diff --git a/newsfragments/645.feature b/newsfragments/645.feature
new file mode 100644
index 000000000..fb38a5121
--- /dev/null
+++ b/newsfragments/645.feature
@@ -0,0 +1 @@
+Add partial support for the Rigaku Oxford Diffraction file format.
diff --git a/src/dxtbx/format/FormatROD.py b/src/dxtbx/format/FormatROD.py
new file mode 100644
index 000000000..417f99f1a
--- /dev/null
+++ b/src/dxtbx/format/FormatROD.py
@@ -0,0 +1,459 @@
+"""Support for the Rigaku Oxford Diffraction image format
+
+NB: Rigaku datasets may use a non-zero-padded image incremental serial number.
+At present, this is not compatible with assumptions in dxtbx. In order to
+import these datasets, the images should be renumbered first.
+
+See https://github.com/cctbx/dxtbx/issues/646 for details."""
+
+# Takanori Nakane took David Waterman's code to parse headers from
+#  https://github.com/cctbx/cctbx_project/commit/b95467f3b2a70a37eeb820ea294128a32551700c
+# and heavily modified it. The original commit in the cctbx_project repository is orphan now.
+from __future__ import annotations
+
+__author__ = "David Waterman, Takanori Nakane"
+__copyright__ = (
+    "Copyright 2018 United Kingdom Research and Innovation & 2022 Takanori Nakane"
+)
+__license__ = "BSD 3-clause"
+
+import re
+import struct
+
+import numpy as np
+
+from scitbx.array_family import flex
+
+from dxtbx.format.Format import Format
+
+
+class FormatROD(Format):
+    """An image reading class for Rigaku Oxford Diffraction images.
+
+    The Rigaku Oxford Diffraction format is used by CrysAlis(Pro).
+    """
+
+    def __init__(self, image_file, **kwargs):
+        """Initialise the image structure from the given file."""
+
+        from dxtbx import IncorrectFormatError
+
+        if not self.understand(image_file):
+            raise IncorrectFormatError(self, image_file)
+
+        Format.__init__(self, image_file, **kwargs)
+
+    @staticmethod
+    def understand(image_file):
+        with FormatROD.open_file(image_file, "rb") as f:
+            try:
+                hdr = f.read(256).decode("ascii")
+            except UnicodeDecodeError:
+                return False
+
+        lines = hdr.splitlines()
+        if len(lines) < 2:
+            return False
+
+        vers = lines[0].split()
+        if len(vers) < 2 or vers[0] != "OD" or vers[1] != "SAPPHIRE":
+            return False
+
+        compression = lines[1].split("=")
+        if compression[0] != "COMPRESSION":
+            return False
+
+        return True
+
+    @staticmethod
+    def _read_ascii_header(image_file):
+        """Read the ASCII header comprising the first 256 bytes of the file"""
+
+        hd = {}
+        with FormatROD.open_file(image_file, "rb") as f:
+            hdr = f.read(256).decode("ascii")
+        lines = hdr.splitlines()
+
+        vers = lines[0].split()
+        if len(vers) < 2 or vers[0] != "OD" or vers[1] != "SAPPHIRE":
+            raise ValueError("Wrong header format")
+        hd["version"] = float(vers[-1])
+
+        compression = lines[1].split("=")
+        if compression[0] != "COMPRESSION":
+            raise ValueError("Wrong header format")
+        hd["compression"] = compression[1]
+
+        # Extract definitions from the 3rd - 5th line
+        defn = re.compile(r"([A-Z]+=[ 0-9]+)")
+        for line in lines[2:5]:
+            sizes = defn.findall(line)
+            for s in sizes:
+                n, v = s.split("=")
+                hd[n] = int(v)
+
+        hd["time"] = lines[5].split("TIME=")[-1].strip("\x1a").rstrip()
+
+        return hd
+
+    @staticmethod
+    def _read_binary_header(
+        image_file,
+        offset=256,
+        nbytes=5120,
+        general_nbytes=512,
+        special_nbytes=768,
+        km4gonio_nbytes=1024,
+        statistics_nbytes=512,
+        history_nbytes=2048,
+    ):
+        """Read the most relevant parameters from the binary header"""
+
+        with FormatROD.open_file(image_file, "rb") as f:
+            # General section
+            f.seek(offset)
+            bin_x, bin_y = struct.unpack("<hh", f.read(4))
+            f.seek(offset + 22)
+            chip_npx_x, chip_npx_y, im_npx_x, im_npx_y = struct.unpack(
+                "<hhhh", f.read(8)
+            )
+            f.seek(offset + 36)
+            num_points = struct.unpack("<I", f.read(4))[0]
+            if num_points != im_npx_x * im_npx_y:
+                raise ValueError("Cannot interpret binary header")
+
+            # Special section
+            f.seek(offset + general_nbytes + 56)
+            gain = struct.unpack("<d", f.read(8))[0]
+            f.seek(offset + general_nbytes + 464)
+            overflow_flag, overflow_after_remeasure_flag = struct.unpack(
+                "<hh", f.read(4)
+            )
+            f.seek(offset + general_nbytes + 472)
+            overflow_threshold = struct.unpack("<l", f.read(4))[0]
+            f.seek(offset + general_nbytes + 480)
+            exposure_time_sec, overflow_time_sec = struct.unpack("<dd", f.read(16))
+            f.seek(offset + general_nbytes + 548)
+            detector_type = struct.unpack("<l", f.read(4))[0]
+            f.seek(offset + general_nbytes + 568)
+            real_px_size_x, real_px_size_y = struct.unpack("<dd", f.read(16))
+
+            # Goniometer section
+            f.seek(offset + general_nbytes + special_nbytes + 284)
+            # angles for OMEGA, THETA, CHI(=KAPPA), PHI,
+            # OMEGA_PRIME (also called DETECTOR_AXIS; what's this?), THETA_PRIME
+            start_angles_steps = struct.unpack("<llllllllll", f.read(40))
+            end_angles_steps = struct.unpack("<llllllllll", f.read(40))
+            f.seek(offset + general_nbytes + special_nbytes + 368)
+            step_to_rad = struct.unpack("<dddddddddd", f.read(80))
+            f.seek(offset + general_nbytes + special_nbytes + 552)
+            # FIXME: I don't know what these are. Isn't the beam along e1 by definition??
+            beam_rotn_around_e2, beam_rotn_around_e3 = struct.unpack("<dd", f.read(16))
+            alpha1_wavelength = struct.unpack("<d", f.read(8))[0]
+            f.seek(offset + general_nbytes + special_nbytes + 640)
+            # detector rotation in degrees along e1, e2, e3
+            detector_rotns = struct.unpack("<ddd", f.read(24))
+            # FIXME: direct beam position when all angles are zero?
+            origin_px_x, origin_px_y = struct.unpack("<dd", f.read(16))
+            # alpha and beta are angles between KAPPA(=CHI) and THETA, and e3.
+            angles_in_deg = struct.unpack(
+                "<dddd", f.read(32)
+            )  # alpha, beta, gamma, delta
+            f.seek(offset + general_nbytes + special_nbytes + 712)
+            distance_mm = struct.unpack("<d", f.read(8))[0]
+
+        return {
+            "bin_x": bin_x,
+            "bin_y": bin_y,
+            "chip_npx_x": chip_npx_x,
+            "chip_npx_y": im_npx_y,
+            "im_npx_x": im_npx_x,
+            "im_npx_y": im_npx_y,
+            "gain": gain,
+            "overflow_flag": overflow_flag,
+            "overflow_after_remeasure_flag": overflow_after_remeasure_flag,
+            "overflow_threshold": overflow_threshold,
+            "exposure_time_sec": exposure_time_sec,
+            "overflow_time_sec": overflow_time_sec,
+            "detector_type": detector_type,
+            "real_px_size_x": real_px_size_x,
+            "real_px_size_y": real_px_size_y,
+            "start_angles_steps": start_angles_steps,
+            "end_angles_steps": end_angles_steps,
+            "step_to_rad": step_to_rad,
+            "beam_rotn_around_e2": beam_rotn_around_e2,
+            "beam_rotn_around_e3": beam_rotn_around_e3,
+            "alpha1_wavelength": alpha1_wavelength,
+            "detector_rotns": detector_rotns,
+            "origin_px_x": origin_px_x,
+            "origin_px_y": origin_px_y,
+            "angles_in_deg": angles_in_deg,
+            "distance_mm": distance_mm,
+        }
+
+    def _start(self):
+        """Open the image file, read useful metadata into an internal dictionary
+        self._header_dictionary"""
+
+        self._txt_header = self._read_ascii_header(self._image_file)
+        self._bin_header = self._read_binary_header(self._image_file)
+
+        return
+
+    # Rigaku/Oxford Geometry:
+    # - e3: parallel to OMEGA (and THETA)
+    # - e1: crystal to source
+    # - e2: completes the right handed coordinate system
+
+    def get_goniometer(self, index=None):
+        return Format.get_goniometer(self)
+
+    def _goniometer(self):
+        # FIXME: sometimes XDS.INP generated by CrysAlisPro has
+        #  tiny deviationis from (0, 1, 0). I don't know how to calculate it.
+        # DIALS' third axis points to the opposite of XDS; thus the direction will be opposite.
+        direction = (0.0, -1.0, 0.0)
+
+        # FIXME: represent kappa axis in fixed_rotation
+        return self._goniometer_factory.known_axis(direction)
+
+    def get_beam(self, index=None):
+        return Format.get_beam(self)
+
+    def _beam(self):
+        return self._beam_factory.make_polarized_beam(
+            sample_to_source=(0.0, 0.0, 1.0),
+            wavelength=self._bin_header["alpha1_wavelength"],
+            polarization=(0, 1, 0),
+            polarization_fraction=0.5,
+        )
+
+    def get_detector(self, index=None):
+        return Format.get_detector(self)
+
+    def _detector(self):
+        gonio_angles = np.array(self._bin_header["start_angles_steps"]) * np.array(
+            self._bin_header["step_to_rad"]
+        )
+        detector_rotns_rad = np.array(self._bin_header["detector_rotns"]) / 180 * np.pi
+        rot_e1 = np.array(
+            [
+                np.cos(detector_rotns_rad[0]),
+                np.sin(detector_rotns_rad[0]),
+                0,
+                -np.sin(detector_rotns_rad[0]),
+                np.cos(detector_rotns_rad[0]),
+                0,
+                0,
+                0,
+                1,
+            ]
+        ).reshape(3, 3)
+        rot_e2 = np.array(
+            [
+                1,
+                0,
+                0,
+                0,
+                np.cos(detector_rotns_rad[1]),
+                np.sin(detector_rotns_rad[1]),
+                0,
+                -np.sin(detector_rotns_rad[1]),
+                np.cos(detector_rotns_rad[1]),
+            ]
+        ).reshape(3, 3)
+        rot_theta = np.array(
+            [
+                np.cos(gonio_angles[1]),
+                0,
+                np.sin(gonio_angles[1]),
+                0,
+                1,
+                0,
+                -np.sin(gonio_angles[1]),
+                0,
+                np.cos(gonio_angles[1]),
+            ]
+        ).reshape(3, 3)
+        detector_axes = rot_theta.dot(rot_e2.dot(rot_e1))
+        # The third axis points to the opposite of XDS.
+        detector_axes[2, :] *= -1
+
+        pixel_size_x = self._bin_header["real_px_size_x"]
+        pixel_size_y = self._bin_header["real_px_size_y"]
+        origin_at_zero = np.array(
+            [
+                -self._bin_header["origin_px_x"] * pixel_size_x,
+                +self._bin_header["origin_px_y"] * pixel_size_y,
+                +self._bin_header["distance_mm"],
+            ]
+        )
+
+        # FIXME: this formula seem to give the right answer but I don't know why.
+        # XDS exporter in IPR's CrysAlisPro seems broken. It writes the same ORGX/Y
+        # regardless of the theta angles.
+        origin = detector_axes.dot(origin_at_zero)
+
+        detector = self._detector_factory.make_detector(
+            "PAD",
+            detector_axes[:, 0],
+            -detector_axes[:, 1],
+            origin,
+            (pixel_size_x, pixel_size_y),
+            (self._txt_header["NX"], self._txt_header["NY"]),
+            (0, self._bin_header["overflow_threshold"]),
+        )  # not sure about min
+
+        return detector
+
+    def get_scan(self, index=None):
+        return Format.get_scan(self)
+
+    def _scan(self):
+        """Return the scan information for this image."""
+
+        for axis in [0, 3]:  # 0 - OMEGA, 3 - PHI: FIXME: is it always PHI scan?
+            start_angle = (
+                self._bin_header["start_angles_steps"][axis]
+                * self._bin_header["step_to_rad"][axis]
+                / np.pi
+                * 180
+            )
+            end_angle = (
+                self._bin_header["end_angles_steps"][axis]
+                * self._bin_header["step_to_rad"][axis]
+                / np.pi
+                * 180
+            )
+            if start_angle != end_angle:
+                break
+
+        return self._scan_factory.single_file(
+            filename=self._image_file,
+            exposure_times=self._bin_header["exposure_time_sec"],
+            osc_start=start_angle,
+            osc_width=end_angle - start_angle,
+            epoch=None,
+        )
+
+    def get_raw_data(self):
+        comp = self._txt_header["compression"].strip()
+        if comp.startswith("TY6"):
+            return self._get_raw_data_ty6()
+        else:
+            raise NotImplementedError("Can't handle compression: {0}".format(comp))
+
+    def _get_raw_data_ty6(self):
+        offset = self._txt_header["NHEADER"]
+        nx = self._txt_header["NX"]
+        ny = self._txt_header["NY"]
+        with open(self._image_file, "rb") as f:
+            f.seek(offset)
+            lbytesincompressedfield = struct.unpack("<l", f.read(4))[0]
+            linedata = np.fromfile(f, dtype=np.uint8, count=lbytesincompressedfield)
+            offsets = struct.unpack("<%dI" % ny, f.read(4 * ny))
+
+            image = np.zeros((ny, nx), dtype=np.int32)
+            for iy in range(ny):
+                image[iy, :] = self.decode_TY6_oneline(linedata[offsets[iy] :], nx)
+            return flex.int(image)
+
+    def decode_TY6_oneline(self, linedata, w):
+        """Decompress TY6 encoded pixels for a single line.
+        w is the number of pixels in the fast axis."""
+
+        BLOCKSIZE = 8
+        SHORT_OVERFLOW = np.int32(254)
+        LONG_OVERFLOW = 255
+        SHORT_OVERFLOW_SIGNED = SHORT_OVERFLOW - 127
+        LONG_OVERFLOW_SIGNED = LONG_OVERFLOW - 127
+
+        ipos = 0
+        opos = 0
+        ret = np.zeros(w, dtype=np.int32)
+
+        nblock = (w - 1) // (BLOCKSIZE * 2)
+        nrest = (w - 1) % (BLOCKSIZE * 2)
+
+        firstpx = linedata[ipos]
+        ipos += 1
+        if firstpx < SHORT_OVERFLOW:
+            ret[opos] = firstpx - 127
+        elif firstpx == LONG_OVERFLOW:
+            ret[opos] = linedata[ipos : (ipos + 4)].view(np.int32)[0]
+            ipos += 4
+        else:
+            ret[opos] = linedata[ipos : (ipos + 2)].view(np.int16)[0]
+            ipos += 2
+        opos += 1
+
+        for k in range(nblock):
+            bittype = linedata[ipos]
+            nbits = (bittype & 15, (bittype >> 4) & 15)
+            ipos += 1
+            # ipos_bit = ipos * 8
+
+            for i in range(2):
+                nbit = nbits[i]
+                # 0: 0
+                # 1: 1 or 0
+                # 2: -1, 0, 1, 2
+                # 3: -3, -2, -1, 0, 1, 2, 3, 4 (= b111 -3 = 7 - 3)
+                zero_at = 0
+                if nbit > 1:
+                    zero_at = (1 << (nbit - 1)) - 1
+
+                v = 0
+                for j in range(nbit):
+                    v |= linedata[ipos] << (8 * j)
+                    ipos += 1
+
+                mask = (1 << nbit) - 1
+                for j in range(BLOCKSIZE):
+                    ret[opos] = ((v >> (nbit * j)) & mask) - zero_at
+                    opos += 1
+
+            for i in range(opos - BLOCKSIZE * 2, opos):
+                offset = ret[i]
+
+                # surprisingly, this IF is very slow!
+                if offset >= SHORT_OVERFLOW_SIGNED:
+                    if offset >= LONG_OVERFLOW_SIGNED:
+                        offset = linedata[ipos : (ipos + 4)].view(np.int32)[0]
+                        ipos += 4
+                    else:
+                        offset = linedata[ipos : (ipos + 2)].view(np.int16)[0]
+                        ipos += 2
+
+                ret[i] = offset + ret[i - 1]
+
+        for i in range(nrest):
+            px = linedata[ipos]
+            ipos += 1
+            if px < SHORT_OVERFLOW:
+                ret[opos] = ret[opos - 1] + px - 127
+            elif px == LONG_OVERFLOW:
+                ret[opos] = (
+                    ret[opos - 1] + linedata[ipos : (ipos + 4)].view(np.int32)[0]
+                )
+                ipos += 4
+            else:
+                ret[opos] = (
+                    ret[opos - 1] + linedata[ipos : (ipos + 2)].view(np.int16)[0]
+                )
+                ipos += 2
+            opos += 1
+
+        return ret
+
+
+if __name__ == "__main__":
+    import sys
+
+    for arg in sys.argv[1:]:
+        if FormatROD.understand(arg):
+            reader = FormatROD(arg)
+            print(reader._txt_header)
+            print(reader._bin_header)
+        else:
+            print("Unsupported format")