A method and apparatus for removing undesired features in an image. The image is converted from a spatial domain to a frequency domain to form a transformed image. A filter is applied to the transformed image to form a filtered image in the frequency domain. The filtered image is converted from the
A method and apparatus for removing undesired features in an image. The image is converted from a spatial domain to a frequency domain to form a transformed image. A filter is applied to the transformed image to form a filtered image in the frequency domain. The filtered image is converted from the frequency domain back into the spatial domain to form a modified image. An intensity of an undesired feature in the modified image is increased as compared to the intensity of the undesired feature in the image. The undesired feature is removed from the image using the modified image to form a processed image.
대표청구항▼
1. A method for removing an undesired feature in an image, the method comprising: converting the image from a spatial domain to a frequency domain to form a transformed image;applying a filter to the transformed image to form a filtered image in the frequency domain;converting the filtered image fro
1. A method for removing an undesired feature in an image, the method comprising: converting the image from a spatial domain to a frequency domain to form a transformed image;applying a filter to the transformed image to form a filtered image in the frequency domain;converting the filtered image from the frequency domain back into the spatial domain to form a modified image, wherein an intensity of the undesired feature in the modified image is increased as compared to the intensity of the undesired feature in the image; andremoving the undesired feature from the image, wherein removing the undesired feature comprises: applying a threshold to the modified image in which pixels in the modified image having values greater than the threshold are retained to form a thresholded image;convolving the modified image with a point spread function for the image to form a convolved image, wherein the convolved image indicates a first location for the undesired feature in the convolved image that corresponds to a second location for the undesired feature in the image; andsubtracting the convolved image from the image to form a processed image, wherein a portion of the image at the second location for the undesired feature in the image is removed from the image to remove the undesired feature. 2. The method of claim 1, wherein the image is of a scene and wherein the undesired feature in the image is a bright spot caused by a glint in the scene. 3. The method of claim 1, wherein converting the image from the spatial domain to the frequency domain to form the transformed image comprises: computing a Fourier transform for the image that converts the image from the spatial domain to the transformed image in the frequency domain. 4. The method of claim 3, wherein applying the filter to the transformed image to form the filtered image in the frequency domain comprises: changing an amplitude component for the Fourier transform of the image without changing a phase component for the Fourier transform of the image to form the filtered image. 5. The method of claim 4, wherein changing the amplitude component for the Fourier transform of the image without changing the phase component for the Fourier transform of the image to form the filtered image comprises: scaling amplitudes for the amplitude component for the Fourier transform of the image without affecting phases in the phase component for the Fourier transform using a root filter to form the filtered image, wherein the root filter scales the amplitudes in a non-linear manner. 6. The method of claim 1, wherein converting the filtered image from the frequency domain back into the spatial domain to form the modified image comprises: computing an inverse Fourier transform for the filtered image that converts the filtered image from the frequency domain to the spatial domain, wherein the intensity of the undesired feature in the modified image is increased as compared to the intensity of the undesired feature in the image. 7. The method of claim 1 further comprising: identifying characteristics of an object in the image using the processed image. 8. The method of claim 1 further comprising: identifying an object in the image using the processed image. 9. The method of claim 1 further comprising: determining whether an inconsistency is present in an object in the image using the processed image. 10. The method of claim 9, wherein determining whether the inconsistency is present in the object using the processed image comprises: comparing the processed image with another image of the object to form a comparison; anddetermining whether the inconsistency is present using the comparison. 11. An apparatus comprising: a computer system configured to convert an image from a spatial domain to a frequency domain to form a transformed image; apply a filter to the transformed image to form a filtered image in the frequency domain; convert the filtered image from the frequency domain back into the spatial domain to form a modified image, wherein an intensity of an undesired feature in the modified image is increased as compared to the intensity of the undesired feature in the image; and remove the undesired feature from the image, wherein in being configured to remove the undesired feature, the computer system is configured to apply a threshold to the modified image in which pixels in the modified image having values greater than the threshold are retained to form a thresholded image; convolve the modified image with a point spread function for the image to form a convolved image, wherein the convolved image indicates a first location for the undesired feature in the convolved image that corresponds to a second location for the undesired feature in the image; and subtract the convolved image from the image to form a processed image, wherein a portion of the image at the second location for the undesired feature in the image is removed from the image to remove the undesired feature. 12. The apparatus of claim 11 further comprising: a sensor system configured to generate the image of a scene, wherein the undesired feature in the image is a bright spot in the image caused by a glint in the scene. 13. The apparatus of claim 11, wherein in being configured to convert the image from the spatial domain to the frequency domain to form the transformed image, the computer system is configured to compute a Fourier transform for the image that converts the image from the spatial domain to the transformed image in the frequency domain. 14. The apparatus of claim 13, wherein in being configured to apply the filter to the transformed image to form the filtered image in the frequency domain, the computer system is configured to scale amplitudes for an amplitude component for the Fourier transform of the image without affecting phases in a phase component for the Fourier transform using a root filter to form the filtered image, wherein the root filter scales the amplitudes in a non-linear manner. 15. A vehicle inspection system comprising: a sensor system configured to generate images of a vehicle;an image processor configured to receive an image from the sensor system; convert the image from a spatial domain to a frequency domain to form a transformed image; apply a filter to the transformed image to form a filtered image in the frequency domain; convert the filtered image from the frequency domain back into the spatial domain to form a modified image, wherein an intensity of an undesired feature in the modified image is increased as compared to the intensity of the undesired feature in the image; and remove the undesired feature from the image wherein in being configured to remove the undesired feature from the image using the modified image to form the processed image, a computer system is configured to apply a threshold to the modified image in which pixels in the modified image having values greater than the threshold are retained to form a thresholded image; convolve the modified image with a point spread function for the image to form a convolved image, wherein the convolved image indicates a first location for the undesired feature in the convolved image that corresponds to a second location for the undesired feature in the image; and subtract the convolved image from the image to form the processed image, wherein a portion of the image at the second location for the undesired feature in the image is removed from the image to remove the undesired feature; andan image analyzer configured to analyze the processed image to determine whether an inconsistency is present in the vehicle. 16. The vehicle inspection system of claim 15, wherein in being configured to analyze the processed image to determine whether the inconsistency is present, the image analyzer is configured to compare the processed image with another image of the vehicle to form a comparison; and determine whether the inconsistency is present using the comparison. 17. The vehicle inspection system of claim 15 further comprising: a display system, wherein the image analyzer is configured to display an indication of whether the inconsistency is present in the vehicle on the display system. 18. The vehicle inspection system of claim 15, wherein the image analyzer is further configured to generate a report indicating whether the inconsistency is present in the vehicle.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (6)
Clarke Donald A. (Windsor CAX) Reynolds Rodger (Windsor CAX) Pryor Timothy R. (Tecumseh CAX), Diffractosight improvements.
Frederick G. Hewitt ; Steven J. Orfield, Methods for assessing visual tasks to establish desirable lighting and viewing conditions for performance of tasks, apparatus; and, applications.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.