整理去年做的小项目，纪念我的图像处理入门。
因为要在DSP上实现，所以完全手工C代码垒起来的，还要保证和PC端跑的结果一样，觉得可能特殊场景会有用，上传github，没有依赖任何库；
格式注释什么的暂时没有整理，没有优化，能用，但是可能不够快，也没有考虑内存开销的优化，直接就在DDR2/SRAM上跑的。
Harris c代码实现 因为要匹配，最后对检测到的角点做了SAD（效率高）
matlab代码网上down的，自己添加下采样和匹配改了个面目全非。
DSP输出：
Harris start: Sample = 1 , Region = 8
max R = 32487164.72,(247, 3)
1 Harris: 178, 64
2 Harris: 186, 83
3 Harris: 191, 85
4 Harris: 195, 86
5 Harris: 67, 88
6 Harris: 217, 90
7 Harris: 58, 91
8 Harris: 52, 95
9 Harris: 241, 98
10 Harris: 17,112
11 Harris: 9,114
12 Harris: 217,115
13 Harris: 43,119
14 Harris: 233,119
15 Harris: 239,123
16 Harris: 228,126
17 Harris: 11,128
18 Harris: 234,130
19 Harris: 64,133
20 Harris: 240,133
21 Harris: 61,135
22 Harris: 225,135
23 Harris: 233,138
24 Harris: 204,139
25 Harris: 176,140
26 Harris: 119,141
27 Harris: 147,141
28 Harris: 60,142
29 Harris: 90,142
30 Harris: 30,143
31 Harris: 241,144
32 Harris: 238,145
33 Harris: 225,150
34 Harris: 233,151
35 Harris: 167,153
36 Harris: 201,153
37 Harris: 134,154
38 Harris: 100,155
39 Harris: 66,156
40 Harris: 221,156
41 Harris: 31,157
42 Harris: 239,163
43 Harris: 24,168
44 Harris: 237,174
45 Harris: 34,179
46 Harris: 65,183
47 Harris: 241,185
48 Harris: 11,188
49 Harris: 40,191
50 Harris: 70,191
51 Harris: 12,200
52 Harris: 30,200
53 Harris: 26,202
54 Harris: 46,207
55 Harris: 165,210
56 Harris: 222,213
57 Harris: 93,215
58 Harris: 218,217
59 Harris: 163,222
60 Harris: 11,224
61 Harris: 233,226
62 Harris: 227,230
63 Harris: 15,232
64 Harris: 18,234
65 Harris: 239,240
66 Harris: 16,243
67 Harris: 31,244
68 Harris: 23,245
69 Harris: 194,245
70 Harris: 210,246
71 Harris: 233,246
max R = 8346880.08,(1, 113)
1 Harris: 137, 9
2 Harris: 154, 10
3 Harris: 140, 24
4 Harris: 171, 64
5 Harris: 175, 65
6 Harris: 148, 67
7 Harris: 49, 75
8 Harris: 88, 80
9 Harris: 174, 82
10 Harris: 117, 83
11 Harris: 178, 83
12 Harris: 184, 83
13 Harris: 80, 84
14 Harris: 188, 85
15 Harris: 104, 86
16 Harris: 193, 86
17 Harris: 69, 87
18 Harris: 63, 88
19 Harris: 205, 89
20 Harris: 104, 90
21 Harris: 212, 90
22 Harris: 55, 91
23 Harris: 67, 93
24 Harris: 48, 94
25 Harris: 42, 95
26 Harris: 175, 96
27 Harris: 38, 97
28 Harris: 122, 97
29 Harris: 236, 98
30 Harris: 32, 99
31 Harris: 205,104
32 Harris: 13,105
33 Harris: 44,105
34 Harris: 197,105
35 Harris: 212,105
36 Harris: 44,107
37 Harris: 34,110
38 Harris: 225,110
39 Harris: 199,111
40 Harris: 10,112
41 Harris: 65,112
42 Harris: 204,112
43 Harris: 83,114
44 Harris: 210,115
45 Harris: 217,117
46 Harris: 35,119
47 Harris: 226,119
48 Harris: 245,120
49 Harris: 192,123
50 Harris: 231,123
51 Harris: 219,125
52 Harris: 228,127
53 Harris: 185,129
54 Harris: 225,131
55 Harris: 54,132
56 Harris: 230,132
57 Harris: 208,134
58 Harris: 46,135
59 Harris: 51,135
60 Harris: 215,135
61 Harris: 201,136
62 Harris: 194,139
63 Harris: 222,139
64 Harris: 137,140
65 Harris: 165,140
66 Harris: 183,140
67 Harris: 108,141
68 Harris: 19,142
69 Harris: 34,142
70 Harris: 49,142
71 Harris: 79,142
72 Harris: 118,142
73 Harris: 203,142
74 Harris: 214,143
75 Harris: 230,144
76 Harris: 178,145
77 Harris: 227,145
78 Harris: 238,145
79 Harris: 163,146
80 Harris: 155,147
81 Harris: 180,149
82 Harris: 213,150
83 Harris: 221,151
84 Harris: 61,152
85 Harris: 154,153
86 Harris: 188,153
87 Harris: 121,154
88 Harris: 232,154
89 Harris: 87,155
90 Harris: 53,156
91 Harris: 208,156
92 Harris: 18,157
93 Harris: 74,157
94 Harris: 189,159
95 Harris: 161,162
96 Harris: 127,163
97 Harris: 166,163
98 Harris: 225,163
99 Harris: 184,164
100 Harris: 33,166
101 Harris: 52,166
102 Harris: 58,168
103 Harris: 9,169
104 Harris: 41,169
105 Harris: 218,170
106 Harris: 27,171
107 Harris: 56,172
108 Harris: 17,173
109 Harris: 135,173
110 Harris: 25,174
111 Harris: 66,174
112 Harris: 50,175
113 Harris: 222,175
114 Harris: 98,176
115 Harris: 56,177
116 Harris: 234,177
117 Harris: 17,178
118 Harris: 29,178
119 Harris: 116,179
120 Harris: 44,180
121 Harris: 131,180
122 Harris: 244,180
123 Harris: 234,181
124 Harris: 32,182
125 Harris: 14,183
126 Harris: 48,183
127 Harris: 167,183
128 Harris: 131,184
129 Harris: 140,184
130 Harris: 241,184
131 Harris: 215,186
132 Harris: 19,187
133 Harris: 223,187
134 Harris: 148,188
135 Harris: 27,189
136 Harris: 23,190
137 Harris: 31,191
138 Harris: 52,191
139 Harris: 120,192
140 Harris: 147,194
141 Harris: 167,195
142 Harris: 187,195
143 Harris: 34,197
144 Harris: 220,197
145 Harris: 193,198
146 Harris: 29,199
147 Harris: 225,199
148 Harris: 180,200
149 Harris: 189,200
150 Harris: 144,201
151 Harris: 16,203
152 Harris: 110,203
153 Harris: 147,203
154 Harris: 230,203
155 Harris: 9,204
156 Harris: 88,205
157 Harris: 182,205
158 Harris: 191,205
159 Harris: 206,206
160 Harris: 26,207
161 Harris: 144,210
162 Harris: 244,210
163 Harris: 194,211
164 Harris: 56,212
165 Harris: 202,212
166 Harris: 95,213
167 Harris: 12,215
168 Harris: 31,215
169 Harris: 72,215
170 Harris: 182,216
171 Harris: 196,216
172 Harris: 104,217
173 Harris: 112,217
174 Harris: 99,219
175 Harris: 141,222
176 Harris: 233,222
177 Harris: 200,223
178 Harris: 9,224
179 Harris: 210,226
180 Harris: 15,229
181 Harris: 147,229
182 Harris: 204,230
183 Harris: 47,234
184 Harris: 126,234
185 Harris: 139,234
186 Harris: 9,235
187 Harris: 53,236
188 Harris: 212,238
189 Harris: 129,239
190 Harris: 206,239
191 Harris: 27,240
192 Harris: 12,244
193 Harris: 190,244
194 Harris: 244,244
195 Harris: 235,245
196 Harris: 29,247
197 Harris: 210,247
Matched points: 29
0: matchedCorner[174] (163,222,141,222) d: 0.03975
1: matchedCorner[160] (165,210,144,210) d: 0.05183
2: matchedCorner[168] ( 93,215, 72,215) d: 0.05275
3: matchedCorner[84] (167,153,154,153) d: 0.05445
4: matchedCorner[89] ( 66,156, 53,156) d: 0.05949
5: matchedCorner[137] ( 70,191, 52,191) d: 0.05978
6: matchedCorner[125] ( 65,183, 48,183) d: 0.06032
7: matchedCorner[69] ( 60,142, 49,142) d: 0.06281
8: matchedCorner[91] ( 31,157, 18,157) d: 0.06552
9: matchedCorner[159] ( 46,207, 26,207) d: 0.06596
10: matchedCorner[13] (191, 85,188, 85) d: 0.06981
11: matchedCorner[178] (233,226,210,226) d: 0.07079
12: matchedCorner[70] ( 90,142, 79,142) d: 0.07585
13: matchedCorner[135] ( 40,191, 23,190) d: 0.07670
14: matchedCorner[46] (233,119,226,119) d: 0.07697
15: matchedCorner[67] ( 30,143, 19,142) d: 0.07735
16: matchedCorner[64] (176,140,165,140) d: 0.07751
17: matchedCorner[181] (227,230,204,230) d: 0.07766
18: matchedCorner[58] ( 61,135, 51,135) d: 0.07878
19: matchedCorner[11] (186, 83,184, 83) d: 0.08188
20: matchedCorner[45] ( 43,119, 35,119) d: 0.08230
21: matchedCorner[54] ( 64,133, 54,132) d: 0.08530
22: matchedCorner[86] (134,154,121,154) d: 0.08576
23: matchedCorner[49] (239,123,231,123) d: 0.08604
24: matchedCorner[88] (100,155, 87,155) d: 0.09002
25: matchedCorner[63] (147,141,137,140) d: 0.09699
26: matchedCorner[82] (233,151,221,151) d: 0.09712
27: matchedCorner[90] (221,156,208,156) d: 0.09803
28: matchedCorner[176] (218,217,200,223) d: 0.09966
29: matchedCorner[43] (217,115,210,115) d: 0.10124
30: matchedCorner[97] (239,163,225,163) d: 0.10264
31: matchedCorner[15] (195, 86,193, 86) d: 0.10453
32: matchedCorner[81] (225,150,213,150) d: 0.10457
33: matchedCorner[39] ( 17,112, 10,112) d: 0.10529
34: matchedCorner[61] (204,139,194,139) d: 0.10692
35: matchedCorner[85] (201,153,188,153) d: 0.10697
36: matchedCorner[20] (217, 90,212, 90) d: 0.10785
37: matchedCorner[66] (119,141,108,141) d: 0.10903
38: matchedCorner[96] ( 24,168,166,163) d: 0.11074
39: matchedCorner[116] ( 34,179, 17,178) d: 0.11335
40: matchedCorner[17] ( 67, 88, 63, 88) d: 0.11478
41: matchedCorner[37] (194,245,225,110) d: 0.11824
42: matchedCorner[74] (241,144,230,144) d: 0.11879
43: matchedCorner[59] (225,135,215,135) d: 0.12120
44: matchedCorner[21] ( 58, 91, 55, 91) d: 0.12153
45: matchedCorner[76] (238,145,227,145) d: 0.12230
46: matchedCorner[ 3] (178, 64,171, 64) d: 0.12550
47: matchedCorner[28] (241, 98,236, 98) d: 0.12557
48: matchedCorner[27] ( 11,188,122, 97) d: 0.13124
49: matchedCorner[23] ( 52, 95, 48, 94) d: 0.14120
50: matchedCorner[164] (222,213,202,212) d: 0.14147
51: matchedCorner[37] ( 30,200,225,110) d: 0.14359
52: matchedCorner[71] ( 11,128,118,142) d: 0.14817
53: matchedCorner[165] ( 18,234, 95,213) d: 0.14993
54: matchedCorner[55] (240,133,230,132) d: 0.15049
55: matchedCorner[188] (241,185,129,239) d: 0.15167
56: matchedCorner[112] (237,174,222,175) d: 0.15237
57: matchedCorner[53] (234,130,225,131) d: 0.15975
58: matchedCorner[94] ( 12,200,161,162) d: 0.16515
59: matchedCorner[130] (210,246,215,186) d: 0.16592
60: matchedCorner[171] ( 15,232,104,217) d: 0.17298
61: matchedCorner[62] (233,138,222,139) d: 0.17425
62: matchedCorner[77] ( 26,202,238,145) d: 0.17961
63: matchedCorner[187] (239,240,212,238) d: 0.18350
64: matchedCorner[157] ( 9,114,191,205) d: 0.18354
65: matchedCorner[19] ( 31,244,104, 90) d: 0.18516
66: matchedCorner[196] (233,246,210,247) d: 0.19022
67: matchedCorner[50] (228,126,219,125) d: 0.19145
68: matchedCorner[174] ( 11,224,141,222) d: 0.19151
69: matchedCorner[167] ( 16,243, 31,215) d: 0.22552
70: matchedCorner[163] ( 23,245, 56,212) d: 0.23696
Harris end!
SURF c代码实现 为了利用积分图最后求特征向量的时候没有旋转图像，基本是参考OpenCV和网上down到的一份matlab以及C#拼起来的
matlab代码可能来自这里，记不清楚了，总之是网上不知名大神共享的，当时连OpenCV都没用过的小白我如获至宝...
DSP输出：
OpenSURF test:
Cycle to get ipt1 is 0
Number of key points in pic1:103
Cycle to describe ipt1 is 0
Number of key points in pic2:128
Cycle to get and describe ipt2 is 0
Time to matched: 0 ms
0: matchedIpt[95] ( 85.99, 85.17, 82.86, 85.30) d: 0.00260
1: matchedIpt[70] ( 47.06, 95.99, 42.83, 95.91) d: 0.00359
2: matchedIpt[82] ( 67.96,194.75, 49.92,194.57) d: 0.00455
3: matchedIpt[13] (214.44,125.26,205.99,125.16) d: 0.00532
4: matchedIpt[69] (202.97, 92.88,198.14, 92.96) d: 0.00570
5: matchedIpt[69] (205.31, 93.83,198.14, 92.96) d: 0.00662
6: matchedIpt[ 5] ( 35.38, 99.79, 29.98, 99.51) d: 0.00670
7: matchedIpt[54] ( 61.49,217.59, 38.79,217.87) d: 0.00806
8: matchedIpt[35] (152.79,179.11,136.93,179.25) d: 0.00806
9: matchedIpt[38] ( 58.68,183.02, 41.89,182.71) d: 0.00832
10: matchedIpt[92] ( 72.85,239.35, 48.47,239.21) d: 0.00904
11: matchedIpt[80] ( 73.11,186.04, 55.19,186.07) d: 0.00931
12: matchedIpt[ 3] (231.97, 93.13,227.17, 93.04) d: 0.00975
13: matchedIpt[25] (132.31,160.72,118.63,160.50) d: 0.01006
14: matchedIpt[30] (149.93,167.20,134.69,166.84) d: 0.01027
15: matchedIpt[ 8] ( 94.27,111.85, 87.44,111.56) d: 0.01030
16: matchedIpt[94] (122.94, 83.84,119.90, 83.62) d: 0.01062
17: matchedIpt[105] (176.30,191.17,157.86,191.50) d: 0.01115
18: matchedIpt[31] (223.54,168.70,208.99,168.44) d: 0.01183
19: matchedIpt[12] ( 46.80,120.28, 38.81,119.73) d: 0.01211
20: matchedIpt[71] (224.16,100.35,219.76,100.07) d: 0.01308
21: matchedIpt[44] (186.16,195.73,166.90,195.93) d: 0.01309
22: matchedIpt[77] (213.58,172.32,198.76,172.91) d: 0.01331
23: matchedIpt[107] (115.48,203.99, 95.71,203.96) d: 0.01420
24: matchedIpt[20] (169.80,146.23,158.25,146.20) d: 0.01550
25: matchedIpt[ 7] (232.83,101.16,227.14,100.89) d: 0.01680
26: matchedIpt[ 9] ( 86.30,113.77, 79.05,113.92) d: 0.01721
27: matchedIpt[66] (122.35, 83.45,119.29, 83.40) d: 0.01780
28: matchedIpt[ 6] ( 27.53,101.58, 22.73,101.44) d: 0.01916
29: matchedIpt[ 1] ( 60.16, 90.06, 56.41, 89.71) d: 0.01963
30: matchedIpt[37] (137.53,181.55,120.93,181.54) d: 0.01971
31: matchedIpt[58] (118.01,226.98, 94.54,227.09) d: 0.02034
32: matchedIpt[112] (171.38,220.73,149.69,221.25) d: 0.02042
33: matchedIpt[17] (230.51,135.97,219.99,135.96) d: 0.02077
34: matchedIpt[61] (110.96,230.61, 87.16,230.17) d: 0.02178
35: matchedIpt[125] (169.41,123.85,166.05,124.65) d: 0.02278
36: matchedIpt[52] ( 65.81,216.05, 46.05,215.23) d: 0.02281
37: matchedIpt[87] (168.93,219.45,148.37,220.08) d: 0.02330
38: matchedIpt[110] ( 70.02,219.18, 48.17,218.89) d: 0.02513
39: matchedIpt[32] (211.27,171.33,197.11,171.64) d: 0.02650
40: matchedIpt[115] ( 81.10,228.88, 57.83,229.34) d: 0.02689
41: matchedIpt[96] (231.62, 91.24,226.57, 88.88) d: 0.02925
42: matchedIpt[64] (108.60,243.74, 82.07,243.78) d: 0.03126
43: matchedIpt[22] (140.21,154.63,127.19,154.34) d: 0.03203
44: matchedIpt[95] ( 91.32, 84.99, 82.86, 85.30) d: 0.03230
45: matchedIpt[93] (146.93, 27.16,146.03, 27.53) d: 0.03469
46: matchedIpt[60] (134.77,229.37,110.14,228.31) d: 0.03591
47: matchedIpt[105] (175.91,189.85,157.86,191.50) d: 0.04027
48: matchedIpt[125] (174.17,124.58,166.05,124.65) d: 0.04138
49: matchedIpt[85] (199.25,207.93,178.58,207.88) d: 0.04370
50: matchedIpt[24] ( 37.63,159.41, 24.60,158.93) d: 0.04504
51: matchedIpt[97] (220.12,128.17,211.02,128.12) d: 0.05070
52: matchedIpt[25] (229.83,128.56,118.63,160.50) d: 0.05942
53: matchedIpt[78] (238.84,169.78,238.67,173.68) d: 0.07877
54: matchedIpt[108] (172.92,208.34,151.75,208.93) d: 0.07884
55: matchedIpt[39] (156.11,186.26,147.67,185.59) d: 0.08015
56: matchedIpt[55] (165.94,219.38,144.14,219.32) d: 0.08813
57: matchedIpt[27] (234.62,162.62,220.43,162.56) d: 0.09373
58: matchedIpt[14] (155.06,171.10,228.58,124.93) d: 0.09682
59: matchedIpt[17] ( 38.87,167.80,219.99,135.96) d: 0.09913
60: matchedIpt[10] (218.30,114.22,211.73,114.14) d: 0.10389
61: matchedIpt[121] (227.64,178.41,205.99,178.25) d: 0.10590
62: matchedIpt[38] ( 58.80,196.00, 41.89,182.71) d: 0.10633
63: matchedIpt[34] ( 36.21,174.12, 20.49,173.70) d: 0.10729
64: matchedIpt[121] (229.02,176.50,205.99,178.25) d: 0.11332
65: matchedIpt[14] (236.82,124.87,228.58,124.93) d: 0.11353
66: matchedIpt[120] (207.85,159.31,194.75,159.18) d: 0.11531
67: matchedIpt[114] ( 51.57,229.20, 33.27,226.18) d: 0.11878
68: matchedIpt[71] (224.11,101.66,219.76,100.07) d: 0.11981
69: matchedIpt[99] ( 43.98,156.20, 31.10,155.11) d: 0.12281
70: matchedIpt[99] (241.88,195.45, 31.10,155.11) d: 0.12402
71: matchedIpt[25] (235.89,145.88,118.63,160.50) d: 0.12459
72: matchedIpt[66] (110.62,216.96,119.29, 83.40) d: 0.12702
OpenSURF end!
其实写那些代码的时候根本不懂SURF的原理，依样画葫芦竟然还画的有模有样，当然理论水平还是要跟上去。
斑点检测 尺度空间 特征点方法综述 觉得这个博客特别好，很形象的解释。
大部分参考了《图像局部不变性特征与描述》。