WO2024029370A1 - Method for labeling particulate matter in optical sorter - Google Patents
Method for labeling particulate matter in optical sorter Download PDFInfo
- Publication number
- WO2024029370A1 WO2024029370A1 PCT/JP2023/026765 JP2023026765W WO2024029370A1 WO 2024029370 A1 WO2024029370 A1 WO 2024029370A1 JP 2023026765 W JP2023026765 W JP 2023026765W WO 2024029370 A1 WO2024029370 A1 WO 2024029370A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- sorting
- particulate matter
- labeling
- quality determination
- optical sorter
- Prior art date
Links
- 238000002372 labelling Methods 0.000 title claims abstract description 68
- 238000000034 method Methods 0.000 title claims abstract description 51
- 239000013618 particulate matter Substances 0.000 title claims abstract description 38
- 230000003287 optical effect Effects 0.000 title claims abstract description 27
- 238000003384 imaging method Methods 0.000 claims abstract description 30
- 238000004364 calculation method Methods 0.000 claims abstract description 18
- 239000008187 granular material Substances 0.000 claims description 27
- 238000002347 injection Methods 0.000 claims description 11
- 239000007924 injection Substances 0.000 claims description 11
- 239000011236 particulate material Substances 0.000 claims description 6
- 230000002950 deficient Effects 0.000 abstract description 2
- 238000001303 quality assessment method Methods 0.000 abstract 1
- 235000013339 cereals Nutrition 0.000 description 11
- 238000010586 diagram Methods 0.000 description 8
- 239000002245 particle Substances 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 7
- 238000007796 conventional method Methods 0.000 description 3
- 235000021329 brown rice Nutrition 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 239000008188 pellet Substances 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/85—Investigating moving fluids or granular solids
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
Definitions
- the present invention relates to a method for labeling granular materials in an optical sorter, which can improve the sorting speed and accuracy of granular materials such as grains and pellets.
- Patent Document 1 discloses a labeling processing method that can perform real-time processing on digital images transmitted from a television camera.
- the present invention provides a labeling method for particulate matter in an optical sorting machine, which enables high-speed and high-accuracy determination of pass/fail of a plurality of particulate matter transferred at high speed in an optical sorting machine.
- the purpose is to provide a method.
- the present invention includes an imaging step in which a plurality of granules transferred in an optical sorter are imaged by an imaging means, and an image taken in the imaging step is stored in a line memory one line image at a time.
- a labeling processing step in which a labeling process is performed using pipeline processing using the granular data
- a feature amount calculation step in which a feature amount is calculated based on the extracted data in the labeling processing step;
- a method for labeling particulate matter in an optical sorter comprising: a step of determining whether the product is good or bad.
- the labeling process can be performed at high speed, and the quality of the pass/fail judgment can be improved. Results and selection signals based on them can be quickly output.
- the distance between the installation position of the CCD camera, which is the imaging means, and the air injection position of the ejector, which is the sorting means can be brought closer. This makes it possible to reduce the chances that the vehicle will deviate from the transport route. As a result, it becomes possible to sort out particulate matter at high speed and with high precision.
- the sorting speed remains unchanged, and production capacity and sorting quality are improved compared to the conventional method.
- FIG. 2 is a flow diagram illustrating a labeling method in the prior art. It is a flow diagram explaining the labeling method in the present invention.
- FIG. 4 is a diagram for comparing and explaining the processing time of the conventional technology and the present invention. It is a flow diagram explaining pipeline processing in an embodiment of the present invention.
- FIG. 1 shows a schematic flow diagram of a method for labeling particulates in the prior art.
- a particulate matter is imaged by an imaging means such as a CCD camera (step S101)
- a frame image is stored in the image memory in units of frames (fixed number of lines) (step S102).
- labeling processing is executed frame by frame (step S103), and the labeling data is stored in the image memory again (step S104).
- step S104 the feature amount of the granule is calculated for each frame based on the labeling data (step S104) (step S105), and the quality of the granule is determined based on the calculated feature amount (step S106). .
- FIG. 2 shows a schematic flow diagram of the method for labeling particulates according to the present invention.
- an imaging step (S1) in which a particulate object is imaged by an imaging means including a line sensor, and the images captured in the imaging step (S1) are pipelined one line at a time using a line memory.
- a labeling processing step (S2) in which a labeling process is performed by processing, a feature amount calculation step (S3) in which a feature amount is calculated based on the extracted data in the labeling processing step (S2), and a feature amount calculation step (S3). It has at least a quality determination step (S4) of determining the quality of the granular material based on the calculation result.
- the process moves to a sorting step in which granules are sorted based on the quality determination result in the quality determination step (S4), and an ejector serving as a sorting means is used.
- the granules will be sorted out.
- the present invention is characterized in that one line image captured by an imaging means such as a line sensor is pipeline-processed one line at a time without being stored in an image memory.
- labeling processing is performed for each line image for particulate number 1, and when the labeling of the final line image of particulate number 1 is completed, the feature amount (for example, feature quantities related to color, shape, and quality) are calculated. Then, based on the calculation result of the feature amount, a comprehensive judgment is made as to whether the particulate material is good or bad.
- the same series of processing is performed on the illustrated granules No. 2 and 3 as well.
- FIG. 3 shows a schematic diagram comparing the processing times of the prior art and the present invention.
- the flow rate of particulates in an optical sorter increases from a small amount (for example, several tens of grains) to a large amount (for example, several hundred grains)
- the filling rate of the particulates within one frame increases, so the labeling process
- the time, feature value calculation time, and pass/fail judgment time become longer. Therefore, the processing time per frame increases depending on the amount of falling particles (detected amount).
- the labeling process, feature value calculation, and pass/fail determination are performed after one frame of image is collected, the time required to make the pass/fail determination also increases.
- the labeling process is performed by pipeline processing using the line memory for each line image without storing the captured image in the image memory. Therefore, irrespective of whether the amount of flowing down (detected amount) of the granules is small or large, it is possible to quickly pass through each processing step and determine the quality of the granules. That is, the time from when the image of the granular material is finished being imaged by the imaging means until the quality determination is made can be extremely shortened, and it is possible to keep the processing time constant regardless of the number of particles.
- FIG. 4 shows a schematic flow diagram of an embodiment of pipeline processing of the present invention. This embodiment is implemented by a labeling circuit and has a line memory with a predetermined circuit configuration.
- images captured by a CCD camera for example, a line sensor
- a lookup table in the labeling circuit Processed as a ring buffer (temporary storage). Therefore, unlike the conventional method described above, there is no need to wait for one frame image to be collected, and there is no need to temporarily store the image in the image memory and then erase it. That is, immediately after imaging, labeling processing is performed by pipeline processing for each line image.
- labeling is performed in units of one frame, so if the number of particles in one frame exceeds the number of labels in the lookup table, processing becomes impossible. In that case, it is necessary to increase the size of the lookup table.
- a label different from the original label is assigned to a particle that spans between frames, a process is required to determine whether it is the same particle or a different particle.
- labeling processing is performed in units of one line image, so as long as the number of grains arranged horizontally on one line at the same time does not exceed the number of labels in the lookup table, the labeling process cannot be performed correctly. It is possible. Therefore, it is possible to reduce the size of the lookup table without increasing it.
- the labels are not interrupted, and it can be regarded as an infinitely large lookup table.
- one lookup table is configured for the temporary label and one for the true label in order to correspond to the speed of the line sensor (clock frequency of one pixel). If the speed of the line sensor is slow, it is also possible to label with one lookup table. Note that if the line memory component (for example, a rewritable IC) has a high speed, temporary label processing and true label processing can be performed using a single lookup table even if the line sensor speed is high. You can also.
- the present invention is applicable to various types of granules such as grains and pellets.
- the imaging means is not necessarily limited to a line sensor, and an area sensor may also be used.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The present invention is characterized by having: an imaging step in which an imaging means is used to capture an image of a plurality of particulate matter being transferred in an optical sorter; a labeling process step in which an image captured in the imaging step is subjected to a labeling process by means of a pipeline process using a line memory, one line image at a time; a feature amount calculation step in which a feature amount is calculated on the basis of data extracted in the labeling process step; and a quality assessment step for assessing whether the particulate matter is good or defective, on the basis of the calculation result in the feature amount calculation step.
Description
本発明は、穀粒やペレットなどの粒状物に対する選別速度や選別精度の向上を図ることが可能な、光学式選別機における粒状物のラベリング方法に関する。
The present invention relates to a method for labeling granular materials in an optical sorter, which can improve the sorting speed and accuracy of granular materials such as grains and pellets.
従来から、被検査物の外観をカメラで撮像し、撮像した画像をデジタル処理することにより、被検査物が不良品か否かを判別する方法が知られている。また、被検査物の良否の判別に際しては、撮像した画像から連結領域を１つのまとまりとして認識して抽出するラベリング処理が、外観検査装置でも一般的に行われている。
Conventionally, there has been known a method of determining whether or not the inspected object is a defective product by capturing an image of the outer appearance of the inspected object with a camera and digitally processing the captured image. Furthermore, when determining the quality of an inspected object, a labeling process is generally performed in visual inspection apparatuses, in which connected regions are recognized and extracted as one group from a captured image.
例えば、特許文献１には、テレビカメラから伝送されるデジタル画像に対してリアルタイム処理が可能なラベリング処理方法が開示されている。
For example, Patent Document 1 discloses a labeling processing method that can perform real-time processing on digital images transmitted from a television camera.
しかしながら、特許文献１に開示されているような従来技術にあっては、フレーム（一定ライン数）単位で処理するため、ラベリング処理の開始が１フレーム収集後になり、ラベリング処理に時間を要していた。また、ラベリング処理に時間を要する他の要因として、ラベリング処理を行うために、ルックアップテーブルの最適化と真ラベリングを、画像メモリへの記憶（保存）とリセット（消去）との繰り返し処理で行っていることが挙げられる。さらに、一つの粒状物がフレームをまたぐ場合もある。この場合、異なるラベル番号がそれぞれのフレームにラベリングされてしまうが、それらは同じ粒状物であると認識させる必要がある。このラベリング処理は、非常に煩雑な処理を要することがあった。
However, in the conventional technology disclosed in Patent Document 1, since processing is performed in units of frames (fixed number of lines), the labeling process starts after one frame has been collected, and the labeling process takes time. Ta. Another reason why the labeling process takes time is that in order to perform the labeling process, lookup table optimization and true labeling are repeated by storing (save) and reset (erase) in the image memory. One example is that Furthermore, a single particle may straddle the frame. In this case, each frame is labeled with a different label number, but it is necessary to recognize that they are the same particulate matter. This labeling process sometimes requires very complicated processing.
また、穀物の選別機においては、穀物の撮像位置とエジェクタのエアー噴射位置を近づけて、選別精度を高めたいというニーズがあるものの、ラベリング処理に時間を要する従来技術では、このようなニーズに対応することが困難であった。
Additionally, in grain sorting machines, there is a need to increase sorting accuracy by bringing the grain imaging position closer to the ejector's air injection position, but conventional technology that requires time for labeling processing cannot meet these needs. It was difficult to do so.
そこで、本発明はこのような課題に鑑み、光学式選別機において高速で移送される複数の粒状物を、高速かつ高精度に良否判定することが可能な、光学式選別機における粒状物のラベリング方法を提供することを目的とする。
Therefore, in view of such problems, the present invention provides a labeling method for particulate matter in an optical sorting machine, which enables high-speed and high-accuracy determination of pass/fail of a plurality of particulate matter transferred at high speed in an optical sorting machine. The purpose is to provide a method.
本発明は上述の課題を解決するために、光学式選別機において移送される複数の粒状物を撮像手段によって撮像する撮像工程と、前記撮像工程によって撮像した画像を、１ライン画像ずつラインメモリを使用してパイプライン処理によってラベリング処理を行うラベリング処理工程と、前記ラベリング処理工程における抽出データに基づいて特徴量を算出する特徴量算出工程と、前記特徴量算出工程の算出結果に基づいて前記粒状物の良否を判定する良否判定工程と、を有することを特徴とする光学式選別機における粒状物のラベリング方法を提供する。
In order to solve the above-mentioned problems, the present invention includes an imaging step in which a plurality of granules transferred in an optical sorter are imaged by an imaging means, and an image taken in the imaging step is stored in a line memory one line image at a time. a labeling processing step in which a labeling process is performed using pipeline processing using the granular data; a feature amount calculation step in which a feature amount is calculated based on the extracted data in the labeling processing step; Provided is a method for labeling particulate matter in an optical sorter, comprising: a step of determining whether the product is good or bad.
本発明によれば、従来技術のように撮像した画像を画像メモリに一旦保存して再び読み出す時間が生じず、１ライン画像ずつラベリング処理するので、高速でラベリング処理することができ、良否判定の結果とそれに基づく選別信号を素早く出力することができる。このようなラベリング処理の高速化により、撮像手段であるＣＣＤカメラの設置位置と、選別手段であるエジェクタのエアー噴射位置との距離をより近づけることができるので、撮像位置から選別位置までに粒状物が移送経路から外れる機会を減らすことが可能となる。ひいては、高速かつ高精度に粒状物を選別することが可能となる。しかも、粒状物の流量を増やした場合でも、選別速度が変わることがなく、生産能力と選別品質が従来よりも向上する。
According to the present invention, there is no time required to temporarily store a captured image in an image memory and then read it out again as in the conventional technology, and the labeling process is performed one line at a time. Therefore, the labeling process can be performed at high speed, and the quality of the pass/fail judgment can be improved. Results and selection signals based on them can be quickly output. By speeding up the labeling process, the distance between the installation position of the CCD camera, which is the imaging means, and the air injection position of the ejector, which is the sorting means, can be brought closer. This makes it possible to reduce the chances that the vehicle will deviate from the transport route. As a result, it becomes possible to sort out particulate matter at high speed and with high precision. Moreover, even when the flow rate of granular material is increased, the sorting speed remains unchanged, and production capacity and sorting quality are improved compared to the conventional method.
以下に、図面を参照して本発明の光学式選別機における粒状物のラベリング方法について、穀粒をラインセンサによって撮像し、当該穀粒の良否判定を行う例について説明する。
Below, with reference to the drawings, an example of a method for labeling grains in an optical sorter of the present invention in which grains are imaged by a line sensor and quality determination of the grains is determined will be described.
図１には、従来技術における粒状物のラベリング方法が概略フロー図で示されている。図１において、ＣＣＤカメラなどの撮像手段によって粒状物が撮像されると（ステップＳ１０１）、フレーム（一定ライン数）単位で画像メモリにフレーム画像が保存される（ステップＳ１０２）。そして１フレームの画像が収集されると、フレーム単位でラベリング処理が実行され（ステップＳ１０３）、再び画像メモリにラベリングデータが保存される（ステップＳ１０４）。
FIG. 1 shows a schematic flow diagram of a method for labeling particulates in the prior art. In FIG. 1, when a particulate matter is imaged by an imaging means such as a CCD camera (step S101), a frame image is stored in the image memory in units of frames (fixed number of lines) (step S102). When one frame of images is collected, labeling processing is executed frame by frame (step S103), and the labeling data is stored in the image memory again (step S104).
その後、ラベリングデータ（ステップＳ１０４）に基づいて、粒状物の特徴量が１フレームごとに算出され（ステップＳ１０５）、算出された当該特徴量に基づいて粒状物の良否判定が行われる（ステップＳ１０６）。これら一連の処理は演算回路によって実行している。
Thereafter, the feature amount of the granule is calculated for each frame based on the labeling data (step S104) (step S105), and the quality of the granule is determined based on the calculated feature amount (step S106). . These series of processes are executed by an arithmetic circuit.
図２には本発明における粒状物のラベリング方法が概略フロー図で示されている。具体的には、ラインセンサ等を含む撮像手段によって、粒状物を撮像する撮像工程（Ｓ１）と、当該撮像工程（Ｓ１）によって撮像した画像を、１ライン画像ずつラインメモリを使用してパイプライン処理によってラベリング処理を行うラベリング処理工程（Ｓ２）と、当該ラベリング処理工程（Ｓ２）における抽出データに基づいて特徴量を算出する特徴量算出工程（Ｓ３）と、当該特徴量算出工程（Ｓ３）の算出結果に基づいて粒状物の良否を判定する良否判定工程（Ｓ４）とを少なくとも有している。なお、図示していないが、良否判定工程（Ｓ４）の後には、当該良否判定工程（Ｓ４）における良否の判定結果に基づいて、粒状物を選別する選別工程に移行し、選別手段となるエジェクタによって粒状物は選別されることとなる。
FIG. 2 shows a schematic flow diagram of the method for labeling particulates according to the present invention. Specifically, there is an imaging step (S1) in which a particulate object is imaged by an imaging means including a line sensor, and the images captured in the imaging step (S1) are pipelined one line at a time using a line memory. A labeling processing step (S2) in which a labeling process is performed by processing, a feature amount calculation step (S3) in which a feature amount is calculated based on the extracted data in the labeling processing step (S2), and a feature amount calculation step (S3). It has at least a quality determination step (S4) of determining the quality of the granular material based on the calculation result. Although not shown in the figure, after the quality determination step (S4), the process moves to a sorting step in which granules are sorted based on the quality determination result in the quality determination step (S4), and an ejector serving as a sorting means is used. The granules will be sorted out.
より詳細に説明すると、ラインセンサ等の撮像手段で撮像した１ライン画像を、画像メモリに保存することなく、１ライン画像ずつパイプライン処理することを特徴としている。図２に示される実施形態では、例えば、粒状物番号１について１ライン画像ずつラベリング処理を行い、粒状物番号１の最終ライン画像のラベリングが終了した段階で、当該粒状物番号１の特徴量（例えば、色や形状、品質に関する特徴量）が算出される。そして、当該特徴量の算出結果に基づいて、粒状物に対する良否の総合判定が行われる。このような一連の処理は、図示される粒状物番号２及び粒状物番号３に対しても同様の処理が行われる。
To explain in more detail, the present invention is characterized in that one line image captured by an imaging means such as a line sensor is pipeline-processed one line at a time without being stored in an image memory. In the embodiment shown in FIG. 2, for example, labeling processing is performed for each line image for particulate number 1, and when the labeling of the final line image of particulate number 1 is completed, the feature amount ( For example, feature quantities related to color, shape, and quality) are calculated. Then, based on the calculation result of the feature amount, a comprehensive judgment is made as to whether the particulate material is good or bad. The same series of processing is performed on the illustrated granules No. 2 and 3 as well.
より理解を深めるために、図３には、従来技術と本発明の処理時間を比較する概略の説明図が示されている。従来技術では、光学式選別機における粒状物の流量が少量（例えば、数十粒）から多量（例えば、数百粒）になるほど、１フレーム内における粒状物の充填率が高くなるため、ラベリング処理時間や特徴量算出時間、良否判定時間が長くなる。したがって、粒状物の流下量（検出量）に応じて１フレームあたりの処理時間が長くなる。加えて、従来技術では、１フレームの画像が収集されてからラベリング処理及び特徴量算出、良否判定を行うため、良否判定までにかかる時間も長くなる。
For better understanding, FIG. 3 shows a schematic diagram comparing the processing times of the prior art and the present invention. In conventional technology, as the flow rate of particulates in an optical sorter increases from a small amount (for example, several tens of grains) to a large amount (for example, several hundred grains), the filling rate of the particulates within one frame increases, so the labeling process The time, feature value calculation time, and pass/fail judgment time become longer. Therefore, the processing time per frame increases depending on the amount of falling particles (detected amount). In addition, in the conventional technology, since the labeling process, feature value calculation, and pass/fail determination are performed after one frame of image is collected, the time required to make the pass/fail determination also increases.
一方、本発明では、撮像した画像を画像メモリに保存することなく、１ライン画像ずつラインメモリを使用してパイプライン処理によってラベリング処理を行う。したがって、粒状物の流下量（検出量）が少量であるか多量であるかに拘わらず、素早く各処理工程を経て粒状物の良否判定を実行することが可能となる。すなわち、粒状物を撮像手段で撮像し終わってから良否判定するまでの時間を非常に短くすることができ、粒数によらずに処理時間を一定にすることが可能となる。
On the other hand, in the present invention, the labeling process is performed by pipeline processing using the line memory for each line image without storing the captured image in the image memory. Therefore, irrespective of whether the amount of flowing down (detected amount) of the granules is small or large, it is possible to quickly pass through each processing step and determine the quality of the granules. That is, the time from when the image of the granular material is finished being imaged by the imaging means until the quality determination is made can be extremely shortened, and it is possible to keep the processing time constant regardless of the number of particles.
図４には、本発明のパイプライン処理における実施例が概略のフロー図で示されている。本実施例はラベリング回路によって実行され、所定の回路構成によってラインメモリを有している。
FIG. 4 shows a schematic flow diagram of an embodiment of pipeline processing of the present invention. This embodiment is implemented by a labeling circuit and has a line memory with a predetermined circuit configuration.
図４に示されるように、ラベリング回路のパイプライン処理でラベル作成を行うため、ＣＣＤカメラ（例えば、ラインセンサ）で撮像した画像を、ラベリング回路内のルックアップテーブルを使い、１ライン画像ずつ、リングバッファ（一時保存）として処理する。したがって、前述した従来法のように、１フレーム画像の収集を待つことなく、画像メモリへ画像を一旦記憶させて消去する処理を行うこともない。つまり撮像した直後に、１ライン画像ずつパイプライン処理によってラベリング処理を行っている。
As shown in FIG. 4, in order to create labels through pipeline processing in the labeling circuit, images captured by a CCD camera (for example, a line sensor) are processed one line at a time using a lookup table in the labeling circuit. Processed as a ring buffer (temporary storage). Therefore, unlike the conventional method described above, there is no need to wait for one frame image to be collected, and there is no need to temporarily store the image in the image memory and then erase it. That is, immediately after imaging, labeling processing is performed by pipeline processing for each line image.
より詳細に説明すると、従来技術では、１フレーム単位でラベリングを実施しているので、１フレーム内の粒状物の粒数がルックアップテーブルのラベル数を越えると処理できなくなる。そうなると、ルックアップテーブルのサイズを大きくする必要がある。加えて、フレーム間をまたぐ粒状物に本来のラベルとは異なるラベルが割り当てられてしまうため、同一の粒状物なのか別の粒状物なのかを判定する処理が必要になる。一方、本発明では、１ライン画像の単位でラベリング処理を行うので、１ライン上に横一列に同時に並ぶ粒状物の粒数が、ルックアップテーブルのラベル数を越えない限りは正しく処理することが可能である。したがって、ルックアップテーブルのサイズを大きくすることなく、むしろ小さくすることが可能となる。さらに、上記したように、本実施例ではルックアップテーブルをリングバッファにすることで、ラベルが途切れることが無く、無限に大きなルックアップテーブルとみなすことができる。
To explain in more detail, in the conventional technology, labeling is performed in units of one frame, so if the number of particles in one frame exceeds the number of labels in the lookup table, processing becomes impossible. In that case, it is necessary to increase the size of the lookup table. In addition, since a label different from the original label is assigned to a particle that spans between frames, a process is required to determine whether it is the same particle or a different particle. On the other hand, in the present invention, labeling processing is performed in units of one line image, so as long as the number of grains arranged horizontally on one line at the same time does not exceed the number of labels in the lookup table, the labeling process cannot be performed correctly. It is possible. Therefore, it is possible to reduce the size of the lookup table without increasing it. Furthermore, as described above, in this embodiment, by using a ring buffer as the lookup table, the labels are not interrupted, and it can be regarded as an infinitely large lookup table.
このような本実施例によれば、ラベリング処理の高速化が実現できるので、撮像手段であるＣＣＤカメラの設置位置と、選別手段であるエジェクタのエアー噴射位置との距離をより近づけることが可能となり、精度の高い粒状物の選別が可能となる。加えて、１ライン画像ずつラベリング処理することで、粒状物がフレームをまたぐような状態が発生しない。したがって、従来法のような煩雑な処理が不要になり、処理の高速化に寄与している。なお、上記エジェクタとして、バルブの開閉速度が速いピエゾバルブを使用することにより、粒状物の選別精度をさらに高めることが可能となる。
According to this embodiment, it is possible to speed up the labeling process, so it is possible to bring the installation position of the CCD camera, which is the imaging means, closer to the air injection position of the ejector, which is the sorting means. , it becomes possible to sort out particulate matter with high precision. In addition, by performing the labeling process on each line image, a situation in which particles straddle frames does not occur. Therefore, the complicated processing required in the conventional method is no longer necessary, contributing to faster processing. Note that by using a piezo valve with a fast valve opening/closing speed as the ejector, it is possible to further improve the accuracy of sorting particulate matter.
また、図４に示す実施例では、ラインセンサの速度（１画素のクロック周波数）に応じるため、ルックアップテーブルを、仮ラベル用に一つ、真ラベル用に一つ、構成している。ラインセンサの速度が遅い場合は一つのルックアップテーブルでラベリングすることも可能である。なお、ラインメモリ部品（例えば、書き換え可能なＩＣ）が高速度化したものであれば、ラインセンサの速度が速くても、一つのルックアップテーブルにより、仮ラベル処理と真ラベル処理を行うようにすることもできる。
Furthermore, in the embodiment shown in FIG. 4, one lookup table is configured for the temporary label and one for the true label in order to correspond to the speed of the line sensor (clock frequency of one pixel). If the speed of the line sensor is slow, it is also possible to label with one lookup table. Note that if the line memory component (for example, a rewritable IC) has a high speed, temporary label processing and true label processing can be performed using a single lookup table even if the line sensor speed is high. You can also.
（他の実施形態）
以上、本発明における実施例について説明したが、必ずしも上記実施例に限定されるものではない。 (Other embodiments)
Although the embodiments of the present invention have been described above, the present invention is not necessarily limited to the above embodiments.
以上、本発明における実施例について説明したが、必ずしも上記実施例に限定されるものではない。 (Other embodiments)
Although the embodiments of the present invention have been described above, the present invention is not necessarily limited to the above embodiments.
例えば、図示される実施例では、粒状物として穀粒のラインセンサによる撮像態様を例に示したが、粒状物は穀粒のほかペレットなど、種々の粒状物に対して本発明は適用可能である。また、必ずしも撮像手段はラインセンサに限定されるものではなく、エリアセンサを使用することも可能である。
For example, in the illustrated embodiment, an example is shown in which grains are imaged by a line sensor, but the present invention is applicable to various types of granules such as grains and pellets. be. Further, the imaging means is not necessarily limited to a line sensor, and an area sensor may also be used.
ここまで、本発明の実施態様の構成について、本発明の実施例に対する試験結果も参照しながら説明してきたが、本発明は、図面を参照して説明した上述の実施態様及び実施例に限定されるものではない。添付の請求の範囲及びその要旨を逸脱することなく、様々な変更、置換が可能であり、または上述の実施態様の構成と本質的に同等な構成を含む玄米の処理方法または食用玄米が具現化され得ることは、当業者にとって明らかである。
Up to this point, the configuration of the embodiments of the present invention has been explained with reference to the test results for the examples of the present invention, but the present invention is limited to the above-described embodiments and examples described with reference to the drawings. It's not something you can do. Various changes and substitutions can be made without departing from the scope and gist of the appended claims, or a brown rice processing method or edible brown rice that includes a configuration essentially equivalent to the configuration of the above-described embodiments is embodied. It will be clear to those skilled in the art that this can be done.
Ｓ１ 撮像工程
Ｓ２ ラベリング処理工程
Ｓ３ 特徴量算出工程
Ｓ４ 良否判定工程 S1 Imaging process S2 Labeling process S3 Feature amount calculation process S4 Quality determination process
Ｓ２ ラベリング処理工程
Ｓ３ 特徴量算出工程
Ｓ４ 良否判定工程 S1 Imaging process S2 Labeling process S3 Feature amount calculation process S4 Quality determination process
Claims (17)
- 光学式選別機において移送される複数の粒状物を撮像手段によって撮像する撮像工程と、
前記撮像工程によって撮像した画像を、１ライン画像ずつラインメモリを使用してパイプライン処理によってラベリング処理を行うラベリング処理工程と、
前記ラベリング処理工程における抽出データに基づいて特徴量を算出する特徴量算出工程と、
前記特徴量算出工程の算出結果に基づいて前記粒状物の良否を判定する良否判定工程と、を有する
ことを特徴とする光学式選別機における粒状物のラベリング方法。 an imaging step of imaging a plurality of granules transferred in the optical sorter using an imaging means;
a labeling processing step of labeling the image captured in the imaging step by pipeline processing using a line memory one line image at a time;
a feature amount calculation step of calculating a feature amount based on the extracted data in the labeling processing step;
A method for labeling particulate matter in an optical sorter, comprising: a quality determining step of determining whether the particulate material is good or bad based on the calculation result of the feature value calculation step. - 前記特徴量算出工程では、１ライン画像ごとに前記ラベリング処理されたデータに基づいて特徴量を算出する
請求項１に記載の光学式選別機における粒状物のラベリング方法。 The method for labeling particulate matter in an optical sorter according to claim 1, wherein in the feature amount calculation step, a feature amount is calculated for each line image based on the labeled data. - 前記良否判定工程では、１ライン画像ごとに前記ラベリング処理されたデータに基づく特徴量の算出結果に基づいて前記粒状物の良否を判定する
請求項１又は２に記載の光学式選別機における粒状物のラベリング方法。 The quality of the particulate matter in the optical sorter according to claim 1 or 2, wherein in the quality determination step, the quality of the particulate matter is determined based on a calculation result of a feature amount based on the labeled data for each line image. labeling method. - 前記良否判定工程では、前記粒状物の画像の最終ラインにおいて前記特徴量の算出結果に基づく良否の総合判定を行う
請求項１又は２に記載の光学式選別機における粒状物のラベリング方法。 3. The method of labeling a particulate material in an optical sorter according to claim 1, wherein in the quality determination step, a comprehensive determination of quality is performed based on the calculation result of the feature amount in the last line of the image of the particulate material. - 前記良否判定工程では、前記粒状物の画像の最終ラインにおいて前記特徴量の算出結果に基づく良否の総合判定を行う
請求項３に記載の光学式選別機における粒状物のラベリング方法。 4. The method of labeling a particulate material in an optical sorter according to claim 3, wherein in the quality determination step, a comprehensive determination of quality is performed based on the calculation result of the feature amount in the last line of the image of the particulate material. - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物を選別するエジェクタとしてピエゾバルブを使用する
請求項１又は２に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 1 or 2, wherein in the sorting step, a piezo valve is used as an ejector for sorting the particulate matter. - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物を選別するエジェクタとしてピエゾバルブを使用する
請求項３に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 3, wherein in the sorting step, a piezo valve is used as an ejector for sorting the particulate matter. - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物を選別するエジェクタとしてピエゾバルブを使用する
請求項４に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 4, wherein in the sorting step, a piezo valve is used as an ejector for sorting the particulate matter. - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物を選別するエジェクタとしてピエゾバルブを使用する
請求項５に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 5, wherein in the sorting step, a piezo valve is used as an ejector for sorting the particulate matter. - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物の選別精度を維持しつつエジェクタのエアー噴射位置と前記撮像手段との距離をより近づけることが可能である
請求項１又は２に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
In the optical sorting machine according to claim 1 or 2, in the sorting step, it is possible to make the distance between the air injection position of the ejector and the imaging means closer while maintaining the accuracy of sorting the particulate matter. Labeling method. - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物の選別精度を維持しつつエジェクタのエアー噴射位置と前記撮像手段との距離をより近づけることが可能である
請求項３に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 3, wherein in the sorting step, the distance between the air injection position of the ejector and the imaging means can be made closer while maintaining the accuracy of sorting the particulate matter. . - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物の選別精度を維持しつつエジェクタのエアー噴射位置と前記撮像手段との距離をより近づけることが可能である
請求項４に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 4, wherein in the sorting step, the distance between the air injection position of the ejector and the imaging means can be made closer while maintaining the accuracy of sorting the particulate matter. . - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物の選別精度を維持しつつエジェクタのエアー噴射位置と前記撮像手段との距離をより近づけることが可能である
請求項５に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 5, wherein in the sorting step, it is possible to bring the distance between the air injection position of the ejector and the imaging means closer while maintaining the accuracy of sorting the particulate matter. . - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物の選別精度を維持しつつエジェクタのエアー噴射位置と前記撮像手段との距離をより近づけることが可能である
請求項６に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 6, wherein in the sorting step, the distance between the air injection position of the ejector and the imaging means can be made closer while maintaining the accuracy of sorting the particulate matter. . - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物の選別精度を維持しつつエジェクタのエアー噴射位置と前記撮像手段との距離をより近づけることが可能である
請求項７に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 7, wherein in the sorting step, it is possible to bring the distance between the air injection position of the ejector and the imaging means closer while maintaining the accuracy of sorting the particulate matter. . - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物の選別精度を維持しつつエジェクタのエアー噴射位置と前記撮像手段との距離をより近づけることが可能である
請求項８に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 8, wherein in the sorting step, it is possible to bring the distance between the air injection position of the ejector and the imaging means closer while maintaining the accuracy of sorting the particulate matter. . - 前記良否判定工程における良否の判定結果に基づいて前記粒状物を選別する選別工程を有し、
前記選別工程では前記粒状物の選別精度を維持しつつエジェクタのエアー噴射位置と前記撮像手段との距離をより近づけることが可能である
請求項９に記載の光学式選別機における粒状物のラベリング方法。 a sorting step of sorting the granular material based on the quality determination result in the quality determination step;
The method for labeling particulate matter in an optical sorter according to claim 9, wherein in the sorting step, it is possible to bring the distance between the air injection position of the ejector and the imaging means closer while maintaining the accuracy of sorting the particulate matter. .
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2022-124953 | 2022-08-04 | ||
| JP2022124953A JP2024021833A (en) | 2022-08-04 | 2022-08-04 | Method for labeling granular material in optical sorter |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2024029370A1 true WO2024029370A1 (en) | 2024-02-08 |
Family
ID=89848881
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2023/026765 WO2024029370A1 (en) | 2022-08-04 | 2023-07-21 | Method for labeling particulate matter in optical sorter |
Country Status (2)
| Country | Link |
|---|---|
| JP (1) | JP2024021833A (en) |
| WO (1) | WO2024029370A1 (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005055245A (en) * | 2003-08-01 | 2005-03-03 | Seirei Ind Co Ltd | Apparatus and method for sorting grain |
| KR20100068815A (en) * | 2008-12-15 | 2010-06-24 | 삼성전기주식회사 | Method for grouping pixels in 2d digital image |
| JP2013178670A (en) * | 2012-02-28 | 2013-09-09 | Canon Inc | Information processor and information processing method |
-
2022
- 2022-08-04 JP JP2022124953A patent/JP2024021833A/en active Pending
-
2023
- 2023-07-21 WO PCT/JP2023/026765 patent/WO2024029370A1/en unknown
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005055245A (en) * | 2003-08-01 | 2005-03-03 | Seirei Ind Co Ltd | Apparatus and method for sorting grain |
| KR20100068815A (en) * | 2008-12-15 | 2010-06-24 | 삼성전기주식회사 | Method for grouping pixels in 2d digital image |
| JP2013178670A (en) * | 2012-02-28 | 2013-09-09 | Canon Inc | Information processor and information processing method |
Non-Patent Citations (1)
| Title |
|---|
| ANONYMOUS: "Developing a new type of valve with ultra-high-speed response - Utilizing piezo elements to improve the performance of optical sorters", NEWS RELEASES/ANNOUNCEMENTS, SATAKE CO., LTD., 18 September 2018 (2018-09-18), XP093136282, Retrieved from the Internet <URL: https://www.satake-japan.co.jp/news/new-release/post-57.html> [retrieved on 20240229] * |
Also Published As
| Publication number | Publication date |
|---|---|
| JP2024021833A (en) | 2024-02-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7248732B2 (en) | Pattern inspection method and inspection apparatus | |
| JP2004077164A (en) | Defect inspection method | |
| JPH04166751A (en) | Method and apparatus for inspecting defect in bottle and the like | |
| CN103297718A (en) | Defective pixel correction apparatus and control method thereof and program for executing the method | |
| JP2006170809A (en) | Device and method for detecting defect | |
| CN112102281B (en) | Truck brake cylinder fault detection method based on improved Faster Rcnn | |
| WO2024029370A1 (en) | Method for labeling particulate matter in optical sorter | |
| CN110108713A (en) | A kind of Superficial Foreign Body defect fast filtering method and system | |
| CN117252864B (en) | Steel production device smoothness detection system based on identification analysis | |
| JP3316400B2 (en) | Container inspection by field programmable array logic | |
| CN115532620A (en) | Pulp molding product quality inspection device and method | |
| JP4408358B2 (en) | Grain sorting device | |
| CN114331944A (en) | Artificial intelligence flaw image classification method and system | |
| JP2007059858A (en) | Chip image inspection method and its system | |
| JP4528509B2 (en) | Grain sorting device | |
| JP2500649B2 (en) | IC foreign matter inspection device | |
| JP3374818B2 (en) | Defect detection apparatus and method | |
| JP2009244058A (en) | Quality inspection device and quality inspection method | |
| CN114453262B (en) | Method, apparatus, storage medium and detection device for detecting device | |
| CN111933543B (en) | Appearance detection method of axial diode | |
| CN114425524B (en) | Material receiving control method and device for detection equipment and detection equipment | |
| JP2756738B2 (en) | Apparatus for visual inspection of semiconductor devices | |
| JPH05157523A (en) | Inspecting method of shape of glass bottle at hot end | |
| JPS62262192A (en) | Mark inspection method | |
| CN116840244A (en) | Software and hardware coordinated packaged chip detection device and method and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23849925 Country of ref document: EP Kind code of ref document: A1 |