标签：开发人员 使用 spin 技术 伸缩性 切换 https 半导体 自旋
* 这个例程演示了如何用Gmm分类器用面积和圆度来对柑橘类水果惊醒识别和分类，以及显示识别到的2d特征空间
*This example program shows how to apply a general GMM
* classification to distinguish citrus fruits using the
* features ‘area‘ and ‘circularity‘. Additionally, the
* 2D feature space for the extracted fruits is visualized.
* 读取图像
read_image (Image, ‘color/citrus_fruits_01‘)
*获取图像指针
get_image_pointer1 (Image, Pointer, Type, Width, Height)
*关闭窗体
dev_close_window ()
*打开新窗体
dev_open_window (0, 0, Width, Height, ‘white‘, WindowHandle)
*设置窗体显示字体样式
set_display_font (WindowHandle, 12, ‘mono‘, ‘true‘, ‘false‘)
*设置填充样式
dev_set_draw (‘margin‘)
*设置线宽
dev_set_line_width (2)
*显示图像
dev_display (Image)
*关闭窗体更新
dev_update_window (‘off‘)
*关闭’pc‘更新
dev_update_pc (‘off‘)
*变量不更新
dev_update_var (‘off‘)
*
FeaturesArea := []
FeaturesCircularity := []
ClassName := [‘orange‘,‘lemon‘]
* 创建GMM分类器句柄
* Create a GMM classifier
create_class_gmm (2, 2, 1, ‘spherical‘, ‘normalization‘, 10, 42, GMMHandle)
* 分别加入训练样本
* Add training samples
for I := 1 to 4 by 1
read_image (Image, ‘color/citrus_fruits_‘ + I$‘.2d‘)
dev_display (Image)
* ‘Add Samples‘添加样本
get_regions (Image, SelectedRegions)
*显示选择区域
dev_display (SelectedRegions)
*获取数量
count_obj (SelectedRegions, NumberObjects)
for J := 1 to NumberObjects by 1
select_obj (SelectedRegions, ObjectSelected, J)
*获取圆度和面积特征参数
get_features (ObjectSelected, WindowHandle, Circularity, Area, RowRegionCenter, ColumnRegionCenter)
*添加到数组
FeaturesArea := [FeaturesArea,Area]
FeaturesCircularity := [FeaturesCircularity,Circularity]
FeatureVector := real([Circularity,Area])
*前两副图像为橘子的训练图像
if (I <= 2)
*添加样本到橘子的分类ID
add_sample_class_gmm (GMMHandle, FeatureVector, 0, 0)
*显示信息
disp_message (WindowHandle, ‘Add to Class:‘ + ClassName[0], ‘window‘, RowRegionCenter, ColumnRegionCenter - 100, ‘black‘, ‘true‘)
*后两幅为柠檬的驯良图片
else
*添加样本到柠檬的分类ID
add_sample_class_gmm (GMMHandle, FeatureVector, 1, 0)
*显示信息
disp_message (WindowHandle, ‘Add to Class:‘ + ClassName[1], ‘window‘, RowRegionCenter, ColumnRegionCenter - 100, ‘black‘, ‘true‘)
endif
endfor
*右下角显示‘Press F5 to continue‘这个信息
disp_continue_message (WindowHandle, ‘black‘, ‘true‘)
stop ()
endfor
*清除窗体
dev_clear_window ()
* 分别显示橘子和柠檬的2D特征空间
* Visualize the feature space
visualize_2D_feature_space (Cross, Height, Width, WindowHandle, FeaturesArea[0:5], FeaturesCircularity[0:5], ‘dim gray‘, 18)
* ‘oranges‘, 40, 440
visualize_2D_feature_space (Cross, Height, Width, WindowHandle, FeaturesArea[6:11], FeaturesCircularity[6:11], ‘light gray‘, 18)
* ‘lemons‘, 70, 440
*右下角显示‘Press F5 to continue‘这个信息
disp_continue_message (WindowHandle, ‘black‘, ‘true‘)
stop ()
* 前面是添加样本数据，现在是训练样本
* Train the classifier
train_class_gmm (GMMHandle, 100, 0.001, ‘training‘, 0.0001, Centers, Iter)
*
* Classify
for I := 1 to 15 by 1
*读取图像
read_image (Image, ‘color/citrus_fruits_‘ + I$‘.2d‘)
*显示图像
dev_display (Image)
* ‘Classify Image‘, 10, 10
get_regions (Image, SelectedRegions)
*显示分割的区域
dev_display (SelectedRegions)
*数区域数量，一个一个识别
count_obj (SelectedRegions, NumberObjects)
for J := 1 to NumberObjects by 1
*选择区域
select_obj (SelectedRegions, ObjectSelected, J)
*获取圆度和面积特征
get_features (ObjectSelected, WindowHandle, Circularity, Area, RowRegionCenter, ColumnRegionCenter)
FeaturesArea := [FeaturesArea,Area]
FeaturesCircularity := [FeaturesCircularity,Circularity]
FeatureVector := real([Circularity,Area])
*执行飞来识别
classify_class_gmm (GMMHandle, FeatureVector, 1, ClassID, ClassProb, Density, KSigmaProb)
*显示识别到的物体信息
disp_message (WindowHandle, ‘Class: ‘ + ClassName[ClassID], ‘window‘, RowRegionCenter, ColumnRegionCenter - 100, ‘black‘, ‘true‘)
disp_message (WindowHandle, ‘KSigmaProb: ‘ + KSigmaProb, ‘window‘, RowRegionCenter + 30, ColumnRegionCenter - 100, ‘black‘, ‘true‘)
endfor
if (I != 15)
disp_continue_message (WindowHandle, ‘black‘, ‘true‘)
endif
stop ()
endfor
* 清除句柄，释放内存
* Clear the classifier from memory
clear_class_gmm (GMMHandle)
标签：开发人员 使用 spin 技术 伸缩性 切换 https 半导体 自旋
原文地址：https://blog.51cto.com/14618340/2518639
