Search

CN-121982060-A - Self-adaptive infrared detection tracking method, system, equipment and storage medium

CN121982060ACN 121982060 ACN121982060 ACN 121982060ACN-121982060-A

Abstract

The invention discloses a self-adaptive infrared detection tracking method, a system, equipment and a storage medium based on discrete cosine transform, wherein the method comprises the following steps of inputting an image, and performing two-dimensional discrete cosine transform on the input image to obtain a frequency domain signal; the method comprises the steps of carrying out frequency domain filtering on an obtained frequency domain signal, then carrying out inverse discrete cosine transform to obtain a saliency map, carrying out edge extraction on the saliency map to obtain a target contour, carrying out statistics on gray scale distribution in the target contour to obtain a mean value and a standard deviation in a target area, setting a mean value and a standard deviation threshold value, carrying out threshold segmentation on an image in the target area, segmenting all pixels of the target, and obtaining a centroid of the segmented target pixels for detection tracking of the target. The invention discloses a discrete cosine transform-based self-adaptive infrared detection tracking method, a discrete cosine transform-based self-adaptive infrared detection tracking system, discrete cosine transform-based self-adaptive infrared detection tracking equipment and a storage medium, which solve the problem that parameter self-adaptation is difficult to realize in the infrared detection tracking process, and improve the detection tracking robustness.

Inventors

  • Jiang Tongyuan
  • Zhu Shilun
  • JIANG YUANSONG
  • XIE YALI
  • LU JIANG
  • LI XIUJING
  • XIA PING

Assignees

  • 华中光电技术研究所(中国船舶集团有限公司第七一七研究所)

Dates

Publication Date
20260505
Application Date
20251224

Claims (10)

  1. 1. The self-adaptive infrared detection tracking method is characterized by comprising the following steps of: Step 1, inputting an image, and performing two-dimensional discrete cosine transform on the input image to obtain a frequency domain signal; step 2, performing frequency domain filtering on the obtained frequency domain signal, and then performing inverse discrete cosine transform to obtain a saliency map; step 3, extracting edges of the saliency map to obtain a target contour; step4, counting gray distribution in the target contour to obtain a mean value and a standard deviation in the target area; step 5, setting a mean value and a standard deviation threshold value, and carrying out threshold segmentation on the image in the target area to segment all pixels of the target; and 6, calculating the mass center of the segmented target pixels for detection and tracking of the target.
  2. 2. The method of claim 1, wherein the two-dimensional discrete cosine transform is performed by performing a one-dimensional discrete cosine transform in a column direction and then performing a one-dimensional discrete cosine transform in a row direction on the input image.
  3. 3. The method of claim 1, wherein in step 2, a sign (Q) function is used for frequency domain filtering.
  4. 4. The method of claim 1, wherein in step 2, the two-dimensional inverse discrete cosine transform is performed by performing one-dimensional inverse discrete cosine transform in a column direction and then performing one-dimensional inverse discrete cosine transform in a row direction.
  5. 5. The method of claim 1, wherein in step 3, a soble algorithm is used to perform edge extraction on the saliency map to obtain a target contour.
  6. 6. The method of claim 1, wherein in step 5, the mean value and standard deviation of the pixel statistics in the contour are set as a threshold value to perform threshold segmentation, and the decision of the mean value and standard deviation being greater than the threshold value is regarded as the target, otherwise the decision is regarded as the background.
  7. 7. The method of claim 1, wherein in step 6, the centroid calculation method is as follows: Wherein X ij ,Y ij is the coordinate value of the image line direction, and I ij is the pixel value of the image coordinate (I, j) position.
  8. 8. An adaptive infrared detection tracking system for implementing the infrared detection tracking method of any one of claims 1-7, said system comprising: The two-dimensional discrete cosine transform module receives an external input image, performs two-dimensional discrete cosine transform on the input image to obtain a frequency domain signal, and outputs the frequency domain signal to the frequency domain filtering module; The frequency domain filtering module performs frequency domain filtering on the input frequency domain signal, performs inverse discrete cosine transform to obtain a saliency map, and outputs the saliency map to the edge extraction module; The edge extraction module is used for extracting edges of the obtained saliency map to obtain a target contour and outputting the target contour to the threshold calculation module; The threshold calculation module is used for carrying out gray distribution statistics on the input target contour, acquiring the mean value and standard deviation in the target area and outputting an image segmentation module; The image segmentation module is used for constructing a threshold value according to the input mean value and standard deviation, carrying out threshold segmentation on the image in the target area, and outputting the segmented target pixels to the centroid calculation module; And the centroid calculation module is used for calculating the centroid of the input target pixel and outputting the result for target detection tracking.
  9. 9. An electronic device comprising a memory for storing a computer program and a processor for loading the computer program in the memory, characterized in that the operation of the adaptive infrared detection tracking system of claim 8 is enabled when the electronic device is in operation.
  10. 10. A computer scale storage medium having a computer program stored thereon, wherein the program, when executed by a processor, is capable of performing the operation of the adaptive infrared detection tracking system of claim 8.

Description

Self-adaptive infrared detection tracking method, system, equipment and storage medium Technical Field The invention belongs to the field of image processing, and particularly relates to a discrete cosine transform-based self-adaptive infrared detection tracking method, a system, equipment and a storage medium. Background The infrared image detection and tracking utilizes the infrared radiation difference between the target and the background in the infrared image to form a thermal image, thereby realizing the detection and tracking of the target. With the maturation of uncooled detector technology and the development of algorithms, this technology is moving towards higher performance, lower cost and wider application. Infrared target detection based on discrete cosine transform is an effective frequency domain processing method, which transforms an image from a spatial domain to a frequency domain, and detects the target by utilizing the difference between the target and a background in the energy distribution of the frequency domain. Disclosure of Invention In order to solve the technical problems, the invention provides a discrete cosine transform-based adaptive infrared detection tracking method, a discrete cosine transform-based adaptive infrared detection tracking system, discrete cosine transform-based adaptive infrared detection tracking equipment and a discrete cosine transform-based storage medium, so that the problem that parameter adaptation is difficult to realize in the infrared detection tracking process is solved, and the detection tracking robustness is improved. The invention aims at realizing the following technical scheme, and discloses a self-adaptive infrared detection tracking method, which comprises the following steps of: Step 1, inputting an image, and performing two-dimensional discrete cosine transform on the input image to obtain a frequency domain signal; step 2, performing frequency domain filtering on the obtained frequency domain signal, and then performing inverse discrete cosine transform to obtain a saliency map; step 3, extracting edges of the saliency map to obtain a target contour; step4, counting gray distribution in the target contour to obtain a mean value and a standard deviation in the target area; step 5, setting a mean value and a standard deviation threshold value, and carrying out threshold segmentation on the image in the target area to segment all pixels of the target; and 6, calculating the mass center of the segmented target pixels for detection and tracking of the target. Preferably, in step 1, the two-dimensional discrete cosine transform is performed by first performing one-dimensional discrete cosine transform in a column direction and then performing one-dimensional discrete cosine transform in a row direction on the input image. Preferably, in step 2, a sign (Q) function is used for frequency domain filtering. Preferably, in the step 2, the two-dimensional discrete cosine inverse transformation is performed by performing one-dimensional discrete cosine inverse transformation in the column direction and then performing one-dimensional discrete cosine inverse transformation in the row direction. Preferably, in step3, edge extraction is performed on the saliency map by adopting soble algorithm to obtain the target contour. Preferably, in step 5, the mean value+standard deviation of the pixel statistics in the contour is set as a threshold value to perform threshold segmentation, and the decision that the mean value+standard deviation is greater than the threshold value is regarded as the target, otherwise the decision is regarded as the background. Preferably, in step 6, the centroid calculation method is as follows: Wherein X ij,Yij is the coordinate value of the image line direction, and I ij is the pixel value of the image coordinate (I, j) position. The invention further provides an infrared detection tracking system for realizing the method, besides providing a method for adapting to infrared detection tracking, the system comprises: The two-dimensional discrete cosine transform module receives an external input image, performs two-dimensional discrete cosine transform on the input image to obtain a frequency domain signal, and outputs the frequency domain signal to the frequency domain filtering module; The frequency domain filtering module performs frequency domain filtering on the input frequency domain signal, performs inverse discrete cosine transform to obtain a saliency map, and outputs the saliency map to the edge extraction module; The edge extraction module is used for extracting edges of the obtained saliency map to obtain a target contour and outputting the target contour to the threshold calculation module; The threshold calculation module is used for carrying out gray distribution statistics on the input target contour, acquiring the mean value and standard deviation in the target area and outputting an image segmentation module; The