Fig. 1. Experimental setup and principle. (a) An object $O_1$ placed in the far field of a 1-mm-thick $\beta$-Barium Borate (BBO) nonlinear crystal is illuminated by spatially entangled photon pairs produced via type-I spontaneous parametric down conversion (SPDC), while an object $O_2$ is illuminated by diffused classical light. A lens $f_1 = 50$ mm is positioned after the crystal to direct the photon-pairs towards $O_1$. Both objects are composed of an absorptive pattern layer on a reflective surface. They are imaged onto the SPAD camera using lens $f_2 = 100$ mm, $f_3 = 50$ mm, $f_4 = 100$ mm and an unbalanced beam splitter (0.1R/0.9T). (b) When the SPAD gate window is set to capture only photon pair pulses reflected by the quantum object, the “skater-shape” object appears in the intensity image and a peak is detected in the spatially-averaged correlation image, which shows the number of photon coincidences spatially averaged over all pair of pixels ${r_1}$ and ${r_2}$ separated by a given distance ${r_1}+{r_2}$. The correlation peak confirms the presence of photon pairs among the detected photons. (c) When the SPAD gate window is set to capture only classical light, the “car-shape” object appears in the intensity image whereas no peak obtained is visible in the spatially-averaged correlation image. Intensity and spatially-averaged correlation images were reconstructed from $N = 2000$ frames (8-bit) acquired using an exposure time of $350$ ns (1-bit). Intensity image coordinate units are in pixels.

Fig. 2. Results with synchronous classical light interference. (a) The reflected light from objects $O_1$ (“person") and $O_2$ (“bike") are both synchronous with the camera. (b) shows the selected intensity and spatially-averaged correlation images ($9\times 9$ central data) at the gate positions $0.09$ ns, $7.2$ ns, $14.58$ ns and $23.4$ ns covered with none reflected light, reflected light from only $O_1$, $O_1$ $\&$ $O_2$, and only $O_2$ respectively. Correlation peaks are obtained at $14.58$ ns and $23.4$ ns gate positions when there is quantum light reflected to the camera. The measurement is implemented over a time range of $27$ ns corresponding to $1500$ continuous gate positions with a proper time offset initially to the pump laser trigger. (c) Average intensity over all pixels (blue curve) and the peak coincidences (red curve) values along the measured time range. The peak coincidences are the normalized coincidence values at the center position (0, 0) of the spatially-averaged correlation images. The four positions in (b) are also marked on the horizontal axis of the curve. (d) The subtracted intensity image of $O_2$ (classical) and its arrival time ($16.110$ ns) to the camera by locating the first falling edge of the average intensity profile. (e) The subtracted intensity image of $O_1$ (quantum) and its arrival time ($24.462$ ns) to the camera by locating the falling edge of the correlation peak profile. Experiments are performed by $N = 5000$ frames (8-bit) acquired in $13.5$ s at each gate position using an exposure time of $350$ ns for 1-bit frame. The time step between two successive gate windows is $18$ ps. Intensity image coordinate unites are in pixels. See Visualization 1 for the entire scanning results.

Fig. 3. Measurement of spatially-resolved correlation images over time. The shape of the object illuminated by photon pairs (“person") is retrieved by measuring spatially-resolved correlation images at the gate positions $0.09$ ns, $7.2$ ns, $14.58$ ns and $23.4$ ns. Each image is obtained by acquiring $5$ million frames (8-bit), which corresponds to approximately $3.8$ hours of acquisition.

Fig. 4. Results with asynchronous spurious light. (a) Photon pairs reflected by object $O_1$ (“STOP traffic sign") is synchronous with the camera, while classical photons reflected by $O_2$ (“50 traffic sign") arrives at the camera in temporally random sequences as the classical laser is asynchronous. (b) The camera scanned over a time range of $27$ ns (1500 continuous gate positions). Intensity and spatially-averaged correlation images (central 9$\times$9 pixels area) are shown for three different gate positions ($2.16$ ns, $13.5$ ns and $24.66$ ns). The correlation peak only appears at the gate window covered with photon pairs reflected by $O_1$. (c) The corresponding three gate positions are marked in the curve of the average intensity and correlation peak responses over the detected time range. (d) Intensity image reconstructed by subtracting intensity image at $13.5$ ns by this at $24.66$ ns. At each gate position, $N = 3000$ frames (8-bit) were acquired in $8.1$ s using an exposure time of $350$ ns (1-bit). The time step between two successive gate positions is $18$ ps. Intensity image coordinate unites are in pixels. See Visualization 2 for the entire scanning results.