文档

JPEGD

The JPEG decoder (JPEGD) can decode .jpg, .jpeg, .JPG, and .JPEG images into YUV images. For details about the JPEGD functions and restrictions, see Functions and Restrictions.

This section describes the API call sequence of JPEGD, and sample code is also provided to help you better understand the sequence.

API Call Sequence

If JPEG image decoding is involved during app development, ensure that your app contains the code logic for such image decoding. For details about the API call sequence, see pyACL API Call Sequence.

Figure 1 JPEG image decoding

The current system supports .jpg, .jpeg, .JPG, and .JPEG image decoding. Different source image encoding formats lead to different output formats. The key APIs are described as follows:

  1. Call acl.media.dvpp_create_channel_desc to create a channel description.
  2. Call acl.media.dvpp_create_channel to create a channel for image data processing.
  3. Before decoding JPEG images, call acl.media.dvpp_malloc to allocate device buffers for storing the input or output data.

    Before allocating an output buffer, call acl.media.dvpp_jpeg_predict_dec_size to predict the required output allocation size based on the JPEG image size.

  4. Call the asynchronous API acl.media.dvpp_jpeg_decode_async for decoding.

    Also call acl.rt.synchronize_stream to block the application until all tasks in the specified stream are complete.

  5. After the decoding is complete, call acl.media.dvpp_free to free the input and output buffers in a timely manner.
  6. Call acl.media.dvpp_destroy_channel to destroy the image data processing channel.

    Call acl.media.dvpp_destroy_channel_desc to destroy the channel description after the channel is destroyed.

Sample Code

You can view the complete code in Sample Obtaining.

After APIs are called, add an exception handling branch, and record error logs and warning logs. The following is a code snippet of key steps only, which is not ready to use.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
 
import acl
# ......

# 1. Initialize pyACL.
ret = acl.init()

# 2. Allocate runtime resources.

# 3. Create a description of the data processing channel. self.dvpp_channel_desc is of type acldvppChannelDesc.
self.dvpp_channel_desc = acl.media.dvpp_create_channel_desc()

# 4. Create a data processing channel.
ret = acl.media.dvpp_create_channel(self.dvpp_channel_desc)

# 5. Allocate buffers.
# Load the YUV image data.
# Allocate the device buffer inputDevBuff.
# Copy image data from the host to the device by calling acl.rt.memcpy. After data copy is complete, free the host memory in a timely manner.
np_jpg = np.fromfile(self.path, dtype=np.byte)
self.in_buffer_size = np_jpg.itemsize * np_jpg.size
bytes_data = np_jpg.tobytes()
np_jpg_ptr = acl.util.bytes_to_ptr(bytes_data)
self.in_buffer_dev, ret = acl.media.dvpp_malloc(self.in_buffer_size)
ret = acl.rt.memcpy(self.in_buffer_dev, self.in_buffer_size, np_jpg_ptr,
                     self.in_buffer_size, 1)

# 6 Calculate and allocate the output device buffer self.out_buffer_dev in 16-pixel-aligned mode to store the encoded output data.
height_stride = ((h + 1) // 2) * 2
width_stride = ((w + 15) // 16) * 16
# malloc for output
self.out_buffer_size = (width_stride * height_stride * 3) // 2
self.out_buffer_dev, ret = acl.media.dvpp_malloc(self.out_buffer_size)

# 7. Create the description of the decoded image and set the attribute values.
# Set the output buffer and size to the picture description for data placeholder.
self.pic_desc = acl.media.dvpp_create_pic_desc()
ret = acl.media.dvpp_set_pic_desc_data(self.pic_desc, self.out_buffer_dev)
ret = acl.media.dvpp_set_pic_desc_size(self.pic_desc, self.out_buffer_size)
ret = acl.media.dvpp_set_pic_desc_format(
            self.pic_desc, PIXEL_FORMAT_YUV_SEMIPLANAR_420)
ret = acl.media.dvpp_set_pic_desc_width(self.pic_desc, w)
ret = acl.media.dvpp_set_pic_desc_height(self.pic_desc, h)
ret = acl.media.dvpp_set_pic_desc_width_stride(self.pic_desc, width_stride)
ret = acl.media.dvpp_set_pic_desc_height_stride(self.pic_desc, height_stride)

# 8. Perform asynchronous decoding and call acl.rt.synchronize_stream to block the host until all tasks in the specified stream are complete.
ret = acl.media.dvpp_jpeg_decode_async(self.dvpp_channel_desc,
                                           self.in_buffer_dev,
                                           self.in_buffer_size,
                                           self.pic_desc, self.stream)
ret = acl.rt.synchronize_stream(self.stream)

# 9. After the decoding is complete, destroy the allocations, including the description of the output image, output buffer, description of the channel as well as the channel.
ret = acl.media.dvpp_destroy_pic_desc(self.pic_desc)
ret = acl.media.dvpp_free(self.in_buffer_dev)
ret = acl.media.dvpp_free(self.out_buffer_dev)
ret = acl.media.dvpp_destroy_channel(self.dvpp_channel_desc)
ret = acl.media.dvpp_destroy_channel_desc(self.dvpp_channel_desc)

# 10. Destroy runtime allocations.

# Deinitialize pyACL.
ret = acl.finalize()

# ......
Parent topic: Media Data Processing V1