JPS60195686A - Circuit for extracting feature of picture - Google Patents

Circuit for extracting feature of picture

Info

Publication number
JPS60195686A
JPS60195686A JP59050919A JP5091984A JPS60195686A JP S60195686 A JPS60195686 A JP S60195686A JP 59050919 A JP59050919 A JP 59050919A JP 5091984 A JP5091984 A JP 5091984A JP S60195686 A JPS60195686 A JP S60195686A
Authority
JP
Japan
Prior art keywords
picture
image
smoothing
feature
noises
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
JP59050919A
Other languages
Japanese (ja)
Inventor
Yuriko Ogawa
小川　優理子
Koji Kameshima
亀島　鉱二
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hitachi Ltd
Original Assignee
Hitachi Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hitachi Ltd filed Critical Hitachi Ltd
Priority to JP59050919A priority Critical patent/JPS60195686A/en
Publication of JPS60195686A publication Critical patent/JPS60195686A/en
Pending legal-status Critical Current

Links

Classifications

    • D—TEXTILES; PAPER
    • D05—SEWING; EMBROIDERING; TUFTING
    • D05B—SEWING
    • D05B39/00—Workpiece carriers

Landscapes

  • Engineering & Computer Science (AREA)
  • Textile Engineering (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

PURPOSE:To obtain a device which extracts a feature pattern, from which noises are removed, at a high speed, by extracting the feature point of a taken picture in such a way that isolated points in the picture are removed by performing a differentiating process and smoothing process on the picture and the feature point is removed after the picture is stored. CONSTITUTION:An input device 1 takes the picture of the corner section A in a building and its real picture image fo becomes the differentiated picture fk which is a candidate for a feature point containing noises by a differentiation arithmetic means 2. The differentiated picture fk becomes a differentiated picture fk', in which dark parts and light parts are emphasized at a light-dark converting means 3 and outputted through a binary coding means 7 after repeatedly subjected to adding operations with the output fa of a smoothness operating means 5 at an adder. Since the smoothing process acts on isolated points of noises, etc., stronger than line segments, isolated points of noises, etc., are successively removed when the smoothing process is repeated. Therefore, the picture fm' obtained after performing the above-mentioned processes becomes on which is formed by fetching feature line segment information only.

Description

【発明の詳細な説明】 〔発明の利用分野〕 本発明は、ノイズを含む画像の特徴抽出装置に係り、特
に孤立点状のノイズを除去し、鮮明なエツジ分布を抽出
するための画像の特徴抽出装置に関するものである。[Detailed Description of the Invention] [Field of Application of the Invention] The present invention relates to a feature extraction device for an image containing noise, and in particular to an image feature extraction device for removing isolated point noise and extracting a sharp edge distribution. This relates to an extraction device.

〔発明の背景〕[Background of the invention]

近年、生産工場、ロボット等には、人間の眼の働きに相
当する視覚手段によって物体を認識する技術が種々採用
されつつあ、る。そしてこの物体認識においては、画像
情報の処理が重要となっている。この画像情報の処理の
基本的問題は画像データから明暗の急変点、すなわち特
徴点をめることである。例えば、・画像の特徴点を抽出
し、その特徴点の配置をめることができれば、方向、距
離などの３次元的な物体配置に関する知識を得ることが
できる。このとき特徴点は、それの属する集合である特
徴パターンとの関連ではじめて決定されることに注意を
要する。すなわち、特徴点を抽出するためには、画像の
広い領域にわたる処理を行う必要がある。したがって、
この種の画像情報処理においては、一つの特徴パターン
を抽出するために多大な計算を行う必要がある。In recent years, various technologies for recognizing objects using visual means equivalent to the function of the human eye have been adopted in production factories, robots, and the like. In this object recognition, processing of image information is important. The basic problem in processing this image information is to find sudden points of change in brightness, ie, feature points, from image data. For example, if feature points of an image can be extracted and the arrangement of the feature points determined, knowledge regarding three-dimensional object arrangement such as direction and distance can be obtained. At this time, it should be noted that the feature points are determined only in relation to the feature pattern to which they belong. That is, in order to extract feature points, it is necessary to perform processing over a wide area of the image. therefore,
In this type of image information processing, it is necessary to perform a large amount of calculation in order to extract one feature pattern.

近年では上述の画像処理の高速化が要求されているが、
上述の処理方策では並列演算が不可能なため、処理の高
速化が図れないものであった。In recent years, there has been a demand for faster image processing as described above.
With the above-mentioned processing strategy, parallel calculations are not possible, so it is not possible to speed up the processing.

〔発明の目的〕[Purpose of the invention]

本発明の目的は、ノイズを除去した特徴パターンを高速
で抽出する画像の特徴抽出装置を提供することにある。An object of the present invention is to provide an image feature extraction device that extracts feature patterns from which noise has been removed at high speed.

〔発明の概要〕[Summary of the invention]

本発明は上記の目的を達成するために、撮影画像の特徴
点を抽出する装置において、撮影画像を微分処理する微
分演算手段と、この手段からの微分画像を平滑化処理し
てその画像中の孤立点を除去する平滑化演算手段と、こ
の手段からの画像を記憶すると共にこの画像を前記平滑
演算手段に出力する記憶手段とを備え、画像中の孤立的
なノイズを除去して特徴点を抽出するようにしたもので
ある。In order to achieve the above object, the present invention provides an apparatus for extracting feature points from a photographed image, which includes a differential calculation means for performing differential processing on the photographed image, and a differential image from the means for smoothing. A smoothing calculation means for removing isolated points, and a storage means for storing an image from this means and outputting this image to the smoothing calculation means, removes isolated noise in the image and extracts feature points. It was designed to be extracted.

〔発明の実施例〕[Embodiments of the invention]

以下本発明の実施例を図面を参照して説明する。 Embodiments of the present invention will be described below with reference to the drawings.

第１図は本発明の装置の一実施例を示すもので、この図
において、ビジコンカメラ等の入力手段１は例えば建屋
内の隅部等の実環境Ａを撮影しその実画像ｆ。を出力す
る。微分演算手段２は実画像ｆ、を微分し、ノイズを含
む特徴点候補の微分画像＜ｋを得る。明暗変換手段３は
特徴点候補の微分画像ｆｋの明暗度を強調した微分画像
ｆｋ′を得るもので、例えば変換テーブルをもつ変換器
で構成される。加算器４は微分演算手段２からの微分画
像ｆｋ′と平滑演算手段５からの出力値ｆａとの画素毎
の加算を行ない、ノイズを含む２値画像ｆ８を得る。メ
モリ６は加算器４からの２値画像ｆｅを記憶する。この
メモリ６に記憶された２値画像ｆ８は平滑演算手段５お
よび２値化手段７に出力される。この平滑演算手段５に
よる平滑化処理は画像の全画面にわたって実行される。FIG. 1 shows an embodiment of the apparatus of the present invention. In this figure, an input means 1 such as a vidicon camera photographs a real environment A, such as a corner of a building, and generates a real image f. Output. The differential calculation means 2 differentiates the real image f, and obtains a differential image <k of feature point candidates including noise. The brightness conversion means 3 obtains a differential image fk' in which the brightness of the differential image fk of the feature point candidate is emphasized, and is composed of, for example, a converter having a conversion table. The adder 4 performs pixel-by-pixel addition of the differential image fk' from the differential calculation means 2 and the output value fa from the smoothing calculation means 5 to obtain a binary image f8 including noise. The memory 6 stores the binary image fe from the adder 4. The binary image f8 stored in the memory 6 is output to the smoothing calculation means 5 and the binarization means 7. This smoothing process by the smoothing calculation means 5 is executed over the entire screen of the image.

これにより得られた特徴抽出画像ｆｌｌｌを逐次２値化
手段に出力する。平滑化処理はノイズの多い画像に対し
その近傍の傍で各画素を平均化するものであり、これに
より孤立的な画像ノイズがその近傍の影響を受け、除去
傾向に処理される。この２値化手段７は特徴抽出画像ｆ
Ｉｌｌ′　を出力する。The feature extraction image fllll thus obtained is sequentially output to the binarization means. Smoothing processing averages each pixel in the vicinity of a noisy image, and as a result, isolated image noise is influenced by its vicinity and tends to be removed. This binarization means 7 uses the feature extraction image f
Output Ill'.

次に上述した本発明の装置の実施例における画像の特徴
抽出方法を第２図に示すフローチャートおよび第３図〜
第６図に示す処理画像を用いて説明する。Next, a flowchart showing the image feature extraction method in the embodiment of the apparatus of the present invention described above is shown in FIG. 2, and FIGS.
This will be explained using the processed image shown in FIG.

なお、第３図〜第６図には処理画像と共にその水平中間
部（Ｘ−Ｘ）における明暗度も併せて示しである。In addition, FIGS. 3 to 6 also show the brightness at the horizontal intermediate portion (X-X) of the processed image.

いま、第１図に示すように、入力装置１が建屋内の隅部
Ａの状況を撮影すると、その実画像ｆ。Now, as shown in FIG. 1, when the input device 1 photographs the situation at a corner A in the building, an actual image f is obtained.

は第３図に示すように平面的な画像として得られる。こ
の実画像ｆ。は微分演算手段２によって第４図に示すよ
うにノイズを含む特徴点候補の微分画像ｆｋとなる。次
にこの微分画像ｆｋは明暗変換手段３により第５図に示
すように明暗が強調されたノイズを含む微分画像ｆＦｒ
′となる。この微分画像ｆヒ′はさらに加算器４におい
て平滑演算手段５からの出力ｆａとの加算操作のくり返
しにより、２値化手段７を通して出力される。これによ
り得られる画像ｆ１１′は第６図に示すように特徴線分
情報のみを取り出した画像となる。前述した加算操作す
なわち平滑化処理は全画面にわたって実行される。この
平滑化処理は線分よりもノイズ等の孤立点に強く作用す
るので、前述したように平滑化処理の縁り返しにより、
第５図に示す画像ｆｋ′中のノイズ等の孤立点は逐次除
去される。is obtained as a planar image as shown in FIG. This real image f. The differential calculation means 2 produces a differential image fk of the feature point candidate including noise as shown in FIG. Next, this differential image fk is converted into a differential image fFr containing noise with enhanced brightness and darkness by the brightness conversion means 3 as shown in FIG.
'. This differential image fhi' is further repeatedly added to the output fa from the smoothing calculation means 5 in the adder 4, and is outputted through the binarization means 7. The image f11' thus obtained is an image in which only characteristic line segment information is extracted, as shown in FIG. The above-described addition operation, ie, the smoothing process, is performed over the entire screen. This smoothing process has a stronger effect on isolated points such as noise than on line segments, so as mentioned above, by turning the edges of the smoothing process,
Isolated points such as noise in the image fk' shown in FIG. 5 are successively removed.

このため、２値化手段７はその閾値設定値に関する依存
度を高くする必要がないものである。また、このため、
場合によっては、２値化手段７は必ずしも必要としない
ことがある。Therefore, the binarization means 7 does not need to be highly dependent on its threshold setting value. Also, for this reason,
In some cases, the binarization means 7 may not necessarily be necessary.

以上述べたように、上述の本発明の実施例によれば、ま
ず局所的な演算により画像の特徴点の候補を検出し、そ
の特徴点の分布状況に応じて真の特徴点のみを抽出する
という２段階の処理を行うが、後者の処理を局所平滑化
演算の繰り返しにより実行し得るので、並列演算器を用
いることにより高速処理が可能であると共に、画像中の
特徴部分を高速で抽出することができる。As described above, according to the embodiment of the present invention described above, candidates for feature points in an image are first detected by local calculation, and only true feature points are extracted according to the distribution status of the feature points. The latter process can be performed by repeating local smoothing operations, so high-speed processing is possible by using parallel computing units, and feature parts in images can be extracted at high speed. be able to.

〔発明の効果〕〔Effect of the invention〕

以上述べたように、本発明によれば、撮影画像の特徴点
抽出を高速で得ることができ、その処理の高速化を図る
ことができる。As described above, according to the present invention, it is possible to extract feature points from a photographed image at high speed, and the processing speed can be increased.

【図面の簡単な説明】[Brief explanation of the drawing]

第１図は本発明の装置の一実施例の構成を示す。 第２図は本発明の画像処理動作を示すフローチャート図
、第３図〜第６図は本発明の装置によって得られる処理
画像の一例を示す図である。 ｌ・・・入力装置、２・・・微分演算手段、３・・・明
暗変換手段、４・・・加算器、５・・平滑演算手段、６
・・・メモリ、７・・・２値化手段。 聾１因 子２０ ギ３ｚ　＄４０ 子Ｓ図　＄乙口FIG. 1 shows the configuration of an embodiment of the apparatus of the present invention. FIG. 2 is a flowchart showing the image processing operation of the present invention, and FIGS. 3 to 6 are diagrams showing examples of processed images obtained by the apparatus of the present invention. l... Input device, 2... Differential calculation means, 3... Brightness conversion means, 4... Adder, 5... Smoothing calculation means, 6
...Memory, 7...Binarization means. Deafness 1 factor 20 Gi 3z ＄40 Child S figure ＄Otsuguchi

Claims (1)

【特許請求の範囲】 １、撮影画像の特徴点を抽出する装置において、撮影画
像を微分処理する微分演算手段と、この手段からの微分
画像を平滑化処理してその画像中の孤立点を除去する平
滑演算手段と、この手段からの画像を記憶すると共にこ
の画像を前記平滑演算手段に出力する記憶手段とを備え
たことを特徴とする画像の特徴抽出装置。 ２、平滑演算手段は微分画像の全域にわたって平滑化処
理を繰り返し実行することを特徴とする特許請求の範囲
第１項記載の画像の特徴抽出装置。 ３、平滑演算手段は並列演算器で構成したことを特徴と
する特許請求の範囲第２項記載の画像の特徴抽出装置。[Scope of Claims] 1. A device for extracting feature points from a photographed image, comprising a differential calculation means for performing differential processing on the photographed image, and a smoothing process for the differential image from this means to remove isolated points in the image. 1. An image feature extraction device comprising: a smoothing calculation means for performing smoothing; and a storage means for storing an image from the smoothing calculation means and outputting the image to the smoothing calculation means. 2. The image feature extraction device according to claim 1, wherein the smoothing calculation means repeatedly executes the smoothing process over the entire area of the differential image. 3. The image feature extracting device according to claim 2, wherein the smoothing calculation means is constituted by a parallel calculation unit.
JP59050919A 1984-03-19 1984-03-19 Circuit for extracting feature of picture Pending JPS60195686A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP59050919A JPS60195686A (en) 1984-03-19 1984-03-19 Circuit for extracting feature of picture

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP59050919A JPS60195686A (en) 1984-03-19 1984-03-19 Circuit for extracting feature of picture

Publications (1)

Publication Number Publication Date
JPS60195686A true JPS60195686A (en) 1985-10-04

Family

ID=12872198

Family Applications (1)

Application Number Title Priority Date Filing Date
JP59050919A Pending JPS60195686A (en) 1984-03-19 1984-03-19 Circuit for extracting feature of picture

Country Status (1)

Country Link
JP (1) JPS60195686A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4821659A (en) * 1986-12-01 1989-04-18 Brother Kogyo Kabushiki Kaisha Workpiece holding device for a sewing machine

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4821659A (en) * 1986-12-01 1989-04-18 Brother Kogyo Kabushiki Kaisha Workpiece holding device for a sewing machine

Similar Documents

Publication Publication Date Title
CN109325954B (en) Image segmentation method and device and electronic equipment
Sánchez-Nielsen et al. Hand gesture recognition for human-machine interaction
CN110059642B (en) Face image screening method and device
CN111583097A (en) Image processing method, image processing device, electronic equipment and computer readable storage medium
Goel et al. Specific color detection in images using RGB modelling in MATLAB
CN113191216B (en) Multi-user real-time action recognition method and system based on posture recognition and C3D network
JP7403562B2 (en) Method for generating slap/finger foreground masks
Arsic et al. Improved lip detection algorithm based on region segmentation and edge detection
CN112839167A (en) Image processing method, image processing device, electronic equipment and computer readable medium
CN111222355A (en) Method and system for positioning bar code on PCB
JPS60195686A (en) Circuit for extracting feature of picture
Varkonyi-Koczy Fuzzy logic supported corner detection
CN112907553A (en) High-definition image target detection method based on Yolov3
CN114581448B (en) Image detection method, device, terminal equipment and storage medium
JP2960468B2 (en) Method and apparatus for binarizing grayscale image
Murali et al. Speech Processing and Recognition of Sign Language using Tensor flow Classifier
Metkar et al. A MODERN ERA IN IMAGE PROCESSING
JP2003173436A (en) Image processing device, image processing method, and recording medium recording program readably by computer
JPH0322181A (en) Contour extracting method with variable threshold value
Nishad et al. Portable Data Compiler
CN111783790A (en) Hand-drawn picture preprocessing method and device
JPH03269782A (en) Character area extracting method for character reader
JPH0652355A (en) Character segmenting and character recognizing device
JPS61117678A (en) Average density calculating circuit
JPS63155274A (en) Zero-cross detection system for image