IMAGE LEARNING DEVICE, RADAR DEVICE, IMAGE LEARNING METHOD AND PROGRAM

• Main Authors: ADACHI KAZUTO, YAMAZAKI YOSHIAKI, SEKIYA KENICHI, SUKEGAWA HIROSHI, SHIRAI TOMOJI, MIKAMI SHIGERU, YOKOI KENTARO
• Format: Patent
• Language: English; Japanese
• Published: 25.05.2022
• Subjects: ANALOGOUS ARRANGEMENTS USING OTHER WAVES; CALCULATING; COMPUTING; COUNTING; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION ORRERADIATION OF RADIO WAVES; MEASURING; PHYSICS; RADIO DIRECTION-FINDING; RADIO NAVIGATION; TESTING

To efficiently execute deep learning in order to identify a target from an SAR image or an ISAR image.SOLUTION: An image learning device comprises: an image processing unit (15C) which inputs radar image data subjected to FFT processing (S01), performs IFFT processing on the radar image data (S02), increases the data number by generating data having different setting conditions of the image from the data subjected to the IFFT processing (S03), obtains the radar image data by performing FFT processing on the increased data (S04), and increases the data number by generating the radar image data having the different setting conditions of the image from the increased radar image data (S05); and a learning unit (15B) which inputs the increased radar image data together with type identification information indicating the target in each image and recognizes the target from the radar image to execute deep learning of deriving the type identification information.SELECTED DRAWING: Figure 2 【課題】ＳＡＲ画像またはＩＳＡＲ画像から目標物を識別するための深層学習を効率よく実行すること。【解決手段】ＦＦＴ処理されたレーダ画像データを入力し(Ｓ０１)、レーダ画像データをＩＦＦＴ処理し(Ｓ０２)、ＩＦＦＴ処理したデータから画像の設定条件が異なるデータを生成してデータ数を増加させ(Ｓ０３)、増加させたデータをＦＦＴ処理してレーダ画像データを得(Ｓ０４)、増加させたレーダ画像データから画像の設定条件が異なるレーダ画像データを生成してデータ数を増加させる(Ｓ０５)画像加工部(１５Ｃ)と、増加させたレーダ画像データをそれぞれ画像中の目標物を示す類識別情報と共に入力し、レーダ画像から目標物を認識してその類識別情報を導出する深層学習を実行する学習部(１５Ｂ)とを備える。【選択図】 図２