摘要:
What is disclosed is a system and method for identifying materials comprising an object captured in a video and for using the identified materials to track that object as it moves across the captured video scene. In one embodiment, a multi-spectral or hyper-spectral sensor is used to capture a spectral image of an object in an area of interest. Pixels in the spectral planes of the spectral images are analyzed to identify a material comprising objects in that area of interest. A location of each of the identified objects is provided to an imaging sensor which then proceeds to track the objects as they move through a scene. Various embodiments are disclosed.
摘要:
A method of, and system for, selective video frame compression and decompression is presented. The disclosed technique is particularly suited for efficient event-driven searching in large databases. The technique involves designating certain frames in a video sequence as reference frames, to be compressed using a first compression technique, where the remaining frames are compressed using a second compression technique. The first compression technique may be implemented without requiring data from any neighboring frames.
摘要:
In an inkjet printer, a method for of compensating for defects in printed images identifies a cross-process direction location of a defect in a printed image and a candidate inkjet corresponding to the location of the defect. The method modifies the operation of the candidate inkjet to form a second ink image. The method identifies a second inkjet that actually formed the first image defect in response to identifying a second defect in the second ink image located proximate to the first defect. The method enables identification and compensation of inoperable inkjets when image data do not correspond perfectly to inkjets in the printer.
摘要:
A method and system for camera calibration comprises configuring a calibration target comprising calibration reflectors on a test vehicle. Video of a test scene is collected. Next the test vehicle is identified as it enters the test scene and recorded as it passes through the test scene. The position of the calibration target in each frame of the video is determined and the corresponding individual position of each calibration reflector for each frame of the recorded frames is used to construct a camera calibration map to calibrate the video camera.
摘要:
A method has been developed for detecting defects in printed images. The method includes operating a printer to print a first printed image and a second image, where the second image is a duplicate of the first printed image. The method further includes generating image data that correspond to the first printed image and the second printed image, and identifying image defects using differences between the image data generated for the first image and the image data generated for the second image.
摘要:
This disclosure provides a video camera and video processing alert system for detecting a vehicle in reverse. According to one exemplary embodiment, the system operates according to the following guidelines or steps: (1) Acquire video containing features relevant to reverse detection, (2) Identify feature(s) within the video frame that are relevant to a vehicle in reverse, (3) Examine identified features to extract the evidence of vehicle backing up for a current frame, (4) Apply temporal filtering on the frame-to-frame evidence, (5) Use filtered evidence for decision on triggering the alarm, (6) Triggering an alarm if indicated by the decision. The system can be implemented with relative low cost and complexity, due to the affordability of video cameras, and the fact that many drive-through locations have existing video capture infrastructure.
摘要:
An embodiment generally relates to systems and methods for processing license plate image data. A processing module can receive an image comprising a license plate and, from the image, localize a region of the license plate itself. The license plate region can be extracted and license plate distortion parameters can be inferred. The image can be geometrically corrected using the inferred license plate distortion parameters, and the license plate image can be iteratively sharpened to optimize the quality of the license plate image. Further, processing parameters can be modified and applied to the image data in a feedback-based loop for further optimization.
摘要:
An approach for detecting and correcting printing system print quality defects is disclosed that allows defects to be corrected based on an analysis of image data collected over time, across many separate images, under a wide variety of system status and environmental conditions to allow detection of print quality defects that would otherwise be undetectable. The approach supports the detection and correction of defects related to colorant appearance effects that are correlated to colorant interactions. Techniques are described by which allow image data to be efficiently collected and stored to support a wide range of defect processing techniques. Defect processing may be performed in parallel with print job operations using spare processor CPU cycles and/or may be performed off-line by either the printing system processor or a separate defect processing system. Detected print quality defects may be corrected for using one or more interpolation techniques.
摘要:
A method and system for automatically determining an optimal re-training interval for a fault diagnoser based on online monitoring of the performance of a classifier. The classifier generates a soft measure of membership in association with a class based on a training data. The output of the classifier can be utilized to assign a label to new data and then the members associated with each class can be clustered into one or more core members and potential outliers. A statistical measure can be utilized to determine if the distribution of the outliers is sufficiently different than the core members after enough outliers have been accumulated. If the outliers are different with respect to the core members, then the diagnoser can be re-trained; otherwise, the output of the classifier can be fed to the fault diagnoser.
摘要:
A system for recording an image as digital data comprises a photosensor array and a plenoptic set, including a main lens and a lenticular grating, to enable the photosensor array to obtain plenoptic image data from the sheet. A subset of the plenoptic image data is selected, such as according to a predetermined object distance, to obtain image data for a focused image. The subset of the plenoptic image can be selected after the scanning, so that prescanning of an unusual-shaped image-bearing object, such as an open book, is avoided.