This application claims the benefit of the U.S. provisional application No. 60/030,648, filed on Nov. 12, 1996, which is incorporated herein by reference.

FIELD OF THE INVENTION

The present specification relates to automatically-triggered imaging systems. More particularly, the present disclosure describes a self-triggered X-ray sensor.

BACKGROUND OF THE INVENTION

Medical X-ray imaging devices usually include an X-ray source emitting X-ray radiation that passes through a patient's body part and a solid-state image sensor receiving the transmitted X-ray radiation to form an image of the body part. Many conventional X-ray imaging systems use a semiconductor charge-coupled device (CCD) as the image sensor. A layer of scintillator material is thus disposed in front of the CCD to convert the transmitted X-ray radiation into photoncarriers in a visible spectrum. Photogates in the pixel array of the CCD capture the visible photoncarriers and generate photoelectrons. A CCD signal representing a line-by-line scan of the image is produced by sequentially transferring the photoelectrons in the pixel array to a readout register.

X-ray imaging systems using CCD sensors are widely used in dental imaging applications, such as a compact device to photograph teeth. For example, U.S. Pat. Nos. 5,434,418 to Schick and 5,331,166 to Yamamoto et al. disclose this type of dental X-ray imaging devices. Both Schick and Yamamoto describe an automatic mechanism to trigger CCD readout at the end of exposure. Specifically, Schick uses a computer-aided electronic system to monitor photoinduced charges, i.e., a readout is triggered if the CCD or the photodiode array is no longer conducting. Yamamoto et al. devise an X-ray dose monitoring system to control the exposure of the X-ray. Auxiliary X-ray detectors located on the outer casing of the image sensor are used to generate a control signal to shut off the X-ray source and trigger a CCD readout if the X-ray dose reaches a predetermined dose value.

However, the triggering of the X-ray source is asynchronous with respect to the operation of the image sensor in many prior-art systems. One conventional method in operating the CCD sensor is to continuously read out the sensor. The X-ray source is triggered manually by the operator. The X-rays go through the patient's teeth and strike the sensor. When "bright" pixels are detected, a full frame of data is read out and captured. A dark reference frame, captured either during a calibration step or close in time (e.g., less than 1 second) to the illuminated frame, is subtracted from the image to produce the digital X-ray image with minimized background noise.

SUMMARY OF THE INVENTION

The inventor recognized certain limitations to a conventional X-ray imaging system, which are at least in part due to the lack of synchronous triggering.

For example, one problem is that a conventional system requires the computer to constantly scan for the presence of "bright" pixels. This can lead to the possibility of false triggering due to noise.

Lack of synchronous triggering may also introduce additional noise by integrating dark current signal when there is no photo-induced signal.

Furthermore, the image can be smeared if the CCD is clocked during the X-ray exposure since the photoinduced charge in one pixel can be mixed with that from other pixels during the charge transfer process.

The inventor realized that, a more straightforward way of initiating the readout upon triggering of the X-ray source is desirable.

The present application discloses a self-triggered imaging system to synchronize the operation of the image sensor with an input signal. In particular, the imaging system of the present invention can be used to synchronize the operation of the X-ray source and the operation of the image sensor in a X-ray imaging system, thereby providing advantages over the prior-art systems.

One aspect of the present invention is an event detector responsive to the input photons. An event detector preferably includes a photosensitive element for converting received photons into electrical signals and a thresholding circuit for comparing the photoinduced electrical signals to a preset threshold value. The pixels in the image sensor remain in a non-integrating reset state if the electrical signal in the event detector is below the threshold. The image sensor is switched into an integrating state (i.e., activated) if the electrical signal in the event detector is above the threshold. This provides an autonomous triggering mechanism for an imaging system.

Another aspect of the invention is the capability of minimizing background noise by subtracting a dark reference frame from the integrated signal. This increases the signal to noise ratio.

In a preferred embodiment, a plurality of event detectors are deployed relative to one another on the chip of the image sensor. In particular, an active pixel sensor ("APS") device can be used as the image sensor instead of the conventional CCD sensor. The event detectors are therefore made of additional photosite pixels that are located on the same chip of the APS sensor, preferably in the peripheral locations relative to the active pixel array therein.

Yet another aspect of the present invention is the use of APS as the image sensor, thus resulting in a number of advantages over the prior-art imaging systems based on CCDs. The preferred APS image sensor receives and processes input signals with the active pixels, thus eliminating the charge transfer over macroscopic distances which are inherent in CCDs. Consequently, many drawbacks associated with CCDs are avoided by using the disclosed APS sensor. For example, the readout rate of an APS is usually higher than that of CCDs. On-chip controlling and processing circuits can be integrated with active pixels with the complementary-metal-oxide-semiconductor (CMOS) process or any fabrication technique that is compatible with CMOS, e.g., NMOS. Power consumption of an APS sensor can be significantly reduced with the CMOS circuitry. Non-destructive readout can be implemented with the APS sensor. Moreover, the active pixels allow random access and the performance of APS devices can be maintained as the array size increases.

The sophistication and advantages of the present invention will become more apparent in light of the following detailed description of the preferred embodiment thereof.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a block diagram of an image sensor with on-chip event detection.

FIG. 2 is a schematic of the preferred event detector including a thresholding circuitry.

DESCRIPTION OF THE PREFERRED EMBODIMENT

FIG. 1 shows a preferred image sensor 100 with an event detector circuit with a plurality of event detectors 102. Preferably, all components are formed on a substrate 101. The image sensor 100 includes a pixel array 104 having photosensitive pixels arranged with respect to one another in a one or two dimensional array. CMOS active pixel sensors ("APS") are preferred for the array 104 but other pixel arrays can also be used including CCD arrays and photodiode arrays. In a typical APS device, each active pixel includes a light sensing element and one or more active transistors within the pixel itself. The active transistors amplify and buffer the signals generated by the light sensing elements in the pixels. One type of such APS devices is disclosed in U.S. Pat. No. 5,471,515 to Fossum et al., the disclosure of which is incorporated herein by reference.

The image sensor further includes on-chip logic and timing circuits 106 which control the operation of the sensor including the autonomous sequencing of the sensor 100 triggered by the event detectors 102. Extra jumpers can be implemented for adjusting the length of the delay and other sequencing parameters. These jumpers can be integrated as a part of the chip packaging and add insignificant volume and mass to the final sensor. The event detectors 102 are preferably in the peripheral locations relative to the imaging pixel array 104. Four event detectors 102 are shown in FIG. 1.

The event detectors 102 are scanned while the pixels in the imaging array 104 of the sensor 100 are in a nonintegrating reset state. When a signal is detected by any of the event detectors 102 to be above a preset threshold, an "event" is detected, i.e., there is an input light signal to the image sensor 100. In X-ray imaging systems, the presence of an event is used to indicate that the X-ray source has been triggered. The sensor chip 100 then causes all pixels in the array 104 to be placed into an integrating state, collecting photoelectrons generated by the input light signal. For an X-ray imaging system, the input light signal is the light converted from the incident X-rays by the scintillator layer. The input light converted by the scintillator layer is at a frequency to which the pixel array 104 is responsive. In some cases, this converted light is in the visible spectrum. The sensor 100 is placed into a readout mode at the end of a programmable integration period. In this mode, the pixels are sequentially read out. Optionally, at the end of this mode, the sensor 100 is placed into a second integration period and second readout to capture a dark frame. The dark frame is subtracted from the bright frame to minimize background noise.

The synchronization of the X-ray source and the operation of the image sensor by use of the event detectors obviates the aforementioned drawbacks in the prior-art X-ray imaging systems. For example, the image smearing caused by erroneous timing between the exposure of the X-ray and the CCD operation is greatly reduced. Such a smearing effect can be further minimized by using a CMOS APS pixel array as the sensor 104 rather than a CCD array since the problematic charge transfer inherent in CCDs is eliminated in APS devices. The event detectors also negative the need for constantly scanning the computer to search for "bright pixels" as in the prior-art system. Use of CMOS APS pixels is not required to practice the self-triggering system with the event detectors in accordance with the present invention. However, APS architecture is preferred in order to benefit from the advantages thereof including easy implementation of on-chip control and processing circuits and improved readout rate.

One preferred embodiment of the event detector 102 is shown in FIG. 2. Each event detector 102 has a photodiode 210, a transfer gate 212, an integrating floating diffusion node 214, a reset transistor 220, and a source follower 232 with a load transistor 234. There is also a comparator circuit 240 that generates the actual event pulse. The processing of such event detector is compatible with the widely used CMOS process.

An input light signal 202, which is generated by X-rays in a X-ray imaging system, is incident on the n+ photodiode 210. Photogenerated carriers spill over the potential formed by the transfer gate 212, typically biased at 1.25 V, to the floating n+ diffusion node 214. The potential well of the floating diffusion 214 fills. The voltage on the floating diffusion 214 is buffered by a source follower circuit 230 having transistors 232 and 234. The output voltage 236 is fed to the comparator 240. As the voltage on the node 214 drops, the buffered output voltage 236 also drops accordingly. After a specified sampling time (e.g., typically 20-50 msec), the comparator 240 is strobed by pulsing STRB 242. This activates the comparator 240. If the buffered signal 236 is less than VREF 244, a logic "0" is output, corresponding to a positive event detection. No event will result in a logic "1", that is, the buffered signal 236 is still above VREF 244. The threshold value of VREF 244 is adjusted to discriminate between dark current accumulation and the photosignal. The threshold level for VREF 244 referred back to the diffusion node 214 is indicated by level 216. After the strobe signal 242, the floating diffusion 214 is reset to a high voltage by pulsing the reset transistor RST 220.

The automatic triggering mechanism in accordance with the invention can be used to improve the accuracy of measuring the background noise (i.e., the dark reference frame). A dark reference frame is not correlated with the input photons of an image and therefore needs to be subtracted out from the final image. This dark reference frame can be obtained either in a calibration process or in a short time (e.g., less than 1 sec) prior to capturing an image. The latter is preferred since the background noise captured can more accurately reflect the actual noise level at the time of detection under the actual operating conditions (e.g., temperature and background illumination) of the sensor. The may be accomplished by sweeping in the electrons in the pixels in a pre-sampling step prior to triggering an integration state. Generally speaking, the closer the time for pre-sampling to the triggering of an integration, the more accurate the dark reference frame is. The event detection circuit of the present invention can be used to not only automatically obtain a dark reference frame prior to an integration but also allow this to be done right before the triggering of the integration. For example, the triggering mechanism can be used to terminate capturing a dark reference frame in addition to its role of initiating capture of an image. With use of APS devices, subtraction of a dark reference frame can be implemented with on-chip circuitry to speed up the processing speed and save memory space.

Although the present invention has been described in detail with reference to the preferred embodiment, one ordinarily skilled in the art to which-this invention pertains will appreciate that various modifications and enhancements may be made without departing from the spirit and scope of the following claims.