基于预处理图像压缩的病害诊断应用研究

Research on Disease Diagnosis Based on Image Preprocessing and Compression

在手机移动端进行病害图像远程诊断时,因大容量图像会导致上传速度缓慢、增大远程服务端数据处理难度,影响服务端对图像识别的及时性和准确性。针对此问题,提出基于预处理图像压缩方法,即对随时采集的图像进行适度裁剪,然后进行灰度化、自适应阈值分割并提取分割后的病斑区域,最后对病斑区域进行JPEG压缩即经过DCT变换、量化、编码等操作。以黄瓜白粉病、斑点病等为例,在移动端进行病害诊断,实现对病害图像的预处理和压缩过程。测试数据发现,病害图像经过预处理压缩后,在获得良好图像质量的前提下,大大减少图像数据。结果表明,在移动端应用预处理压缩方法能加快图像上传速度,降低远程服务端图像数据处理难度,提高诊断及时性,具有较高的实用性。

When remote diagnosis of image disease is carried out on mobile phone,high-capacity images can slow uploading speed and increase the data processing difficulty on the remote server,and affect the timeliness and accuracy of image recognition on the server at the same time. To solve this problem,a method of image compression based on pre-processing is proposed,that is,the image collected at any time is moderately clipped,and then gray-scale and adaptive threshold segmented and extracted the lesion area. Finally,JPEG compression of the lesion area is carried out through DCT transformation,quantization,coding and other operations. Taking cucumber powdery mildew and spotted disease as examples,the disease diagnosis operation is carried out at the mobile,and the disease image is pre-processed and compressed at the same time. The test data shows that the capacity of image data is greatly reduced under the premise of better image quality after preprocessing and compression. The results show that the application of the proposed method in the mobile can speed up image upload and reduce the difficulty of image data processing at remote server,and improve the timeliness of diagnosis with high practicability.