Active pixel intraoral radiological image sensor and related image capture method - Patents.com

Abstract

To provide an active pixel dental radiation image sensor which starts integration at a timing at which X-ray flash is started in an oral radiation image sensor.SOLUTION: In a 3T pixel matrix image sensor, a switching circuit MUX1 connects the input of an electric current-voltage conversion detection circuit DTX1 to a common connection node NC corresponding to a drain of a photodiode initialization transistor M1 of M-PIX in the first stage of detecting the start of X-ray flash, and provides as the output a signal for detecting the start of the X-ray flash when the electric current generated by all the photodiodes exceeds a predetermined threshold. The switching circuit MUX1 performs control so as to connect the common connection node NC of the pixel to a photodiode re-initialization voltage source VRS.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to the field of medical imaging, and more particularly to an active pixel intraoral dental radiological image sensor using CMOS technology. [Background technology]

[0002] These active pixel structures are based on a photosensitive element (photodiode, photogate) associated with transistors that allow to control the different stages of image capture: initializing the photosensitive element, integrating the charge and reading the pixel. In this readout stage, for each pixel of a row of pixels of the matrix, a voltage level corresponding to the amount of charge accumulated in the pixel during the integration stage is transferred to a column conductor, using a transistor mounted as a voltage follower associated with the node for reading the pixel. The reading is then carried out by a readout circuit at the base of the column. This in fact samples two voltage levels, the voltage level corresponding to the amount of charge accumulated in the node for reading the pixel and the voltage level corresponding to the reinitialization of the readout node, and subtracts them from each other, thus improving the signal-to-noise ratio (double sampling and subtraction of correlated noise).

[0003] Dental radiography is typically performed as follows: A sensor is placed in the patient's mouth and positioned posterior to the anatomical zone to be observed. An X-ray source is positioned and activated to emit an X-ray flash towards the active surface of the sensor through the tissues and materials of the anatomical zone to be radiographed. A sequence circuit of the sensor triggers the image capture when it detects the start of the generation of an X-ray flash at the active surface of the sensor. The X-rays reaching the active surface of the sensor are converted into electrical signals by the pixels, either directly or after being converted into visible light by a scintillator, depending on the type of photodiode. After reading the pixels, an image of the radiographed anatomical zone can be displayed, for example, on a computer screen.

[0004] Detecting the start of the generation of the X-ray flash forms part of the necessary measures to be able to reduce the radiation dose applied to the patient. Detecting the start of the generation of the X-ray flash forms part of the necessary measures to be able to reduce the radiation dose applied to the patient. From the point of view of medical imaging, the recommendation to standardize the body in terms of the radiation dose received by the patient and the practitioner is indeed very strict. With regard to the X-ray source, it includes adjusting the intensity and duration of the X-ray flash for each image acquisition according to the morphology and age of the patient and the anatomical zone to be X-rayed, so that the exposure is as low as possible and there are no unnecessary losses, making it possible to obtain a high quality image in one acquisition. With regard to the image sensor, it includes synchronizing in the best possible way the stage of integrating the charge with the effective exposure of the active surface to the useful radiation, so that the signal-to-noise ratio is optimal with respect to the intensity and duration of the radiation. If the integration stage starts too early, a charge corresponding to a dark current will accumulate. If it starts too late, the useful signal will be lost. Generally, provisions are also made to detect the end of the X-ray flash. This improves the signal-to-noise ratio (no integration of dark current) and the readout phase is also triggered more quickly, improving the speed and efficiency of image acquisition compared to approaches where the integration time is preset to a set value that is longer than the duration of the X-ray flash (X-ray source calibration data).

[0005] Thus, the sensor must integrate a detection circuit, the purpose of which is to allow optimal setting of the start of the integration phase when the X-ray flash begins to occur in the active surface of the sensor. The detection circuit generally uses horizontal and / or vertical bars of photodetectors that surround the matrix of pixels. These bars of photodetectors are thus located immediately behind absorbing anatomical zones (teeth, gums), which impair or delay the detection of the occurrence of the X-ray flash by the sensor. In view of this, the optimal detection zone in the active surface of the sensor can be a free zone (i.e. free of absorbing obstacles) between the upper and lower jaws of the patient, which corresponds to a detection band located in the center of the matrix of pixels of the sensor, as the practitioner wants to see a complete image of the X-ray zone. This also means a loss of image quality, since the interpolation creates inaccuracies.

[0006] The use of pixels distributed in a matrix as reference pixels is known for detection, as described, for example, in US Pat. No. 6,404,854. During detection, these reference pixels are each read individually in a non-destructive manner (without reinitialization between readings) for comparison with a threshold value. This occurs continuously until an image acquisition decision is made if enough reference pixels provide a signal above a certain threshold. A drawback of this approach is that a certain sequence of these pixels is required. In particular, this approach is imperfect in detecting the occurrence of X-rays, regardless of the positioning of the sensor. That is, it is inevitable that, depending on the position of the sensor, a large proportion of the reference pixels will be located behind a highly absorbing zone, which slows down the detection process.

[0007] Furthermore, in order to help reduce the X-ray dose required for radiographic imaging, it is necessary to be able to detect very low signal levels that mark the start of the exposure of the receiving zone to radiation in order to trigger the integration phase without (or as little as possible) losing useful information. In other words, the intention is that the moment at which the detection signal is emitted is very close to the actual arrival of the X-rays on the active surface of the sensor. For this reason, it is more expedient to carry out a direct detection of the currents provided by all the photodetectors using a capacitive transimpedance amplifier, since the significant gain of the amplifier allows the detection threshold to be lowered, i.e. the arrival of the X-ray flash can be determined earlier and therefore the integration can be started earlier. Such amplifiers are well known and are widely used to read the pixels of infrared image sensors. Furthermore, in the field of medical radiology, patent application WO 2017 / 121728 describes a circuit based on such an amplifier for detecting the occurrence of an X-ray flash and for reading the currents provided by a set of detection photodiodes arranged on both sides of a pixel matrix.

[0008] Therefore, there is a significant need to find a technical approach that allows efficient detection of X-ray occurrences independent of the sensor position, without sacrificing or losing image data accuracy, and that is easy to implement in existing topologies, i.e., without requiring a complete redesign and without affecting the active imaging surface. Summary of the Invention [Problem to be solved by the invention]

[0009] The subject of the present invention is an active pixel dental radiographic image sensor with integrated X-ray occurrence detection, which allows efficient detection of X-ray occurrences independent of the position of the sensor in the patient's mouth and maximizes the active surface (matrix) given the size of the sensor. [Means for solving the problem]

[0010] Specifically, the present invention relates to an intraoral radiographic image sensor using MOS technology, comprising: A matrix of light-sensitive pixels arranged in rows and columns, each pixel including a photodiode and a transistor, and including a photodiode initialization transistor (M1) connected between the photodiode node of the pixel and a first connection node common to the pixels. a sequence circuit providing signals for controlling the transistors of the pixel to control an image capture sequence including the steps of a global initialization of the photodiode of the pixel during exposure to an X-ray flash, integrating the charge during an integration period, and reading out the pixel.

[0011] According to the present invention, the sensor includes a first connection switch controlled by a first logic signal for connecting the first connection node to a signal input of a first current detection circuit or to a photodiode initialization voltage source, the first logic signal being in a first logic state or a second logic state, respectively; The sequence circuit of the sensor is configured to control detection by the first detection circuit of the start of exposure to an X-ray flash for initiating an image capture sequence, the image capture sequence (300) including the following operations: a) Commanding the photodiode initialization transistors to switch to the on state in all pixels simultaneously. b) establishing a first logic signal at said first logic state, the effect of which is to inject a current collected at said first connection node originating from the photodiode of the pixel at a signal input of a first detection circuit. c) when an output logic signal of the first detection circuit switches from a first logic state to a second logic state in response to detection of an input current level exceeding a predetermined threshold, establishing said first logic signal at said second logic state, the effect of which is to connect said initialization transistor, which is always on, to said initialization voltage source and to initiate a global initialization phase of the image capture sequence for initializing the photodiode prior to said integration phase.

[0012] In one refinement, the initialization transistors of a set of pixels of the matrix are connected not to the first connection node but to a second connection node electrically isolated from the first connection node, and the sensor includes a second current detection circuit of the same type as the first detection circuit and controlled by the same clock signal, with an input signal coupled to the second connection node, the initialization transistors of the pixels of the set being in the on state at least until detection of the end of exposure to X-rays by the second detection circuit, corresponding to a current injected at its input and collected from the second connection node falling below a predetermined threshold. Upon detecting the end of the exposure to the X-ray flash, the phase of integrating the image acquisition sequence for triggering the read-out phase is stopped.

[0013] Conveniently, a second connection circuit is provided for connecting the second connection node to a signal input of a second detection circuit or to an initialization voltage source as a function of a second control logic signal, which is configured in the sensor to incorporate said set of pixels into end of exposure detection pixels or image capture pixels.

[0014] According to one aspect of the invention, the gain of a capacitive transimpedance amplifier and / or the voltage threshold of a comparator are adjustment parameters of the sensor tuned in the first detection circuit and the second detection circuit to detect the start and end of an X-ray flash, respectively.

[0015] According to another aspect of the invention, the sequence circuit is configured to trigger other image capture sequences executed during exposure to an X-ray flash and applying the same integration period and using the same image capture pixels as a previous image capture sequence that provided first image data, allowing measurement and subtraction of dark noise in said first image data.

[0016] The sensor according to the invention is particularly applicable to active pixels with more than two transistors, in particular with more than four transistors.

[0017] According to one embodiment of the invention, the current detection circuit of the sensor comprises a capacitive transimpedance amplifier with a non-inverting input connected to a common mode voltage and an inverting input connected to the common connection node of the pixel and forming a signal input, said amplifier being controlled by a clock signal to periodically generate at its signal output a voltage ramp as a function of the level of current injected at the signal input, this ramp being applied to a comparator for comparison with a voltage threshold.

[0018] The invention also relates to a corresponding radiographic imaging method. [Brief description of the drawings]

[0019] The invention will be better understood and further advantages will become apparent on reading the detailed description of embodiments given by way of example and illustrated with the aid of the attached drawings, in which:

[0020] [Figure 1] FIG. 2 is a functional block diagram of the detection of the start of an X-ray flash according to the present invention, applied to an image sensor based on an active pixel structure of the three-transistor (3T) type. [Diagram 2] FIG. 1 is a block diagram of such a 3T pixel matrix image sensor including a device for detecting the start of exposure to an X-ray flash according to the present invention. [Diagram 3] FIG. 1 is a block diagram of a capacitive transimpedance amplifier sensing device. [Figure 4] 4 is a timing chart of control signals for detecting the occurrence of an X-ray flash and an image capture sequence in an image sensor having an X-ray flash start detection circuit according to the present invention. [Diagram 5] 4 is a timing diagram of the signals of an X-ray flash detection circuit used in accordance with the present invention to detect the start and end of an X-ray flash detection circuit of the same type. [Figure 6] FIG. 1 is a functional block diagram of the present invention applied to an active pixel of at least five transistor (5T) type. [Figure 7]FIG. 1 is a block diagram of a 3T pixel sensor according to an embodiment of the present invention that enables detection of the start and end of an X-ray flash. [Figure 8] FIG. 8 is a block diagram of a 5T pixel sensor corresponding to the embodiment of FIG. 7. [Figure 9] 4 is a timing diagram of control signals for radiographic imaging according to the present invention, including detection of the start and end of an x-ray flash to respectively activate and deactivate steps of an integrated radiographic imaging sequence; DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0021] It should be noted that the drawings are simplified diagrams illustrating the present invention and are not to scale. Only the functional elements or signals necessary to understand the present invention are shown. For clarity and simplicity, the same notations or references are used throughout the description and figures to denote the same or similar elements.

[0022] Additionally, in the following description, the terms "coupled" or "connected," when used, indicate a direct or indirect electrical connection. The term "connected" indicates a direct electrical connection.

[0023] The present invention relates to radiological image sensors, and in particular to intraoral dental image sensors using CMOS technology with active pixels.

[0024] An active pixel is a pixel, as shown by way of example in figure 1, containing a light-receiving element, typically a photodiode PH, and several transistors (MOS), all of the same type, used as switches (on or off state) or voltage followers, to control the different stages of the image capture sequence. Strictly speaking, CMOS technology (combination of N-type and P-type transistors) is concerned in particular with the electronic circuits around the pixel (control circuits, readout circuits, interface electronics, etc.).

[0025] The present invention is more particularly illustrated in the context of an image sensor fabricated in a P-type doped semiconductor substrate that is brought to a ground reference potential, typically zero potential.

[0026] The photodiode PH is preferably of the “pinned” type, i.e. the junction PN is connected to a lightly doped substrate (P - ) and this region N (the cathode) is formed by a heavily doped P-type surface layer (P ++ ) and is tied to the ground reference potential of the substrate, which sets the potential of the photodiode well N (region N) to a potential V pin (This is set to the rest voltage of the photodiode, which depends only on the concentrations of the N and P dopants.

[0027] The transistors of the active pixels are N-type. They are commanded to switch to the on-state (off or on) by applying a positive potential Von to their gate g, which generally corresponds to the positive supply voltage Vdd of the sensor's electronics. The level of this voltage Vdd (for example 3 volts (Figure 4)) is a function of the CMOS technology used. They are commanded to switch to the off-state (open or blocked) by applying to their gate g a potential generally corresponding to the ground reference potential of the substrate (zero potential) or a more negative potential.

[0028] The sensor's logic control signal can be one of two logic values, "0" or "1", corresponding to the substrate ground reference potential and Vdd, respectively.

[0029] The figures and descriptions reflect these conventions, and those skilled in the art will know how to easily replace all of these with active pixels that include P-type MOS transistors.

[0030] In dental radiology, active pixels are generally used with three control transistors (3T), which facilitates the storage capacity of the pixel. These pixels can only generate what are called "sliding shutter window" images. The reason is that the end of the integration period is offset for each row and coincides with the sequential reading of the pixels by scanning the rows. As for the start of the integration period, in the context of dental radiology, it is the same for all pixels, since it is synchronized upon the detection of the occurrence of an X-ray flash (or more simply "X-flash") on the active surface of the sensor. The effective integration period of a row therefore increases with the row scanning direction. Although the difference in terms of the total integration time is small, this causes a line effect in the image, which can be attenuated by image processing. This is well known.

[0031] The invention describes in detail a sensor using a matrix of such 3T pixels. However, it will be seen that it is easily applicable to structures with more transistors (in particular 5T structures, i.e. including (at least) 5 transistors), with the particular advantage of allowing instantaneous image capture ("snapshot") with the start and end of integration being identical for all pixels. It will be noted that in the context of the invention, the sensor is only sensitive (by construction) to X-rays. It is therefore blind to visible light. During the integration period, it integrates only the charge corresponding to the received X-rays.

[0032] Figures 1 and 2 show the principle of detecting the start of an X-ray flash according to the invention in an image sensor based on a matrix M-PIX of 3T active pixels, in this example comprising N rows and P columns of pixels. Figure 1 shows this by focusing on the pixels of the matrix, while Figure 2 shows the implementation of the invention in a sensor, using all the pixels of the matrix, making it possible to first detect the start of exposure to the flash and then, once the start of the flash has been detected, to carry out an image acquisition sequence.

[0033] A pixel row R of rank i i and the pixel column of rank j, Col j A pixel P located in the matrix ij (Figure 1) has the following: Photodiode node KN i A photodiode PH (preferably of the "pinned" type) connected between its cathode and a reference ground voltage. Photodiode node KN i A photodiode initialization transistor M1 connected to the gate g of the photodiode initialization transistor M1 outputs a signal RSkn for initializing the pixel row of rank i to command the discharge of the charge (electrons) accumulated in the photodiode towards a reference voltage source Vrs (hereinafter also called photodiode initialization voltage). i Receive. Transistor M2 is mounted as a voltage follower. Its gate is connected to the photodiode node KN i (capacitive), which converts the charge stored in the photodiode PH into a voltage level and, when the pixel is selected for reading, is connected to the column conductor Col j will be forwarded to. Transistor M2 and a column conductor Col connected to a read circuit RD j A transistor M3 is connected in series between the pixel row R i Signal Sel for selecting the reading i Receive.

[0034] Regarding notation, KN i denotes the photodiode node of the pixel in row i of the row rank. This node is then denoted by RSkn i , which can be easily related to the photodiode reinitialization signals for each row indicated by .

[0035] The read circuit RD makes it possible to obtain, for each pixel of the associated column, a digital value representative of the amount of charge integrated by the pixel.

[0036] Therefore, as shown in FIG. i Two control signals are provided for each pixel: a signal RSkn for initializing the photodiode; i and a signal Sel for selecting a readout applied to the gates of the transistors M1 and M3 of all pixels in the row of rank i, respectively. i There is.

[0037] These signals are commanded by a sensor sequence circuit DM to control the dental radiography sequence, which conventionally comprises the following successive steps (see FIG. 4): A stage 301 of global initialization RSG of the photodiodes, i.e. all signals RSkn applied simultaneously to all pixels of the matrix and activated by commanding all transistors M1 to switch to the on state simultaneously. i (i=1~N) is activated (set to Von). All signals RSKN i An integration phase 302 starts simultaneously for all pixels, commanded by deactivating (returning to VOFF). A sequential read-out step 303 by scanning the rows of pixels. This has an integration period d INT (the elapsed time from the beginning of the integration phase). During this phase, the reading period d r During the signal Sel i are activated (set to Von) one after the other (one at a time). The read circuit RD is generally arranged to take two samples, namely a sample of the signal level (SHS) corresponding to the amount of charge integrated by the photodiode since the beginning of the integration phase, and an initialization signal RSkn of the current row selected for reading. i Reinitialization of the photodiode by activating RS i This is a later baseline sample.

[0038] In fact, electrically, the drain electrode d of the initialization transistor M1 of each pixel is connected to a connection node NC common to all pixels and, typically, to a reference voltage source VRS used to reinitialize the photodiodes (to discharge any charge accumulated in the potential well of each photodiode).

[0039] Topologically, as shown in FIG. 2, for each pixel, this common connection node NC corresponds to a connection point (shown as a black dot) to a conductive gate GRS (shown as a thick line) generated on the conductive topology level of the sensor chip. The term "gate" is understood in terms of a periodic 2D mesh in line with the matrix. The gate GRS is appropriately electrically connected to a reference voltage source VRS and is responsible for the global initialization phase RSG and the selective initialization phase (RS i , R.S. i+1 , . . . ).

[0040] According to the invention, as shown in figures 1 and 2, it is provided that the common connection node NC (or gate GRS) is not directly connected to the initialization voltage source VRS, but is connected to this voltage source VRS by a switching circuit as a function of the logic state "0" or "1" of the control logic signal PROB1. This switching circuit MUX1 comprises two channels for connecting the node NC, the first channel X1 being connected to the voltage source VRS and the second channel X2 being connected to a signal input IN-DTX1 of a current detection circuit DTX1 comprising a current-to-voltage converter of the sensor. The purpose of this circuit DTX1 is to provide a signal OUT-DTX1 as an output indicating the detection of the start of the exposure of the sensor to an X-ray flash, when the current I generated by all the photodiodes of the matrix (of the pixels) and collected at the common connection node NC through the initialization transistors M1, all activated to the on state, and the circuit MUX1, exceeds a predetermined threshold value.

[0041] In detail, as shown in the timing diagram of FIG. 4, the sequence circuit DM commands a stage 200 of detecting the start of the X-ray flash (upon completion of the standby stage STANDBY100) by the following series of steps: By putting the signal PROB1 in a first logic state, in this example a high state ("1") for selecting the channel X2 of the switch MUX1, the common connection node NC of all the pixels is connected to the current input IN-DTX1; and All photodiode initialization signal RSKN i By simultaneously activating i (cathode) and a common connection node NC. All charges from all photodiodes of the matrix are thus collected at the node NC, generating a photocurrent I which is injected at the input IN-DTX1 of the circuit DTX1. This makes the detection of the occurrence of an X-ray flash according to the invention efficient and accurate, since it takes into account the current contribution of all pixels of the light-receiving matrix of the sensor. The matrix of pixels is then used as a vast well of charge to detect the beginning of exposure to an X-ray flash. Thus, regardless of the position of the sensor in the patient's mouth, there will always be a pixel in the matrix where the X-rays reach with little or no attenuation.

[0042] When the current detection circuit DTX1 detects that a current passes above a predetermined threshold, the output logic signal OUT-DTX1 changes state, in this example from a high logic state "1" to a low logic state "0", which is then detected by the sequence circuit DM and triggers the image capture sequence 300 by the following steps: By putting the signal PROB1 in the other logic state, in this example the low state ("0") for selecting the channel X1 of the switch MUX1, the common connection node NC of all the pixels is connected to the initialization voltage VRS. During the global initialization RSG phase 301, the initialization transistor M1 is kept active.

[0043] The other stages of the image acquisition sequence controlled by the sequence circuit DM (integration 302, sequential read-out 303) then follow in the usual way, which means that in the present invention a pixel of the matrix is ​​first used as a pixel for detecting the start of exposure to an X-ray flash and then as an image acquisition pixel.

[0044] 3 shows an embodiment of a current detection circuit DTX1 that allows the implementation of the detection stage 200, comprising a capacitive transimpedance amplifier CTIA1 and a voltage comparator COMP1. It will be noted that the use of capacitive transimpedance amplifiers is known for integrating the current provided by detection photodiodes arranged at the edge of a matrix of pixels and for detecting the arrival of an X-ray flash, as described, for example, in the aforementioned patent application WO 2017 / 121728. These capacitive transimpedance amplifiers are also used to read out infrared sensor pixels, as described, for example, in EP 1 399 746.

[0045] The capacitive transimpedance amplifier CTIA1 operates at a clock frequency of Φ ctia ensures the periodic integration of the input current I, and subsequently the voltage comparator COMP1 regulates the signal V out-ctia1 The voltage level of the programmable voltage threshold V th1 The voltage comparator is configured to provide as an output a logic signal OUT-DTX1 and to compare the signal level V out-ctia1 is the threshold V th1 If the threshold voltage is exceeded, it switches from an initial logic state (typically a "0") to another state (a "1").

[0046] The amplifier CTIA1 of Fig. 3 corresponds to a possible embodiment based on the basic configuration of a capacitive transimpedance amplifier, where the feedback loop includes an initialization switch T1 arranged in parallel on the integrating capacitor Cf, however, the invention is not limited to this embodiment and covers other implementations of the prior art.

[0047] The non-inverting input (+) of amplifier CTIA1 is connected to a common-mode voltage V cm1 The inverting input (-) is the signal input IN-DTX1, through which the current I to be measured is injected. The switch T1 discharges the integrating capacitor Cf, resulting in an output voltage V out-ctia1 and input IN-DTX1 to the common mode voltage V cm1 (used as a reference point for the periodic integration of the input current I) ctia The integration starts when switch T1 returns to its open state (OFF) and all the current injected at the input is integrated at the terminals of the capacitor Cf. The gain depends on the ratio of the feedback loop capacitor (Cf in this example) and the capacitor Cp at the signal input (representing the (equivalent) parasitic capacitance of the photodiode).

[0048] As shown in Figure 5, the signal is a voltage ramp at the output V out-ctia1 is obtained.

[0049] A comparator, for example, detects the voltage V out-ctia1 , a configurable threshold V th1 and a stage for shaping the signal to provide as output a logic signal OUT-DTX1. This signal is typically a voltage V out-ctia1 is the threshold V th1 When the input voltage Vcc exceeds 1 V, it switches from a low state to a high state (Fig. 5). It is this switching action that is detected by the sequence circuit DM in order to synchronize (trigger) the image acquisition sequence upon the arrival of the X-ray flash.

[0050] Next, the sequence circuit DM switches the signal PROB1 to the other logic state ("0" in this example) to select the first input channel X1 of the circuit MUX1. Then, the common connection node NC of each pixel is connected to the initialization voltage source VRS of the photodiode. The transistor M1 is in the on mode (RSKN at VON). i ), they are kept in this mode (FIG. 4), the effect of which is to activate a stage 301 of global reinitialization of the pixels RSG, which is the first stage of an image acquisition sequence 300 (I-FX) during exposure to an X-ray flash. The sequence circuit DM then controls an integration stage which starts simultaneously for all pixels and then controls the sequential readout stages in the usual way.

[0051] 4 shows the complete sequence for detecting the start of the x-ray flash and image capture, where the practitioner positions the sensor behind the selected anatomical zone with the active face facing the x-ray source and activates the x-ray source, for example, by starting the x-ray source from a control computer or directly. This activation action causes the sensor to exit standby mode 100 (STANDBY) via an activation signal sent by the computer and / or the x-ray source via a wireless or wired (USB) interface.

[0052] The sensor leaves the standby mode (Standby) and enters an initialization phase 101, allowing the power supply VDD and reference VRS voltages to be established. The amplifier CTIA1 outputs out-ctia1 and the voltage at the input IN-DTX1 is the common mode V cm1 Voltage level (Φ ctia 5), which is a level higher than the rest voltage of the photodiode. In this example, V pin is 0.8 volts, and V cm1 is 1 volt. As shown in Figure 4, this results in a common connection node NC (gate GRS) and a node KN i The potential of this same common mode level V cm1 Increase to.

[0053] The sequence circuit DM then controls a step 200 of detecting an X-ray flash as follows. Initialize all rows RSKN for controlling the gate of transistor M1 i Set to VON. And, Activating the signal PROB1 to its logic state ("1") selecting the second channel X2 of the circuit MUX1, which connects the common connection node NC (and therefore the conductive gate GRS) of all of the transistors M1 of the pixel to the input IN-DTX1 of the current detection circuit DTX1.

[0054] A signal SEL for selecting the reading of a row of pixels i remain inactive (low) for the entire duration of the detection phase 200.

[0055] When a flash X-ray beam reaches the active surface of the sensor, the photodiodes generate an electric charge and generate a current I at the input IN-DTX1, which represents the contribution of all the photodiodes of the matrix. This current is integrated by the amplifier CTIA1, which generates a voltage ramp at the output. The ramp voltage V out-ctia1 is the threshold V th1 (set to 1.5 volts in this example), the output OUT-DTX1 switches to a high (logic) state (arrow No. 1, FIG. 4), and the toggling (detection of a rising edge and / or high logic state) of the signal OUT-DTX1 indicates the detection of the start of an X-ray flash in the sequence circuit DM. In a non-limiting practical embodiment, the comparator COMP1 is driven by a clock frequency Φ ctia (Figure 5) The sampling clock Φ S / H The output stage includes a latch-type output stage triggered by

[0056] When the circuit DM detects a switch in the signal OUT-DTX1 indicating the detection of the start of exposure, it is able to control the successive stages of the image capture sequence 300 (I-FX) during exposure to the flash, as already described above. The signal PROB1 changes logic state (arrow No. 2), in this example to "0", selecting the other channel X1 of the circuit MUX1, and all nodes NC (gate GRS, Fig. 2) are electrically connected to the reference voltage source VRS. The input IN-DTX1 is disconnected from the node NC and no longer receives any current. The output voltage V out-ctia1 is the common mode voltage level V cm1 (and stays on) and switches the output OUT-DTX1 of comparator COMP1 back to the low state (Figures 4 and 5). Signal RSKN that commands the initialization of the photodiode to switch to the active state (VDD) i The photodiode is maintained in an active state (transistor M1 is on) so that the overall initialization RSG step 301 of the image capture sequence can be executed. That is, the photodiode is drained of charge, and the capacitive node KN i is set to a voltage level (2 volts in this example) corresponding to the reference voltage VRS (3 volts) obtained by subtracting the threshold voltage of the transistor M1 (arrow No. 3).

[0057] Next, the sequence circuit DM outputs all the signals RSKN i transitions to an inactive state (FIG. 4), thus blocking all transistors M1, which marks the beginning of the integration pixel stage 302, where the photodiodes start to integrate charges through the effect of electrical-to-optical conversion and accumulate these charges (parasitic photodiode capacitance). i The potential of V as a function of the accumulated charge (as a function of illumination) reaches a minimum value V sat decreases to.

[0058] In the example shown in FIG. 4, the sequential read step 303 has an integration period d 1 preset to a value greater than the emission period dX of the X-ray source used. INT The period d counted (for example by a counter in the sequence circuit DM) from the beginning of the integration phase begins after INTAt the end of this, the sequence circuit DM initiates a sequential read stage 303, enabling the pixels of each of the N rows of the matrix to be read row by row.

[0059] Pixel R i Take the row: Signal SEL to select this row i is activated to switch the select transistor M3 of each pixel in this row to the on state (ON). For each pixel in the row, the voltage level provided by the transistor M2 is connected to the photodiode (capacitive read node KN i ) corresponds to the amount of charge stored at this instant in the column conductor of the pixel, Col j (through M3) and sampled (SHS) by the individual read circuit RD. Then, to reinitialize only this row, the signal RSKN i By activating i+1 Signal RSKN i+1 It can be seen that remains inactive at this point.) A new voltage level, which is the reinitialization level, is applied to the column conductor Col j and sampled (SHR) by the read circuit. The difference between the signal level and the reinitialization level represents the image data provided by the pixel. The same read sequence is repeated successively for each of the rows of the matrix.

[0060] The sequence circuit DM generally performs the same integration period d following the first radiation image acquisition I-FX sequence 300. INT by applying the IR-image acquisition signal 300 to control a second dark noise I-obs image acquisition sequence 300 for measuring the dark current level when the sensor is not exposed to X-rays. The dark noise I-obs image data is then subtracted point-by-point from the data of the first image. A better quality radiographic image is obtained.

[0061] integration period d INTOnce the refresh is set (meaning there is no end of X-ray flash detection in the sensor), the second sequence 300 (I-obs) is not executed systematically every time a radiation image is taken, but only occasionally, in particular to take into account temperature variations. Between two refreshes, the image data of the dark noise is stored, for example, in a memory circuit associated with the control circuit DM and subtracted from the obtained radiation image data.

[0062] Although the invention has been described for a 3T pixel sensor, it can be extended to active pixels with more than three transistors, and through the presence of a read node separated from the photodiode node by at least one transfer transistor, the same integration period can be applied to all pixels, thus reducing the effect of dark noise in the image.

[0063] The invention is particularly applicable to 5T pixels (understood to include at least five transistors) including an initialization transistor M1 specific to the photodiode as for 3T pixels, as shown in figure 6. The invention is implemented in the same way, the current detection circuit DTX1 and the circuit MUX1, at the stage of detecting the start of the exposure to an X-ray flash, can connect the common connection node NC first to the signal input IN-DTX1 of the circuit DTX1 and then to the reference voltage source Vrs, allowing the photodiode to be initialized before the stage of integrating the image capture sequence triggered by this detection.

[0064] Compared to the 3T pixel in Figures 1 and 2, the 5T pixel (Figure 6) has a photodiode node KN i from the transfer control signal TF i A capacitive read node SN separated by a transfer transistor gate (designated M4) controlled by iIt will be noted that the 5T pixel further includes a current integration phase (typically a floating diffusion) which is activated to transfer the charge integrated by the photodiode to the readout node. It is thus possible to control an overall transfer phase which is commanded simultaneously by transistor M4 in all pixels, which marks the end of the current integration phase in all pixels before the sequential readout phase. The 5T pixel also includes a current integration phase which is commanded by transistor M5 in all pixels, which is commanded by transistor M6 in all pixels, which is commanded by transistor M7 in all pixels, which is commanded by transistor M8 in all pixels, which is commanded by transistor M9 in all pixels, which is commanded by transistor M1 in all pixels, which is commanded by transistor M2 ... i The transistor M5 includes an initialization transistor M5 of RSSN i The control signal denoted by is activated in the step of reading the pixel, followed by sampling SHR, which corresponds to the reinitialization level, and subtracting it from the signal level (SHS).

[0065] In an alternative embodiment of the invention, applicable to pixels with more than two transistors, the same principle for current detection of the start of exposure to an X-ray flash is also used to detect the end of exposure to an X-ray flash, which determines the end of the integration period, allowing the integration duration of the pixel to be adjusted as closely as possible to the actual duration dX of the flash, thereby reducing the integrated dark current level.

[0066] In detail, according to the invention, this end-of-exposure detection according to the invention is performed by a detection circuit DTX2, similar to the circuit DTX1, based on a current I' generated by a set of pixels of the matrix configured as detection pixels. As shown in Fig. 7 (3T pixels) and Fig. 8 (5T pixels), these pixels are connected to a different common connection node NCE, electrically isolated from the node NC, and the sensor comprises a second detection circuit DTX2 for receiving the current I' photogenerated by the pixels connected to the node NCE. This circuit DTX2 comprises a capacitive transimpedance amplifier CTIA2 and a settable voltage threshold V th2 The two detectors have the same operating frequency (Φ ctia ) The two amplifiers CTIA1 and CTIA2 can operate at the same common-mode voltage (V cm1 =V cm2).

[0067] Topologically, in this alternative embodiment it is possible to have two conductive gates insulated from each other, one interconnected with node NC and the other with node NCE.

[0068] As shown in the timing diagram of FIG. 5, in the phase of detecting the start of exposure, the two detection circuits DTX1, DTX2 operate in a similar manner: one integrates the current I resulting from the pixels of the matrix used for image capture, and the other integrates the current I' resulting from the pixels of the matrix configured as detection pixels (this will be explained below). In both circuits, the amplifier generates an output voltage ramp and the outputs OUT-DTX1 and OUT-DTX2 of the comparators are set when the ramp voltage exceeds a comparison threshold (V th1 ,V th2 ), in this example it switches to "1".

[0069] The circuit DTX1 for detecting the start of an X-ray flash is then disconnected from the node NC (by the action of the signal PROB1), i.e. the amplifier CTIA1 of the detection circuit DTX1 no longer receives a current as input and the output voltage V out-ctia1 is the common mode voltage V cm1 and at the output of the comparator COMP1 the signal OUT-DTX1 switches again to the "0" state in this example. The pixels of the matrix connected to the node NC are then used to execute an image acquisition sequence, i.e. the sequence circuit, as seen above, first controls the phase of the global reinitialization RSG of the photodiodes and then activates the integration phase (FIG. 9).

[0070] As long as the sensor is exposed to the X-ray flash, the detection circuit DTX2 continues to receive the current I' photogenerated by the pixels of the matrix connected to the node NCE, i.e. the amplifier CTIA2 continues to generate as output a periodic voltage ramp, which is equal to or exceeds the threshold V th2Therefore, the output OUT-DTX2 of the comparator COMP2 remains unchanged, and in this example remains "1".

[0071] The end of exposure to the X-ray flash is represented by the current I' which practically no longer increases, i.e. the slight increase is related to the dark current of the photodiode. The lamp voltage then drops below the threshold V th2 , i.e. the signal OUT-DTX2 switches again to the "0" state in this example. This switch is detected by the sequence circuit DM and is used to stop the phase of integrating the current image acquisition sequence.

[0072] The number of pixels of the matrix used to detect the end of exposure can be small compared to the total number of pixels, and can be in the ratio 1 to 2000. In practice, this can be achieved by adjusting the gains of the amplifiers CTIA1, CTIA2 and / or by using different comparison thresholds (V th1 ,V th2 ) is taken into account by using the operating frequency (Φ ctia ,Φ S / H ), i.e. the patient is no longer irradiated, so the speed of detecting the end of exposure does not have to be as good.

[0073] The number of pixels of the matrix used to detect the end of exposure can be less than the total number of pixels, in a ratio of 1 to 2000. In practice, this can be achieved by adjusting the gains of the amplifiers CTIA1, CTIA2 and / or by using different comparison thresholds (V th1 ,V th2 ) is taken into account by using the operating frequency (Φ ctia ,Φ S / H ) can also be decreased. The speed of detecting the end of exposure does not have to be as good since the patient is no longer irradiated.

[0074] In practice, the detection pixels connected to the node NCE may be pixels of several columns and / or rows of pixels of the matrix and / or pixels distributed in the matrix. In the example shown in Fig. 7 and Fig. 8, the pixels used to detect the end of the X-ray flash are pixels of the first column (set E1) and the last column (set E2) of the matrix. To simplify the representation of the drawings, the sets E1, E2 are consecutive columns at each matrix edge (consecutive rows could be used). In practice, however, rows and / or columns interleaved with the "normal" pixel columns will preferably be selected for image acquisition. For example, one column in two columns, or four columns among the first 20 and the last 20 columns of a matrix containing several hundred columns, or even pixels distributed throughout the matrix. In practice, the pixels used for X-ray flash end detection have their signals lost for reading. Image information has to be reconstructed for each flash end detection pixel by interpolation based on the neighboring pixels. Therefore, to limit the interpolation error, it is better to place these detection pixels far away from each other.

[0075] A common connection node NCE of the pixels used for flash end detection is connected to an input IN-DTX2 of the circuit DTX2, while the pixels of the matrix associated with the node NC are used in the current image capture sequence, and this common connection node NC is connected to a voltage reference source (VRS).

[0076] During the current image capture sequence, the transistors M1 of the pixels connected to the node NCE must be kept in the on state to allow the collection and injection of the current I' originating from the photodiodes of these pixels in the circuit DTX2, while the transistors M1 of the pixels connected to the node NC and performing the current image capture are deactivated at the end of the global initialization RSG phase to allow the integration of the charge.

[0077] Thus, in the pixels connected to the node NCE, a signal denoted RSKN_D is provided to control a separate transistor M1, and this signal is the same for all these pixels. D becomes active as soon as the sensor leaves standby and remains active at least until the end of exposure is detected.

[0078] The exposure end signal OUT-DTX2 is used by the sequence circuit DM to end the integration phase of the current image acquisition sequence. This signal OUT-DTX2 thus ends the current active integration period d INTV Set.

[0079] Advantageously, end of exposure detection can be provided as an optional feature in the sensor, by providing a multiplexer MUX2 controlled by a logic control signal PROB2, as shown in Figures 7 and 8, in order to connect the node NCE to an input IN-DTX2 of the circuit DTX2 (channel X'2) to use the associated pixel as an X-ray flash end of detection pixel, or to connect the node NCE to a voltage reference source VRS (channel X'1) to use the associated pixel only as an image capture pixel. The logic state of the signal PROB2 is then typically set by an operator in a parameter register of the sensor to activate or deactivate the X-ray flash end of exposure detection function. For example, as shown, the parameter register contains, for example, a DTX stop bit, which is set to "0" to deactivate the end of exposure detection option and to "1" to activate said option, resulting in the logic signal PROB2.

[0080] In this case, the transistors M1 of the pixels connected to the node NCE must be appropriately controlled depending on whether they are used for end-of-exposure detection or as image capture pixels. For example, as shown in the figure, a logic circuit such as a multiplexer MUX3 controlled by this same signal PROB2 can be provided, which can be as follows: When the detection option is inactive, PROB2 is equal to 0 and the transistor M1 of each pixel connected to the node NCE receives the control signal RSKN corresponding to the row of the pixel in the matrix. i is controlled by. When the detection option is activated, PROB2 is 1 and all transistors M1 of the pixels connected to node NCE are controlled by the control signal RSKN_D.

[0081] 7 and 8 show that the principles of end of exposure detection according to the present invention described herein are equally applicable to 3T or 5T pixel sensors.

[0082] 9 shows in more detail a timing diagram of the signals for detecting the start and end of exposure to an X-ray flash, and the associated control signals when end-of-exposure detection is implemented (in this example PROB2="1"). In this example, the I-FX and I-obs image capture sequence 400 begins with the sequential readout phase 404, as shown in FIG. 8, for a 5T pixel image sensor, and begins with the sequential readout phase 404, as shown in FIG. i 7, this whole transfer stage 403 does not exist and end of exposure detection triggers a sequential read stage 404.

[0083] In the image capture sequence 400 (I-FX or I-obs) controlled by the sequence circuit DM, after detection of the start of exposure (arrow No. 5), the stage 401 of the global initialization RSG starts with all the signals RSKN i (i=1 to N) simultaneously activates the photodiode nodes KN of the pixels connected to node NC. i The signal RSKN_D is connected to the photodiode node KN of the pixel connected to the node NCE. i Initialize all signals RSSN i is the read node SN for all pixels of the matrix i simultaneously (whether they are connected to node NC or NCE).

[0084] The end of exposure detection (arrow No. 6) is performed by detecting the photodiode node KN i Read from node SN i Activate the overall transfer stage 403TFG for transferring charges to all the image capture pixels (signal TF i ) is applied to all pixels, so that the same integration period, i.e., the signal RSKN i At the same time, the same start corresponds to the end of the phase RSG, and the signal TF i At the same time, when the phases TFG and TFG are inactive, the same end corresponding to the end of the phase TFG is set. The phase 404 of reading pixels starts. The row of rank i selected for reading this phase is subjected to sampling of the signal levels of the pixels of the row (SHS) and the corresponding reference level (SHR) of the signal RSSN of the row. i By activating the read node SN i It will be noted that the resulting data stream contains the data of the pixels connected to the node NCE used to detect the end of the exposure, i.e. the data obtained (read) for these pixels will not actually be used but will be replaced by the data calculated by interpolation. This does not change any of the sequences, which is taken into account in the image processing.

[0085] It will be noted that the effective integration period does not have a pre-determined (adjusted) set value. That is, the value d INTV is defined from the two detection signals OUT-DTX1, OUT-DTX2. Therefore, if it is intended to apply the same integration period to the dark noise image capture sequence I-obs, it must be measured. Then, the sequence circuit is connected to the photodiode node KN i The effective exposure time d between the end of the phase 310 of the global reinitialization RSG and the detection of the end of the flash (OUT-DTX2) which triggers the read phase (3T pixels) or the global transfer phase TFG (5T pixels) INTVA counter is provided for measuring

[0086] Time d INTV can potentially change with each new exposure to an x-ray flash, and dark current imaging is systematic.

[0087] In fact, the sequence circuit DM applies a filter to the detection signals OUT-DTX1, OUT-DTX2 and connects the connection node NC or NCE and the photodiode node KN i This allows to eliminate, i.e. not to take into account, spurious (parasitic) pulses which would be induced in particular when switching channels in the circuit switching circuits MUX1, MUX2 due to potential variations induced in the OUT-DTX1 signal during the image capture sequence 300, and to filter to ignore the signal OUT-DTX2 during the stage 200 of detecting the start of the flash exposure. These various filtering measures ("anti-glitch" filters) are typical measures implemented to avoid false detections.

[0088] The invention described herein allows the quality of radiological images, particularly dental images, to be improved at a lower cost, because it uses a matrix of pixels and current detection circuits known to those skilled in the art.

Claims

1. An intraoral radiographic image sensor using MOS technology, comprising: A matrix of light-receiving pixels (M-PIX) arranged in rows and columns, each pixel (P ｉｊ ) includes a photodiode (PH) and a transistor, and the photodiode node (KN ｉ a matrix (M-PIX) including a photodiode initialization transistor (M1) connected between the photodiodes (M1) and a first connection node (NC) common to the pixels; a sequence circuit (DM) for providing signals to command the transistors of the pixel in order to control an image capture sequence (300) including the steps of global initialization (301) of the photodiode of the pixel during exposure to an X-ray flash (FX), integration of the charge during an integration period (302) and reading of the pixel (303), The sensor includes a first connection switch (MUX1) controlled by a first logic signal (PROB1) for connecting the first connection node (NC) to a signal input (IN-DTX1) of a first current detection circuit (DTX1) or to a photodiode initialization voltage source (VRS), the first logic signal being in a first logic state or a second logic state, respectively; a sequence circuit (DM) of the sensor is configured to control, by means of said first current detection circuit (DTX1), a step (200) of detecting the start of exposure to an X-ray flash for triggering an image acquisition sequence (300); The image capture sequence (300) includes the following operations: a) commanding the photodiode initialization transistor (M1) to switch to an ON state in all pixels simultaneously; b) establishing a first logic signal (PROB1) in said first logic state, the effect of which is to inject, at a signal input (IN-DTX1) of a first current detection circuit, a current (I) collected at said first connection node (NC) originating from a photodiode of the pixel; c) when an output logic signal (OUT-DTX1) of the first current detection circuit switches from a first logic state to a second logic state in response to detection of an input current level exceeding a predetermined threshold, establishing said first logic signal (PROB1) to said second logic state, the effect of which is to connect said photodiode initialization transistor (M1) in a constantly on state to said initialization voltage source (VRS) and to trigger a global initialization phase (301) of an image capture sequence in order to initialize the photodiode prior to said integration phase; The intra-oral radiographic image sensor further comprises a second current detection circuit (DTX2) having a signal input (IN-DTX2) connected to a second connection node (NCE), the second connection node (NCE) being electrically isolated from the first connection node and connected to an initialization transistor (M1) of a set (E) of pixels of the matrix, the sequence circuit (DM) is configured to stop an integration phase of the image capture sequence (300) by an exposure end detection signal (OUT-DTX2) provided by a second current detection circuit (DTX2); The image sensor further comprises a second current detection circuit (DTX2) configured to activate an exposure end detection signal (OUT-DTX2) when a current (I') generated from a photodiode of a pixel of the set (E) falls below a predetermined threshold, the second current detection circuit having a signal output.

2. a first current detection circuit (DTX1) comprising a capacitive transimpedance amplifier (CTIA1) having a non-inverting input connected to a common mode voltage (Vcm1) and an inverting input connected to said first connection node (NC) and forming a signal input (IN-DTX1); The amplifier outputs a signal (V ｏｕｔ－ｃｔｉａ１ ), a clock signal (Φ ｃｔｉａ ), the lamp is controlled by a voltage threshold (V ｔｈ１ 2. The image sensor of claim 1, wherein the first and second inputs are applied to a comparator (COMP1) for comparison with the first and second inputs.

3. Each of the pixels includes the photodiode initialization transistor (M1), the gate of which is connected to the photodiode connection node (KN ｉ ), and the voltage follower transistor (M2) connected to the pixel read circuit (RD) and the individual column conductors (Col ｊ 3. The image sensor of claim 1, further comprising a read select transistor (M3) connected in series between said first input and said second input.

4. Each of the pixels includes the photodiode initialization transistor (M1), the photodiode connection node (KN ｉ ) and the pixel reading node (SN ｉ A charge transfer transistor (M4) is connected in series between the pixel read node (SN), an initialization transistor (M5) of the pixel read node (SN), and a voltage follower is mounted. ｉ ), the voltage follower transistor (M2) and the individual column conductors (Col ｊ 3. The image sensor of claim 1, further comprising a read select transistor (M3) connected in series between said first input and said second input.

5. a second connection circuit (MUX2) for connecting the second connection node (NCE) to a signal input (IN-DTX2) of a second current detection circuit or to an initialization voltage source (VRS); 2. The image sensor of claim 1, wherein the stopping of the integration phase of the image capture sequence (300) is controlled by the activation or deactivation of a second control logic signal (PROB2) applied to the second connection node (NCE).

6. 6. The image sensor of claim 5 in combination with claim 2, wherein the gain of a capacitive transimpedance amplifier and / or the voltage threshold of a comparator are adjustment parameters of the sensor adjusted with the first current detection circuit and the second current detection circuit to detect the start and end of an X-ray flash, respectively.

7. The sequence circuit (DM) is executed during the exposure to the X-ray flash and has the same integration period (d) as the preceding image acquisition sequence (300 (I-obs)) that provided the first image data. ＩＮＴ , d ＩＮＴＶ 7. The image sensor according to claim 1 , configured to apply a first image capture sequence (300(I-obs)) and to trigger other image capture sequences (300(I-obs)) using the same image capture pixels, allowing measurement and subtraction of dark noise of the first image data.

8. A method for capturing a radiographic image using an intraoral radiographic image sensor using MOS technology, comprising: The sensor includes a matrix of light-sensitive pixels (M-PIX) arranged in rows and columns, and each pixel (P ｉｊ ) includes a photodiode (PH) and a transistor, and the photodiode node (KN ｉ ) and a first connection node (NC) common to the pixels, It comprises, in a first step (200), connecting said first connection node (NC) to a signal input (IN-DTX1) of a first current detection circuit (DTX1) provided in the sensor for detecting when a current injected into the signal input (IN-DTX1) exceeds a predetermined threshold, corresponding to the detection of the start of exposure to an X-ray flash; said detection triggers an image capture sequence (300), the effect of which is to connect said first connection node (NC) to a voltage source (VRS), enabling a phase of global initialization of the photodiode (301), followed by an integration phase (302) during an integration period, and then a phase of reading the pixel (303); The method further comprises enabling the control of the integration period, during a current image capture sequence, by an exposure end detection signal provided by a second current detection circuit (DTX2) of the sensor, which has a signal input (IN-DTX2) connected to a second connection node (NCE), which is electrically insulated from said first connection node, which is connected to the photodiode initialization transistors (M1) of a set (E) of pixels of the matrix, a second current detection circuit (DTX2) configured to activate an exposure end detection signal (OUT-DTX2) when a current (I') generated from a photodiode of a pixel of the set falls below a predetermined threshold and having a signal output.

9. The radiation image capturing method of claim 8, wherein the control of the integration period by the exposure end detection signal is activated or deactivated by a second configurable control logic signal (PROB2) applied to a second connection circuit (MUX2) enabling a second connection node (NCE) to be connected to a signal input (IN-DTX2) of a second current detection circuit or an initialization voltage source (VRS).

10. A method according to any one of claims 8 to 9, wherein the first current detection circuit provided in the sensor is of the capacitive transimpedance amplifier and comparator type.

Images