Acta Photonica Sinica, Volume. 53, Issue 2, 0210001(2024)

Dual Optical Target Recognition Method for Collimated Images Based on BLOB Region and Edge Feature Analysis

Wenxuan HE1,2, Zhengzhou WANG1、*, Jitong WEI1, Li WANG1, and Dongchi YI1
Author Affiliations
  • 1Xi'an Institute of Optics and Precision Mechanics,Chinese Academy of Sciences,Xi'an 710119,China
  • 2University of Chinese Academy of Sciences,Beijing 100039,China
  • show less
    Figures & Tables(26)
    Dual optical target recognition data processing flow
    Original image
    Binary image
    Results of the first digital morphological processing
    Large expansion calculation results of the main laser target
    The detailed parameters of two BLOB regions with the largest areas
    Schematic diagram of target recognition process
    Target recognition results
    Original images
    Preprocessed images
    Detect template edge image results from the left
    Detect the result of the edge image to be recognized from the left
    Searching for the best matching sequence in the edge image to be recognized
    Judging whether the image to be recognized is adhesive image from four directions
    Region segmentation effect of adhesive image
    Optical target recognition results for 5 different butt collimation images
    Identification process and results of a single valid target
    Identification process and results of different adhesive images
    • Table 1. First and second BLOB statistics comparison of BLOB details

      View table
      View in Article

      Table 1. First and second BLOB statistics comparison of BLOB details

      No.AreaCenterLenx/LenyRegion
      FirstSecondFirstSecondFirstSecond*First XSecond Y
      110 05310 857(140,154)(139,155)(132,126)(135,129)*(4,274)(26,284)
      24 0065 964(424,272)(425,274)(95,86)(101,89)*(324,526)(184,362)
      3122356(454,302)(416,215)(19,14)(43,20)(373,459)(194,234)
      4109170(424,211)(425,222)(23,7)(23,9(402,448)(213,231)
      570106(425,222)(421,322)(17,6)(16,8(404,436)(313,329)
      63380(421,321)(449,326)(10,5)(15,8)(433,463)(318,334)
      72779(403,221)(484,271)(9,7)(10,9)(474,494)(261,279)
      81528(484,270)(365,273)(4,6)(8,5)(356,372)(268,278)
      9823(449,325)(395,312)(6,3)(6,4)(389,401)(307,315)
    • Table 2. Comparison of target recognition results

      View table
      View in Article

      Table 2. Comparison of target recognition results

      No.AreaCenterLenx/Lenymng_flagiRegionXY
      110 857(139,155)(135,129)3(4,26)(274,284)(4,274)(26,284)
      25 964(425,274)(101,89)3(324,526)(184,362)(324,526)(184 362)
      3356(416,215)/0///
      4170(425,222)/0///
      5106(421,322)/0///
      680(449,326)/0///
      779(484,271)/0///
      828(365,273)/0///
      923(395,312)/0///
    • Table 3. Decision process of target recognition

      View table
      View in Article

      Table 3. Decision process of target recognition

      No.CenterRegionIs the center in a region?Blobcount statisticsRecognition flagRecognition results
      No.1No.2No.1No.2
      1(139,155)4 274][26 284]YesNo101Analog laser*
      2(425,274)[324 526][184 362]NoYes012Main laser*
      3(416,215)[374 195][460 235]NoYes01//
      4(425,222)[403 214][246 232]NoYes01//
      5(421,322)[405 314][338 330]NoYes01//
      6(449,326)[434 441][464 335]NoYes01//
      7(484,271)[483 262][494 280]NoYes01//
      8(365,273)[357 268][373 278]NoYes01//
      9(395,312)[389 308][318 316]NoYes01//
      Total////1**//
    • Table 4. Four directions to determine whether the image to be recognized is an adhered image

      View table
      View in Article

      Table 4. Four directions to determine whether the image to be recognized is an adhered image

      No.Detect directionTemplate edge imageBest match serialIs adhesive image
      Start posLenStart posCorr coefficientjoinleft…joindown
      1LeftY=107129Y=820.999 01Yes
      2RightX=107129X=990.924 60
      3UpY=118134Y=680.988 51
      4DownX=118107X=680.991 31
    • Table 5. Optical target recognition results for 5 different butt collimation images

      View table
      View in Article

      Table 5. Optical target recognition results for 5 different butt collimation images

      No.Two largest BLOBTwo largest BLOB centersFull BLOBCenter in blobcount with two regionsRecognition result
      No.1No.2No.1No.2

      No.1

      blobcount1

      No.2

      blobcount2

      No.1No.2
      111 2341 690(239,235)(473,242)615Analog laserMain laser
      214 0551 203(165,305)(378,246)624Analog laserMain laser
      310 8575 964(139,155)(425,274)918Analog laserMain laser
      414 3042 012(431,152)(186,423)312Analog laserMain laser
      56 0311 990(425,274)(195,424)1091Main laserAnalog light
    • Table 6. Processing time performance analysis of multi-optical target recognition algorithm

      View table
      View in Article

      Table 6. Processing time performance analysis of multi-optical target recognition algorithm

      No.Image nameBinarizationCorrosionExpansionFeature extractionTarget separationTime
      1Image1.bmp0.0160.2340.2190.0780.1090.656
      2Image2.bmp0.0940.0630.1870.0470.0940.485
      3Image3.bmp0.0940.1410.2180.0940.1560.703
      4Image4.bmp0.0160.0630.1870.0470.0310.344
      5Image5.bmp0.0160.0630.2810.0940.1560.610
      Mean/0.047 20.112 80.218 40.070 50.109 20.559 6
    • Table 7. Single valid target recognition process parameters

      View table
      View in Article

      Table 7. Single valid target recognition process parameters

      Correlation coefficientjoin_flags in 4 directionsjoin flagsingle flagRecognize result
      Adhesion image

      rleft=0.999 0

      rright=0.924 6

      rup=0.988 5

      rdown=0.991 3

      joinleft=1

      joinright=0

      joinup=1

      joindown=1

      11Single adhesion image
      Analog laser image

      rleft=0.999 3

      rright=0.999 0

      rup=0.999 7

      rdown=0.995 2

      joinleft=1

      joinright=1

      joinup=1

      joindown=1

      02Single main light target
      Master laser image

      rleft=0.566 2

      rright=0.818 9

      rup=0.789 2

      rdown=0.887 7

      joinleft=0

      joinright=0

      joinup=0

      joindown=0

      03Single analog laser target
    • Table 8. Recognition process parameters of different adhesive images

      View table
      View in Article

      Table 8. Recognition process parameters of different adhesive images

      Correlation coefficientjoin_flags in 4 directionsIs adhesive image
      Image1.bmp

      rleft=0.999 3

      rright=0.903 9

      rup=0.855 1

      rdown=0.994 6

      joinleft=1

      joinright=0

      joinup=0

      joindown=1

      Yes
      Image2.bmp

      rleft=0.999 5

      rright=0.926 2

      rup=0.990 2

      rdown=0.963 9

      joinleft=1

      joinright=0

      joinup=1

      joindown=1

      Yes
      Image3.bmp

      rleft=0.999 8

      rright=0.922 1

      rup=0.910 4

      rdown=0.930 9

      joinleft=1

      joinright=0

      joinup=0

      joindown=0

      Yes
      Image4.bmp

      rleft=0.967 1

      rright=0.990 0

      rup=0.999 7

      rdown=0.922 4

      joinleft=1

      joinright=1

      joinup=1

      joindown=0

      Yes
      Image5.bmp

      rleft=0.966 4

      rright=0.989 4

      rup=0.999 7

      rdown=0.919 5

      joinleft=1

      joinright=1

      joinup=1

      joindown=0

      Yes
    Tools

    Get Citation

    Copy Citation Text

    Wenxuan HE, Zhengzhou WANG, Jitong WEI, Li WANG, Dongchi YI. Dual Optical Target Recognition Method for Collimated Images Based on BLOB Region and Edge Feature Analysis[J]. Acta Photonica Sinica, 2024, 53(2): 0210001

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category:

    Received: Jun. 13, 2023

    Accepted: Aug. 15, 2023

    Published Online: Mar. 28, 2024

    The Author Email: WANG Zhengzhou (azhou_china@126.com)

    DOI:10.3788/gzxb20245302.0210001

    Topics