating holes from the background. A "hole-bounding box" is scaled to fit within the item position coordinates. The number of background-colored pixels inside the hole-bounding box is computed and compared to a user-defined threshold number. An item is classified as defective if the threshold number is exceeded.

Abstract: A system and method for sorting items (16) computes the geometric center ("centroid") (156) of any item containing a defect (26) or multiple defects, and directs an ejection air blast at the centroid of the defective item rather than at the location of the defect. Video data from a scanning camera (24) are transmitted to an "item processor" (32A') and a "defect processor" (32). The item processor builds in memory (108) an image of every acceptable or defective item while the defect processor builds a "defect list" (170) of defect coordinate locations detected only on defective items. The defect processor transmits the defect list to the item processor where the defect list is compared with the stored image of the item. For each item containing at least one defect, the item processor computes a defective item centroid that is added to a defective items list (174) for use by a defect removal process that actuates air blasts directed toward the centers of defective items.

Abstract: In an article handling system that functions to make article discrimination-identification determinations, the possible article orientations A, B, C . . . are stored and compared to establish maximum pixel difference numbers and identification of longitudinal window position along the article where such maximum difference occurs. Each of the possible orientations is compared with all other possible orientations so that a small number of windows is identified as part of a setting procedure. When operating, the work articles are scanned and only the data at window locations are used to make article orientation identifications to reduce time and memory requirements for data processing.

Abstract: Apparatus and methods are disclosed for automatically inspecting two- or three-dimensional objects or subjects. A detector and the object are moved relative to each other. In one form, a detector, such as a camera or radiation receiver, moves around an object, which is supported to be rotatable such that the detector may receive electromagnetic energy signals from the object from a variety of angles. The energy may be directed as a beam at and reflected from the object, as for visible light, or passed through the object, as for x-ray radiation. Alternatively, the detector passively receives energy from the object, as in an infrared detector. The detector generates analog image signals resulting from the detected radiation, and an electronic computer process and analyzes the analog signals and generates digital codes, which may be stored or employed to control a display.

Abstract: An apparatus and a method sort items, such as pickles, having defective shapes, such as holes. A video camera scans moving items and acquires lines of image data including item-colored data and background-colored data. The lines of image data are sent to an image processor and compiled into a 64 line "frame" of image data that are further processed to identify the coordinates of individual items and to compute the size of each item. Because holes in an item are the same color as the background color, the invention provides a device differentiating holes from the background. A "hole-bounding box" is scaled to fit within the item position coordinates. The number of background-colored pixels inside the hole-bounding box is computed and compared to a user-defined threshold number. An item is classified as defective if the threshold number is exceeded.

Abstract: In a method of distributing packages according to addresses, headings or the like applied to their surfaces by means of a separating and conveying device, it is provided that along a conveying path all surface portions of each previously separated package are scanned optically, for each package images are obtained from the surface portions of the package and, for further evaluation, the images are displayed simultaneously or approximately simultaneously. For each package, one or a plurality of the images are selected, and each selected image is assigned a sorting information for the distribution of the package corresponding to that image.

Abstract: A system and method for sorting items (16) computes the geometric center ("centroid") (156) of any item containing a defect (26) or multiple defects, and directs an ejection air blast at the centroid of the defective item rather than at the location of the defect. Video data from a scanning camera (24) are transmitted to an "item processor" (32A') and a "defect processor" (32). The item processor builds in memory (108) an image of every acceptable or defective item while the defect processor builds a "defect list" (170) of defect coordinate locations detected only on defective items. The defect processor transmits the defect list to the item processor where the defect list is compared with the stored image of the item. For each item containing at least one defect, the item processor computes a defective item centroid that is added to a defective items list (174) for use by a defect removal process that actuates air blasts directed toward the centers of defective items.

Abstract: Mode selection switches (76, 96, 98) selectively interconnect a sensor line shift timing generator (76) with one of three signal sources - (1) a conventional raster scan timing signal from a raster scan sync generator (70), (2) variable speed external timing signals from a tachometer (32), and (3) fixed speed timing signals developed from the horizontal timing signals of the raster scan sync generator. In a time delay and integration mode, the sensor line shift timing generator causes the CCD arrays of an image section (22) and storage section (24) to shift pixel values down the CCD arrays at a rate commensurate with the external timing signal. As a spot of light emanating from a portion of an object moving through an examination region moves along the CCD array, a corresponding pixel charge is shifted through the CCD array at the same speed such that the same pixel charge integrates light from the same spot as it moves the entire length of the CCD array.