JPS58205275A - Picture data processor - Google Patents

Abstract

PURPOSE:To process sequentially inputted data such as picture data on real-time basis by accessing all tracks successively when data on plural pictures to be processed are processed. CONSTITUTION:A CPU114 supplies a disk control system 150 with such as indication that the 1st picture data D1n stored in a track (n) of a disk file 202a is read. A CPU152 controls a disk circuit 204 and moves a read/write head to seek the track (n), transferring and storing the data D1n to and in one bank, e.g. 106A of a buffer memory 106. Then when data D2n on scanning lines of the 2nd picture to be processed corresponding to the track (n) are inputted to a port 102, a CPU114 superposes the data D1n in the bank 106A upon the input data D2n and then stores the resulting data in the bank 106A through the disk control system 150.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a data processing device, and more specifically, to processing sequentially input data such as image data in real time, and processing image data stored in a rotating memory such as a magnetic disk. Processing device K: 1ml.

Generally, for image data such as medical images, facsimile/millimeter images, etc., pixel data representing each pixel is formed from a recorded image and is sequentially input to a processing device at a predetermined transfer rate for processing.

The input of such image data to the processing device is performed in synchronization with the scanning of the recorded image by the image reading device. Normally, the image reading device scans the recorded image in the main scanning direction and the sub-scanning direction at a predetermined scanning speed to form image data, and inputs the image data as a time-series signal to a processing device.

These image reading apparatuses and recording apparatuses cannot interrupt the reading or recording operation in the middle of the process once they start processing, because it is necessary to maintain the quality of the image.

For example, a system that uses optical scanning for main scanning and mechanical scanning for secondary scanning uses components that are economically feasible with current technology due to mechanical inertia of the device, etc. In such a case, it is difficult to form a high-quality image even after repeated sudden start-up or stoppages. Therefore, due to some reason, the 7X force
Placing this in the middle will result in missing input data. Also,
The same applies to the input of such data to the processing device via a data transmission interface. Therefore, interruption of input is usually not allowed.

Some image data processing apparatuses process corresponding image data in multiple frames of images of the same subject in real time and file them as single frame image data. For example, Japanese Patent Application Laid-Open No. 56-11399 describes an X-ray image processing method using a stimulable phosphor, in which X-ray images recorded on multiple stimulable phosphor sheets are read out and multiple frames are processed. A method for superimposing image signals is described. This method generates image data by scanning an X-ray image stored on a stimulable stimulable phosphor sheet by irradiating it with excitation light. The corresponding h-purple data in the i7 image data of multiple frames falsified in this way is subjected to 1 alignment processing, and the synergistic data is filed as a single image data. This removes the noise contained in the image and improves the signal-to-noise ratio (S/N) of the final image. Normally, along with this overlay processing, unsharp mask processing and gradation processing are also performed.

Typically, image data processing devices perform such superposition,
Image data processing for multiple frames (frames), such as subtraction and weighted averaging that emphasizes differences between multiple images, is performed in real time in response to image data input, and the processing result data is used as an image. It is stored as a data file, and output as a visible image from an output device such as an image recording device as needed.

Incidentally, in rotary memories such as magnetic disks and optical disks having a movable radar imaging mechanism, the time required to move the head varies depending on the position of the track to be accessed, so the access time differs. For example, in a 6z disk that rotates at a rotation speed of 20 milliseconds per rotation,
It takes about 10 milliseconds for the head to move between adjacent tracks. Therefore, the further the object is from the track it is currently accessing to the track it will access next, the faster it will take the head to move. It takes time.

For example, in the case of an image data processing device that performs overlapping processing of two images, conventionally, the data of the first image is accumulated in one logical area of the Rui disk file.
In response to the data of the second image being sequentially input to the processing unit as a serial image signal, the processing unit reads the corresponding data of the first image from the disk file;
Real-time processing such as overlapping processing is performed on the first image data and the second image data, and the processing results are stored in separate logical areas of the disk file.

During this processing, the first image data from one of the two-curved areas of the disk file is transferred track by track to the buffer memory in the stem.
z) The image data is stored track by track in the other logical area of the disk file, and is sent to the other logical area of the disk file. Therefore, each time the read/write head of a magnetic disk processes one track of image data, it must go back and forth between one logical area and the other area.

For example, one scanning line in the main scanning direction of an image to be processed includes 2,000s pixels and is scanned in 30 milliseconds. 7
Assuming that the main amount of pixel data can be accumulated, the processing device will transfer one track's worth of data from the magnetic disk within 210 milliseconds, which is the time it takes for seven scanning lines' worth of data to be transferred from an input device such as an image reading device. It must be read, processed, and stored back on the magnetic disk. Therefore, as mentioned above, it took time for the 7-turn operation. S-again,
If it takes time to read/write data onto a magnetic disk, or write the read/write operation (IJ) due to transfer, such an operation and photon may not occur within the above-mentioned data transfer time. In such a case, it is not possible to interrupt the flow of data from the computer, so there is a possibility that the input data will be missing by the amount of data transfer time exceeded.

The present invention aims to eliminate such drawbacks of the prior art and provide an image data processing device that does not cause data loss in real-time processing of image data that is continuously transferred at a predetermined transfer rate. purpose.

Another object of the present invention is to provide an image data processing device that takes a short time to access a rotating memory.

According to the present invention, these objects are achieved by the following image data processing device. That is, this device:
a processing circuit that processes corresponding image data in a plurality of processed images that are input in time series to form resultant image data; and a rotating memory that stores image data.
An image data processing device that includes a control circuit connected between these processing circuits and a rotating memory and controlling the rotating memory, wherein the rotating memory is configured to record a series of data once every year or so.
The processing circuit (rl) includes a buffer memory supporting two banks for temporarily storing image data for one storage unit of a rotating memory, and controls In response to the input of image data of another image to be processed, while the processing circuit is processing the image data of one corresponding storage unit stored in one bank, The image data stored in the bank is written to the storage unit immediately before the one storage unit in the rotating memory, and then the image data stored in the storage unit immediately after the one storage unit is written to the other storage unit. This is what is read out to the bank.

In addition, in this book, one memory unit means rotation! An Il control unit of a storage location to be accessed in a memory, and may include, for example, one or several disks in a magnetic disk.

Next, embodiments of the image data processing apparatus according to the present invention will be explained with reference to the accompanying drawings.

] The figure shows the stroke number data processing device according to the invention, ′, T
)'-1"h is shown in a professional diagram. In the figure, the data processing device has a data processing system 100 that processes a plurality of frames of image data. It has an input/output circuit 110 connected to input/output ports 102 and 104 for input/output.Input port 1 (
An input device (not shown) such as an image reading device is connected to 22, and the image reading device scans the image to be processed in the main scanning direction and the sub-scanning direction, and generates time-series image data corresponding to each pixel. is continuously supplied to the input port 102 at a predetermined data transfer rate.

In one example, one scanning line in the main scanning direction is scanned in 30 milliseconds, and image data corresponding to 2,000 pixels is input to the input, ff=g 102. Output port 10
For example, an output device (not shown) such as an image reading device is connected to 4, and a processed visible image corresponding to the image data is output from the output port 104.
It is played as a copy.

The input/output circuit 110 includes input/output ports 102 and 10.
4, a port circuit including an amplifier, a latch renoster, etc., is connected to the central processing unit (CPU) via bus 112.
) 114K connected. To avoid complicating the diagram, the bus 112 includes a data bus, an address bus, a command bus, and other control lines. This is a processing device that performs data processing such as averaging and hierarchical processing in real time, as well as centrally controlling the operation of the entire device.In a preferred embodiment of the present invention, the CPU
114 is composed of a pit slice type microprocessor.

The bus 112 includes a read-only memory (ROM) 116 that stores control programs for the CPU 114, image data processing programs, fixed data, etc., and a random access memory (RAM) used as high-speed memory.
118, and a buffer memory 106 are connected to the buffer memory 106 (G).As will be described later, two memory banks IC, 6A and 106 each have a capacity that can temporarily store image data for one track of a magnetic disk.
B'i have. In the same figure, the buffer memory 106 is (dR
Although it is shown as if it were a separate unit from the RAM 118, a portion of the area within the RAM 118 may be used, or it may be a separate unit of the same type as the RAM 118.
may be used.

The data processing device shown in FIG. 1 has a disk control system 150, which is a processing device that controls the magnetic disk device 200. The disk control system 150 is a CPU 152
, ROM 154, RAM 156 and input/output circuit 1
58, etc., and these are interconnected via a bus 160. - The gas 160 is also coupled to the buffer memory 106 of the data processing system 100. Input/output circuit 1
58 is connected to the interface unit 206 of the magnetic disk device 20'0 via a data line 162 and a control line 164. The CPU 152 of the disk control system 150 controls the magnetic disk device 200 in accordance with the program sequence stored in the RoM 154 as described later.

A magnetic disk device 200 is shown as a typical example of a rotating memory, and includes a disk unit 202, a disk circuit 204, and an interface section 206. The disk unit 202 is, for example, a disk, a
Data modules, floppy disks, etc. may be used. Disc picking can be advantageously used because a plurality of tracks included in one cylinder can be selected by switching between magnetic and magnetic tracks. The disk circuit 204 includes a magnetic spatula that controls the read/write head and a data read/write control circuit.
The interface section 206 includes a write circuit and the like, and the interface section 206 connects the data line 162 of the input/output circuit 158 of the disk cloth 1j control system 150, and the control line 1 such as an address line and a power supply.
This is a circuit that interfaces with 64.

In this embodiment, image data for seven scanning lines of the image to be processed can be stored in one track of the disk unit 202. According to the present invention, image data corresponding to successive series of scan lines are stored in successive physical sectors of successive physical trunks, excluding unused tracks, unused sectors, nine-domain sectors, etc. Note that in this specification, storage locations of rotating memory such as tracks and sectors are referred to as "a series" or "a series".
Terms such as "continuous" and "successive" refer to physical tracks, excluding unused tracks, unused sectors, etc.
This means that the physical sectors are consecutive. As shown in FIG. 2, in the logical storage area (file) 202a of the disk unit 202, a series of tracks n-2
, n-1, n+n+1. Image data of seven scanning lines is stored in each of n+2 (n is a natural number). The seven scanning lines in each track are continuous in the sub-scanning direction in the image to be processed, and these scanning lines are continuous in the sub-scanning direction between any two adjacent tracks. is retained.

Therefore, by sequentially accessing each track of the disk file 202a, one frame of the processed image can be reproduced.

- As an example, an explanation will be given of an operation in which data of two processed images (frames) of the same subject are superimposed. Note that the operation is the same for other processes such as saputra kungyong. When the first image data is input from the image reading device to the input port 102 at a predetermined data transfer rate, the CPU 114 stores this data in a predetermined storage area 202a of the disk unit 202. Since this data is a time-series signal formed by sequentially scanning images to be processed, the data is sent to the disk file 20 by the CPU 152 in the order in which it is input to the input port 102.
When assigning and writing a series of tracks in 2a,
This is because one frame of processed images is stored in the file 202a in scanning order. In FIG. 2, the first image data is Dl, the second image data is D2, and both data D1 and D2 are processed in data processing 1 in the CPU 114.
The data resulting from the overlapping process according to 14a is represented by D, and the track numbers n-2, n-1, n on which each data is stored are
, n+1, n+2, etc. are indicated by subscripts.

In addition, the process of adjusting 1 to -n i is as follows for the J-th pixel in the first (frame) image: DJ=ΣaIDI
It is represented by J. However, al indicates the weight.

When the first image data is accumulated in the file 202aKi, each track n 2+n 1+n+n+1. +n+
2 has Din-2+ Dlyl-1+ D, respectively.
This depends on the fact that ll +D1n+, , Di□2 was written.

Next, when the second image data is input to the input port 102 in the order of scanning lines at a predetermined transfer rate, the CPU 114 transfers the data to the disk file 20 corresponding to the image data being input.
The second image data is read out from track 2a and superimposition processing is performed on the two pieces of image data. To explain the operation for an arbitrary track n, the CPU 114 first selects the 1 stored in track n of the disk file 202a.
The disk control system 150 is instructed to read out the green monitor D1n of the first image. CPU 152 controls disk circuitry 204 to move a read/write head (not shown) to read trunk n and transfer data D1n to one puncture of buffer anomaly 106, e.g.
Then, data D2n of the scanning line of the second processed image corresponding to track nK is input to the port 102 (302).
The CPU 114 performs superposition processing between the data D1n of the bank 106A and the input data D2n, and stores the resulting data Dn in the puncture 106A via the disk control system 150 (304, 306). Therefore, when the processing for track n is completed, ・tank 106A
All data becomes Dn.

By the way, while the overlapping process is being performed on track n, the disk control system 150 moves the tracks n-11 to soak track n-11, which is one track before n. Next, the data stored in the other link 106B,
After completing the overlapping process for track n-1, the data Dn-+ is transferred to track n-1 and stored therein. When this writing power is finished, the task control 1
1 system 150 (moves C from track n-1 to the next track n+1, and displays the data stored there, that is, the first processed sheet. Statue data D1
Read n+ and transfer it to Kunk 1 6B. Therefore, while the data D1n of the 9-tank 106A is being processed, the data of the Z-tank 6B is
This is because it was rewritten from -1 to Din+1.

Next, when the data of the second image to be processed of the scanning line corresponding to track n+1 is input to the port 102, the CPU
114 reads the first data Dln+ from the other bank 106B on behalf of one bank 106A through the disk control system 150, and processes it with the input data D2n-N. In this way, the same operation as described for the processing of data D1n is performed on
We will also do this for . By repeating the same operation for all tracks of the file 202a, the disk file 202a has one complete frame of processed image data Dn2゜Dn-1+Dl+D
n+i l I) n+2, etc. are accumulated. CPU11
4 reads this data as necessary and outputs it to output port 1.
04 to an output device (not shown) such as an image recording device.

Data of three or more images to be processed can also be processed in the same way. In the above example, disk file 20
2 a K Since the image data for which processing has been completed for the two images has been accumulated, when the third image data is input to the input port 102, it will be processed in the same way as for the second input image data. Let's do the following actions. In this way, it is possible to process any number of image data.

The image data processing device according to the present invention sequentially stores time-series data formed by sequentially scanning an image to be processed in a physically continuous series of tracks excluding unused tracks,
When processing the data of multiple images to be processed, all tracks must be accessed by performing a 7-track operation on the magnetic disk between a series of 3 tracks on the magnetic disk. The system is configured to perform the steps sequentially.

Therefore, since the amount of movement of the input data in the soak operation is small, the soak time is short, and the access time of the magnetic disk is compared to the transfer time of the input data of the corresponding track input at a predetermined transfer rate. is quite short.

This reduced minute time can be used to control other magnetic disks,
For example, it can be assigned to a retry operation due to a read/write error. This kind of time margin makes it possible to avoid a situation where the input image data overflows from the buffer memory and is lost due to the delay in accessing the magnetic disk, and also to avoid the occurrence of a situation where the input image data is lost due to the delay in accessing the magnetic disk. By providing a buffer memory with a storage capacity, it is possible to buffer the difference between the transfer speed of input image data and the access time of the magnetic disk.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of an image data processing apparatus according to the present invention, and FIG. 2 is an explanatory diagram for explaining the operation of the first apparatus. Explanation of symbols for formal parts 100...Data processing system 106...Buffer memory 106A, 106B-Memory bank 114.152...Central processing unit 150...Disk control system 202...Disk unit 204...Disk circuit l゛”

Claims (1)

[Scope of Claims] 1. A processing circuit that processes corresponding image data divided into a plurality of processed images that are input in time series to form resulting image data; A rotating memory that stores image data; An image data processing device including a control circuit connected between the processing circuit and a rotating memory and controlling the rotating memory, wherein the rotating memory stores one processed image in a series of successive storage units. Sequential image data is stored, the processing circuit includes a buffer memory having two banks each temporarily storing image data for one storage unit of the rotary memory, and 11. the Mii control circuit; The @Fc processing circuit responds to the input IC of the image data of another image processing image and performs processing with the image data of one corresponding storage unit stored in the memory unit of 10,000. Then, the image data stored in the other bank is stored in the storage unit immediately before the one storage unit in the rotary memory, and then stored in the storage unit immediately after the one storage unit. An image data processing device characterized in that image data in one bank is read out into another bank. 2. In the image data processing device according to claim 1, the plurality of images to be processed are radiation images recorded on a stimulable phosphor material from the same subject, and the processing in the processing circuit is An image data processing device characterized in that it includes image processing between processed images of a ship's side.